An Engineering Archive

An Engineering Archive

An Engineering Archive A Selection of Papers From the Proceedings of the Institution of Mechanical Engineers

Edited by Professor Desmond Winterbone Archive assistance from Keith Moore

Published by Mechanical Engineering Publications Limited, for the Institution of Mechanical Engineers, Bury St. Edmunds, and London.

First Published in this Edition 1997 This publication is copyright under the Berne Convention and the International Copyright Convention. All rights reserved. Apart from any fair dealing for the purpose of private study, research, criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, no part may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, electrical, chemical, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owners. Reprographic reproduction is permitted only in accordance with the terms of licences issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P 9HE. Unlicensed multiple copying of the contents of this publication is illegal. Inquiries should be addressed to: The Publishing Editor, Mechanical Engineering Publications Limited, Northgate Avenue, Bury St. Edmunds, Suffolk, IP32 6BW, UK. Tel: 01284 763277. Fax: 01284 704006.

The Institution of Mechanical Engineers 1997

Leather Bound Edition Hardback ppc Edition

ISBN 1 86058 052 1 ISBN 1 86058 053 X

A CIP catalogue record for this book is available from the British Library. Printed by Lavenham Press, Suffolk, UK.

The Publishers are not responsible for any statement made in this publication. Data, discussion, and conclusions developed by authors are for information only and are not intended for use without independent substantiating investigation on the part of potential users. Opinions expressed are those of the Author and are not necessarily those of the Institution of Mechanical Engineers or its Publishers.

Contents Foreword Preface Historical notes on the Proceedings of the Institution IMechE Library Bookplate Note on Subjects for Papers Subjects for Papers - Call for Papers for the Proceedings of the IMechE -1849 George Stephenson - On the fallacies of the rotary engine - 1848 John Scott Russell - Notice on the life and character of the late George Stephenson - 1848 Richard Peacock - On the workshops for the locomotive carriage and waggon departments of the Manchester, Sheffield, and Lincolnshire Railway - 1851 William Menelaus - Description of the large blowing engine and new rolling mill at Dowlais Iron Engineering Works - 1857 William George Armstrong - On water pressure machinery - 1858 Charles William Siemens - Description of a machine for covering telegraph wires with India-Rubber - 1860 Henry Bessemer - On the manufacture of cast steel and its application to constructive purposes - 1861 Francis William Crossley - On Otto and Langen's atmospheric gas engine, and some other gas engines - 1875 Thomas Russell Crampton - On an automatic hydraulic system far excavating the Channel Tunnel - 1882 Beauchamp Tower - First report onfrictionexperiments - 1883 Hiram Stephens Maxim - Description of the Maxim Automatic Machine-Gun - 1885 William Thomson - On ship waves - 1887 Charles Algernon Parsons - Description of the compound steam turbine and turbo-electric generator - 1888 Richard Francis Trevithick - Locomotive building in Japan - 1895

vii ix xiii xv xvi 1 5 13 27

37 51 79 95 115 135 151 177 201 233 249

Works Visits The Eiffel Tower - 1889 Daimler motor mills,, Coventry - 1897 The Colonial Consignment and Distributing Co.'s Frozen Australasian Meat Store, Lambeth - 1900 Henry Selby Hele-Shaw - Road locomotion - 1900 Rudolph Christian Karl Diesel - The Diesel oil-engine, and its industrial importance, particularly for Great Britain - 1912 Frederick William Lanchester - Theflyingmachine: the aerofoil in the light of theory and experiment - 1915 William Henry Bragg - The application of x-rays to the study of the crystalline structure of materials - 1927 Arthur Stanley Eddington - Engineering principles in the machinery of the stars - 1929 Herbert Nigel Gresley - High-pressure locomotives - 1931 Ernest Rutherford - Atomic projectiles and their applications - 1932 Reid Anthony Railton - Racing motor car design - 1937 Frank Whittle - The early history of the Whittle jet propulsion gas turbine - 1945 Harry Ralph Ricardo - Piston aeroengines - 1947 John Douglas Cockcroft - The possibilities of nuclear energy fir heat and power production - 1947

270 271 277 279 283 341 381 421 449 473 513 539 581 613 623

Foreword The 150th anniversary of the Institution, offers an excellent opportunity to look back at the contribution made by the Proceedings of the Institution to the whole area of mechanical engineering. The Institution has a number of complete sets of Proceedings, which are a record of the papers and reports presented to the members and others. These date back to 1848, when the founding president, George Stephenson, wrote his paper On the fallacies of the rotary engine. During the succeeding 150 years, most of the famous engineers who have made major contributions to our life and wealth, have published papers in Proceedings. This volume contains a selection of papers which have been chosen from the first 100 years of the Proceedings to exemplify both landmark accounts of engineering technology and science, and contributions from famous authors. This volume concentrates on the first 100 years, as it is possible to gauge the impact of these papers in a true historical context. The papers reproduced here are in facsimile, to show their original form. Some pages have been enlarged, and others slightly reduced in size for presentation in this volume. The quality of the diagrams is impressive, and these were often produced in the Institution's own drawing office. A number of the most memorable plates are included here and show the Institution's contribution to graphic representation. An Engineering Archive is a fitting commemoration of our Institution's history. I hope that those who use this interesting and fascinating book will enjoy reading this selection from the Institution's unique archive. Ernest Shannon, FEng President, IMechE January 1997

vii

This page intentionally left blank

Preface The IMechE reached its 150th anniversary in 1997 and one of the ways proposed to celebrate this occasion was to publish a compilation of papers from the Institution throughout its history. I was invited to act as Editor for this volume because at that time I was Chairman of the meeting of Editors of the Proceedings. It was obvious that the task was too big for one person to undertake, especially one who was not a specialist in the archives of the Institution. I invited Keith Moore, the Institution Archivist, to act as a coeditor. The task of selecting the most appropriate papers was a daunting one, as can be imagined. In the years to 1950 alone, the Proceedings contained written work from over 1,880 engineers, and the rate of publication has increased enormously in recent years. The problem was reduced somewhat when Professor Duncan Dowson, FRS, FEng, and ex-president, suggested that it might be advisable to restrict consideration to the first 100 years of the Proceedings, on the grounds that it is difficult to judge the impact of papers until they have stood the test of time. This proposal rendered the task more manageable. The whole of this period is summarised in "A Brief Index of Papers 1847-1950", which was compiled in 1951 by Alfred Stock, FLA, who was the Institution's Chartered Librarian. The proposal was that the compilation should contain the "best" papers published over this period, but it soon became apparent that it would be impossible to bring together a set of papers that all readers would class as the best. In consultation with a number of senior members of the Institution, I put together the following guidelines to be used in selecting papers, they should: i. ii. iii. iv. v.

come from the first 100 years (i.e. 1847-1947); reflect as fairly as possible the full range of Institution activities and interests; be papers which made a significant contribution to engineering at their time of publication; be distributed reasonably evenly over the first 100 years; contain contributions from eminent engineers.

These terms of reference have given us reasonable freedom to include the papers which are contained herein, and to some extent these papers reflect our interpretation of how various sectors of engineering developed during the formative years of the Institution. We were greatly helped by the Fellows of the

ix

Royal Academy of Engineering to whom we wrote at the beginning of the task. They suggested a large number of papers, many of which are included in this volume. We thank all those who made suggestions for their help and hope that we have not offended anyone by not including their recommendations. We must also thank those at MEP and IMechE, both past and present, for their help in this task. The Librarians at Birdcage Walk have maintained a number of complete sets of proceeding over the past 150 years and during two World Wars, a true testimony to their stewardship during this time. Turning now to the 24 papers which have been selected for this volume. We have taken advantage of this opportunity to include brief biographies of the authors and, where possible, we have included portraits. We hope that these are interesting additions to the material from the Proceedings, and they help to place the individual papers in context. Obviously, the small number has been a major restriction on the papers which can be included, and every attempt has been made to get a good coverage of the subject of mechanical engineering, and of the contribution made by the Institution. We trust that readers will agree with our choice. The papers published during this first hundred years differ significantly from those of today in a number of aspects. The phrasing and presentation obviously mirror those of the era, and contrasting styles of the first two papers are fine examples of their time. The explanation of the operation of the slider crank mechanism by Stephenson is relatively tortuous but shows that even 150 years ago engineers were aware of the fallacy of the argument that reciprocating motion was inefficient. The second paper, the obituary of George Stephenson by Scott Russell is an eloquent appreciation of the impact of one life on a whole way of life. It is difficult now to appreciate how the railway improved communication. Before the advent of rail travel many journeys were unique events in a lifetime. Afterwards they became commonplace. The "information technology" of the 1850s was achieved by transporting the people to meetings and distributing the printed word by post. The great advances in mechanical engineering could not have come about without the production of inexpensive steel. The contribution of the mechanical engineer in this area is exemplified by the descriptions of the Dowlais Iron Works (Menelaus), and the Bessemer process. As Scott Russell says, "most of us can remember when the idea of laying wrought-iron bars of 50, 70, or 90 Ibs. weight per yard, for continuous miles, was an expense so utterly beyond the conception of the time as not to be entertained for a moment... At this early stage power was provided either manually or by steam engine and line drive. This was to change

x

with the development of "water pressure machinery" by W G Armstrong, the introduction of the internal combustion engine by Crossley, and later the steam turbine by Parsons. The motive power to drive the British economy had now been put into place, and there were no major new inventions in powerplant until Whittle made his first successful jet engine, and Cockcroft realised the potential of civil nuclear power. However, British mechanical engineers continued to make major contributions to the development of state of the art through the work of Lanchester on aircraft, Railton on racing cars, and Gresley (and others), on locomotives. In the area of transport in general, it is fascinating to see the presentation by Crampton (1882) on a scheme for excavating the Channel Tunnel - it took well over a century to achieve this dream! In addition to making artefacts, mechanical engineers also took an interest in the "science" of their subject, and this is demonstrated by a number of the papers included. The early work by Beauchamp Tower gives an interesting insight into the study of hydrodynamic lubrication, and laid the foundation of a whole area of tribology. The excellent paper on ship waves by Thomson, better known as Lord Kelvin, shows how engineers were attempting to improve the "efficiency" of marine transport by detailed hull design. The vote of thanks to this paper is fascinating, where it is stated 'lt has always been said that "Britannia rules the waves"; and when it is considered how Great Britain does three-fourths of the carrying trade of the world...'. This shows the enormous influence Britain had in the world at that time, and mechanical engineering played a major role in maintaining this. Later, Bragg described his work on crystallography, which revealed the internal structure of materials to the engineers who needed them to achieve their designs - new materials are still an area of great potential for mechanical engineers, and we must continue to maintain our interest in this area. Two other scientific papers which testify to the catholic interests of engineers are those by Eddington, who described the machinery of the stars, and Rutherford who reported his experiments on the atomic scale. Very soon after these scientific presentations, mechanical engineers had taken advantage of this knowledge and ventured into space on the one hand, and harnessed the power of the atom on the other. Finally, we turn to two completely different papers, those of Siemens and Maxim. The first, by Siemens, who came to Britain from Germany in 1843, and made his fortune from the invention of the water-meter. We have chosen to include his paper describing the manufacture of insulated cables which revolutionised telegraphic communications - again mechanical engineers made a major contribution to the transfer of information. The second is by Maxim on the machine-gun, which epitomises the way in which mechanical engineers approach a problem. Maxim carefully analysed the problems of the existing guns

xi

and devised a fail-safe mechanism by which the energy released by the explosive charge in the bullet could be used to drive the gun. The detailed design shown in his drawings is a fine testament to the draughting skills existing at that time. We hope you will have as much pleasure reading this selection from the Proceedings as we had putting it together. The mechanical engineer played a major role in developing our society, and improving the standard of living for all. Engineers provide the means of generating the wealth that enables politicians to achieve their dreams and ours. This collection of papers shows the impact that was made in the first century which continued through the next fifty years. It is the challenge which we must pass to the next generation. Undoubtedly engineering will become more interdisciplinary, but there is still a major role for mechanical engineers to play in nano-technology, fluid mechanics and heat transfer, the efficient use of energy, the use of new materials, and the protection of the environment. The challenges are still there for all mechanical engineers to tackle. We hope this volume encourages modern engineers to pick them up. We would like to thank the President and Council of The Royal Society for permission to reproduce photographs of W H Bragg and Lord Rutherford, and The Royal Astronomical Society for permission to use the photograph of A S Eddington. Thanks also, to Sandra Cumming of the RS, and Peter Hingley of the RAS for their kind assistance, and to Mike Claxton at the IMechE for all other photography. The editors are also indebted to Mr D Gillespie, of Crossley Engines, for providing biographical material. Special thanks to Judith Entwisle-Baker and her team at Mechanical Engineering Publications for all the hard work needed to turn the idea for the book into this fine volume. Desmond Winterbone, FEng UMIST January 1997

xii

Historical notes on the Proceedings of the Institution The Proceedings of the Institution of Mechanical Engineers had their genesis in the earliest of the Ordinary Meetings of members. At the inaugural session on the 27th of January 1847, rules for the conduct of business at the IMechE were approved. These included, as the principle objective: ...to enable Mechanics and, Engineers engaged in the different Manufactories, railways, and other establishments, to meet, correspond, and by mutual interchange of ideas respecting improvements in the various branches of mechanical science to increase their knowledge and give an impulse to invention likely to be useful to the community at large.

There was no explicit commitment to publish engineers' discussions. However, advertisements for IMechE events, placed in local and national newspapers, and the circulation to members of journalists' accounts of meetings, suggest that this was the intention. George Stephenson's Presidential Address, read at the first meeting, was not printed. His career reminiscences were fascinating, but contained no technical detail. In addition, the Institution's early financial difficulties ensured that his paper did not go into print. Membership subscriptions were sought to remedy the problem and Stephenson himself donated £100 to the fledgling organisation. This prompted Council to authorise the Treasurer to seek further funds in this way. By May 1847, Council was in a position to order the printing of what became the first official Proceedings. To encourage the submission of dissertations, a list of potential topics was circulated. In response, the Secretary received papers including: from Mr Chesshire, Surgeon, of Birmingham - On a system of railway carriage buffers, from Mr Johnson, of the Manchester and Leeds Railway and Henry Bessemer - Models of improved train mechanisms; and from Alfred Knight and Mr Crawford - Notes on Brakes. The choice for the first published paper was William Buckle's Series of experiments relative to the fan blast. The Institution published three further papers in 1847. The Buckle essays were sent out to members by 20 November, in all 42 copies posted at a cost of 6d. each. Proceedings were issued in parts, accompanied by large separate sheets of lithographed illustrations. John Turner of Birmingham was the printer. George Falkner of Manchester provided the artwork. This

xiii

Birmingham-Manchester axis reflected the centres of active membership in these formative years. The Institution experimented with various printers, reaching a regular arrangement in 1854, with Billing's Steam Press Offices where the Proceedings were printed until 1877. By the 1860s, over 1,000 copies of each Proceedings part was being produced for circulation. Responsibility for this workload fell to the Secretary. While the IMechE was based in Birmingham, this was William Prime Marshall (1814-1906). From 1877, in London, Walter Raleigh Browne (1842-1884) and Alfred Bache (1835-1907) suceeded to the post. Although the original manuscript versions of papers have not survived, the Institution's archives contain some information from each of the Secretaries on how technical papers reached the press. A book of "sundry memoranda" from the 1860s and 1870s provides notes on what the Secretary expected of authors. Papers best not to exceed a, dozen pages for reading, say...20 minutes. None to exceed 18 pages or 30 minutes. Any longer paper to be condensed for reading, by preparation of written abstract....

The Proceedings were not limited to papers. Reports of annual provincial meetings and descriptions of engineering works visits were added. Memoirs of deceased members appeared from 1862. The Institution's research committees, commencing in 1878, generated reports on specific topics. The increasing size and complexity of the Proceedings prompted Alfred Bache to draft a memorandum in 1889 to acquire assistants who should aid with calculations, reports, minutes, and correspondence, for Research Committees on mechanical subjects. Abstracting and indexing of papers. Preparation of manuscripts for printer. Correction of printer's proof.

In 1906, the Institution of Automobile Engineers followed IMechE from Birmingham to London. From that time, the Autos read their papers in the Mechanicals' meeting rooms in Birdcage Walk. Papers such as Frederick Lanchester's were heard here, forming a "shadow" Proceedings until the official merger of the two bodies in 1947. This merger gave rise to the first subject-based, independent Institution journal which evolved into the multi-part Proceedings of today.

xiv

IMechE Library Bookplate The library bookplate of the Institution of Mechanical Engineers was commissioned from J R G Exley in 1935. An approved design was completed by 1937. The original version incorporated stars and a biplane. These features were omitted in the now-familiar plate, shown here.

xv

Note on SUBJECTS FOR PAPERS From the first meetings of the Institution of Mechanical Engineers in 1847, catalogues of important engineering topics were compiled, both to guide members' discussions and to solicit Proceedings papers. These "Subjects for papers" were circulated as offprints, and as preliminaries to the printed journal. Growing from nine subjects at first, up to 96 sub-divided headings 30 years later, these lists show the areas which interested the engineers of the day.

xvi

InstitutionofMechanicalEngineers. SUBJECTS FOB PAPERS. STEAM ENGINE BOILERS, particulars of construction—form—heating surface—cost—consumption of fuel—evaporation of water—pres sure of steam—steam gauges, high pressure and low pressure— explosion of boilers and means of prevention—effects of heat on the metal of boilers, low pressure and high pressure—incrustation of boilers and means of prevention—evaporative power and economy of different kinds of fuel, coal, wood, charcoal, peat, patent coal, and coke—smoke consuming apparatus, best plan and results of working. STEAM, expansive force and best means of using it—power obtained by various plans—comparison of double and single cylinder engines— indicator figures from engines, with details of useful effects, consumption of fuel, &c.—contributions of indicator figures for a general book of reference to be kept in the Institution. PUMPING ENGINES, particulars of various constructions—size of cylinder, strokes per minute, and horse power—number and size of pumps and strokes per minute—application of pumps—fen draining engines. BLAST ENGINES, best kind of engine—size of cylinder, strokes per minute, and horse power—number of boilers—size of blowing cylinder and strokes per minute—means of regulating the blast— improvements in blast cylinders. MARINE ENGINES, power of engines in proportion to tonnage—different constructions of engines—comparative economy and durability of different boilers, tubular boilers, &c.—weight of machinery and boilers—kind of paddle wheels—speed obtained in British war steamers, in British merchant steamers, and in Foreign ditto, with particulars of the construction of engines and paddle wheels, &c.— screw propellers, particulars of different kinds, number of arms, material, means for unshipping, horse power applied, speed obtained, section of vessel. ROTARY ENGINES, particulars of construction and practical application— details of the results of working. LOCOMOTIVE ENGINES, express, passenger, and luggage engines—general particulars of construction, details of experiments, and results of working—speed of engines, cost, power, weight, steadiness— consumption of fuel—heating surface, length and diameter of tubes —experiments 011 size of tubes and blast pipe—comparative expense of working and repairing—best make of pistons, valve gear, &c. WATER WHEELS, particulars of construction and dimensions—form and depth of buckets—head of water, velocity, per centage of power obtained—scoop wheels for draining—turbines, construction and practical application, power obtained, comparative effect and economy. 1

2

WIND MILLS, particulars of construction—number of sails, surface and form of sails—velocity, and power obtained. FLOUR MILLS, particulars of improvements—power employed—application of steam power — results of working with an air blast — advantages of regularity of motion. SUGAR MILLS, particulars of the construction and working—results of the application of the hydraulic press. SAW MILLS, particulars of construction—mode of driving—power employed—particulars of work done—best speeds for vertical and circular saws—form of saw teeth. OIL MILLS, facts relating to the construction and working. COTTON MILLS, information respecting the construction and arrangement of the machinery—power employed, and application of power— cotton presses, mode of construction and working, power employed. ROLLING MILLS, improvements in machinery for making iron and steel —mode of applying power—steam hammers—piling of iron— plates—fancy sections STAMPING AND COINING MACHINERY, particulars of improvements, &c. PAPER MAKING AND PAPER CUTTING MACHINES, ditto ditto PRINTING MACHINES, ditto ditto CALICO PRINTING MACHINERY, ditto ditto WATER PUMPS, facts relating to the best construction, means of working, and application—best forms—velocity of piston. AIR PUMPS. ditto ditto ditto HYDRAULIC PRESSES, facts relating to the best construction, means of working, and application. FIRE ENGINES, ditto ditto ditto SLUICES, ditto ditto ditto CRANES, ditto ditto ditto LIFTS FOR RAISING TRUCKS, &c. ditto ditto ditto BLOWING FANS, particulars of the best construction, and results of experiments. LATHES, PLANING, BORING, AND SLOTTING MACHINES, &c., particulars of improvements—description of new self-acting tools. TOOTHED WHEELS, best construction and form of teeth—results of working. DRIVING BELTS AND STRAPS, best make and material, leather, rope, gutta percha, &c.—comparative durability and results of working— power communicated by certain sizes. STRENGTH OF MATERIALS—facts relating to experiments on ditto, and general details of the proof of girders, &c.—girders of cast and wrought iron, particulars of different constructions, and experiments on them—best forms and proportions of girders—best mixtures of metal. DURABILITY OF TIMBER of various kinds—best plans for seasoning timber and cordage—-results of Kyan's, Payne's, and Burnett's processes—comparative durability of timber in different situations. CORROSION OF METALS by salt and fresh water, and by the atmosphere, &c.—facts relating to corrosion, and best means of prevention. ALLOYS OF METALS—facts relating to different alloys. FRICTION OF VARIOUS BODIES—facts relating to friction under ordinary 2

3

circumstances—friction of iron, brass, copper, tin, wood, &c.—proportion of weight to rubbing surface—best forms of journals, &c.— lubrication, best materials and means of application, and results of practical trials—best plans for oil tests. IRON ROOFS, particulars of construction for different purposes—durability in various climates and situations—comparative cost, weight and durability—roofs for slips of cast-iron, wrought-iron, timber, &c., best construction, form, and material. FIRE-PROOF BUILDINGS, particulars of construction—most efficient plan —results of trials. CHIMNEY STACKS of large size, particulars, mode of building, &c. BRICKS, manufacture and durability—fire-bricks and fire-clay. GAS WORKS—best form, size, and material for retorts—construction of retort ovens—quantity and quality of gas from different coals— improvements in purifiers, condensers, and gas-holders—wet and dry gas-meters—pressure of gas, gas exhauster—gas pipes, strength and durability, and construction of joints—proportionate diameter and length of gas mains, and velocity of the passage of gas—experiments on ditto, and on the friction of gas in mains and loss of pressure. WATER WORKS—facts relating to water works—application of power, and economy of working—proportionate diameter and length of pipes—experiments on the discharge of water from pipes, and friction through pipes — strength and durability of pipes, and construction of joints. WELL SINKING and ARTESIAN WELLS, facts relating to. COFFER DAMS and PILING, facts relating to the construction. PIERS fixed and floating, and Pontoons, ditto ditto. PILE DRIVING APPARATUS, particulars of improvements—use of steam power. DREDGING MACHINES, particulars of-improvements — application of dredging machines—power required, and work done. DIVING BELLS, facts relating to the best construction. CAST IRON LIGHTHOUSES, ditto ditto MINING OPERATIONS, facts relating to mining—means of ventilating mines, use of steam jet and ventilating machinery—mode of raising materials. BLASTING, facts relating to blasting under water, and blasting generally —use of gun cotton, &c. BLAST FURNACES — consumption of fuel in different kinds—burden, make, and quality of metal—pressure of blast—horse power required—economy of working—improvements in manufacture of iron—comparative results of hot and cold blast. HEATING FURNACES, best construction—consumption of fuel, &c. SMITHS' FORGES, best construction—size and material—power of blast. COKE and CHARCOAL, particulars of the best mode of making. RAILWAYS—construction of permanent way—section of rails, and mode of manufacture—experiments on rails, deflection, deterioration, and comparative durability—material and form of sleepers, size, and distances—improvements in chairs, keys, and joint fastenings. 3

4

SWITCHES and CROSSINGS, particulars of improvements, and results of working. TURNTABLES, particulars of various constructions and improvements. SIGNALS for Stations and Trains, and self-acting signals. BREAKS for Carriages and Waggons, best construction. BUFFERS for Carriages, &c., and Station Buffers—different constructions and materials. SPRINGS for Carnages, &c., buffing and bearing springs—particulars of different constructions, and results of working. RAILWAY WHEELS, wrought iron, cast iron, and wood—particulars of different constructions, and results of working—comparative expense and durability—wrought iron and steel tires, comparative economy and results of working. RAILWAY AXLES, best description, form, material, and mode of manufacture. The Council invite communications from the Members and their friends, on the preceding subjects, and on any engineering subjects that will be useful and interesting to the Institution; also presentations of engineering drawings, models, and books for the library of the Institution. The communications should be written on foolscap paper, on one side only of each page, leaving a clear margin on the left side for binding ; and they should be written in the third person. The drawings illustrating the communications should be on so large a scale, as to be clearly visible to the meeting at the time of reading the communication, or enlarged diagrams should be sent for the illustration of any particular portions. W. P. MARSHALL, Secretary. Temple Buildings, New Street, Birmingham, 6th October, 1849.

Printed by B. Hunt and Sons. 75, High Street, Birmingham. 4

George Stephenson (1781-1848) Lauded in his own lifetime, he was described as "a humble mechanic (who) had raised himself to a position which kings or emperors might envy". Stephenson's reputation as "Father of the Railways" has tended to obscure his very real achievements as an engineer. Born at Wylam near Newcastle-upon-Tyne, Stephenson followed his father's example by becoming a colliery fireman and, later, brakesman. A talent for the repair and improvement of pumping engines combined with a desire for self improvement allowed him to prosper, and at one stage he planned to broaden his horizons by emigrating to America. Stephenson was courageous too. He would later relate how his action in saving a mine from fire earned him a horse and two additional rooms for his cottage. His early awareness of the effects of fire was behind his first wholly original invention, a miners' safety lamp, tried at Killingworth in 1815. By that time, Stephenson had also witnessed local attempts at introducing locomotive engines for coal haulage and had built his first colliery train. Many design improvements followed, leading to the machines which powered the Stockton to Darlington line (from 1822) and the Liverpool to Manchester railway (from 1830). From the time the Rocket had triumphed in the Rainhill speed trials, Stephenson's life was "a history of the railway progress of the country". He was employed as chief engineer to many railway companies, but never lost his interest in mining. When he later moved to Derbyshire he turned his hand to developing the coal stocks in that county. Throughout a distinguished career, George Stephenson steadfastly refused to be honoured. His acceptance of the Presidency of the Institution of Mechanical Engineers was a rare concession to his chosen profession in the final year of his life. His acceptance speech was directed at young engineers; one of those present would later recollect that Stephenson still spoke in the strong accent of his own youth.

5

PROCEEDINGS. The usual GENERAL MEETING of the Members was held in the Theatre of the Philosophical Institution, Cannon-street, Birmingham, on "Wednesday, the 26th of July, 1848; GEORGE STEPHENSON, Esq., President, took the chair at four p. M. The minutes of the previous meeting were read by the Secretary, Mr. Kintrea, and confirmed. The PRESIDENT said, that the first Paper to be read, according to the programme, would be his own, "On the Fallacies of the Rotary Engine;" and,—having alluded to the model of a Rotary Engine before the meeting, by a Mr. Onion, then present,—he invited the fullest discussion on his Paper, and on that model. He would show that no power was either gained or lost by the crank; and that, in effect, it was precisely the same as a simple lever. He then exhibited a diagram explanatory of his views ; and read the following Paper:—

ON THE FALLACIES OF THE ROTARY ENGINE. " As all levers give out their powers at right angles to their fulcrums, it will be seen that the right angle line I, (referring to the accompanying diagram) from the connecting rod to the centre of the beam, will be the true measure of the length of the beam, when the crank is at half stroke ; therefore,1/20of half the length of the beam will be gained by the piston end of the beam. The crank being 3 feet long, the up and down stroke of the piston will be 12 feet; the crank pin will, of course, have passed through a space of nearly 19 feet. " Now, a weight hanging upon the drum, which is nearly 4 feet in diameter, will balance the same weight on the piston end of the beam; each will move at the same velocity, and pass through the same space in the same time. " It will be observed, that from C to D is a little more than 1/3 longer than from G to D; it will, therefore, be seen, that the weight at the piston end of the beam has a little more than1/3advantage over the weight at the drum. And it will also be seen, that from C to E is half way from half stroke to the bottom centre; at this portion of the stroke, the leverage of the crank will be nearly 2 feet. The increased power that existed in the crank from half stroke to this point, will gradually be lost from E to H : it is, therefore, clearly proved, that no power is lost by the crank motion,—as the weights resolve themselves into a simple lever. 7

2 There will be a little loss of power when the engine is turning the centres, which is compensated for at the connecting rod end of the beam, by the segment of the right angle line I. " Now, a rotary engine can only give out its power on the arm, like any other lever; and if the piston passes through a space of 19 feet, it will just balance a weight equal to the same power passing through the same space. " GEOEGE STEPHENSON." Tapton House, 23rd July, 1848.

The PRESIDENT went on to observe, that the fallacy of Mr. Onion's principle was pretty conclusively proved by the fact, that fifty patents, at least, had been taken out for Rotary Engines, every one of which had failed. No man who ever lived could improve on the lever principle, as there was no power but in the lever. He would now be glad to hear the opinions of the members, and also any explanation that Mr. Onion might wish to offer. Mr. ONION then stated, that his engine had been working for some weeks at the Derby station, by permission of Mr. Kirtley, the Locomotive Superintendant of the Midland Railway; and, during that trial, experiments with his and another engine had proved that his effected a material saving in fuel. A statement to that effect, authenticated by Mr. Kirtley, was now in the possession of Mr. M'Connell, at whose suggestion he attended that meeting. The PRESIDENT said, that it appeared to him to be impossible that such could be the case. The engine might have answered at one trial, but it might fail at the next; and one trial was by no means a sufficient proof. Mr. SLATE observed, that there was one important desideratum which he desired to see obtained in the Rotary Engine,—namely, some method of packing tightly. That had never yet been found. He had paid much attention to the Rotary Engine, and had seen approaches made to an efficient system of packing, but none had been so perfect as to render the rotary principle equal to the crank. Mr. Onion had told them that his engine was more simply packed than the common engine, and he should like to have that made quite clear to the meeting. Mr. ONION said, that Mr. Scott Russell, who had written on and patented several Rotary Engines, confessed to him, that he 8

3

(Mr. Onion) had succeeded in overcoming difficulties which had hitherto been found to be insurmountable; such as making his engine steam-tight; and, also, doing away with the usual noise of the Rotary Engine. He was satisfied that his engine would bear a comparison with one upon any other principle. Mr. JOSEPH MILLER said, that one great advantage of the Rotary Engine, supposing it to be thoroughly efficient, was the small space which it occupied. If it were made as perfectly tight as the ordinary engine,—admitting, that tightness was one of the advantages of the common crank,—a useful result would be accomplished. He had never yet seen a Rotary Engine rendered sufficiently tight, but he would not go the length of saying, that it could not be done. As a practical man, however, he saw great difficulties in the way. He had never seen a Rotary Engine which remained tight for any length of time; and he should as soon expect to discover the perpetual motion as to make one which would. The PRESIDENT observed, that if he believed there was any thing in Mr. Onion's engine, he would be very happy to give his assistance in bringing it before the public ; but he really could not see any thing of value in it. Mr. MILLER thought that the question of the crank and the Rotary Engine ought now to be finally settled. It was very desirable that that should be done. On being referred to, Mr. HENRY ROBINSON said, that the Government had a Rotary Engine (Lord Dundonald's) working in the Portsmouth dock-yard for the last seven years. Mr. Onion claimed the credit of being the first who had ever succeeded in packing efficiently, but it was only the same as he (Mr. Robinson) had been in the habit of using for years; Mr. Onion had not therefore, advanced anything at all new. If Mr. Onion would call upon him in London, he would show him an engine similar to his own, and packed in the very same way. It was one that was applied to a locomotive, and commonly known as the " Jim Crow Engine,"—from its having been painted black. The difficulty with Rotary Engines had hitherto been in keeping them tight9

4

The difference between Mr. Onion's engine and the one at Portsmouth dock-yard was,, that in the latter, the packing did not depend upon springs. All that he (Mr. Robinson) was prepared to say about that engine (Lord Dundonald's) was, that it had hitherto done the work which it was intended to do. The PRESIDENT asked if any member was of opinion that there was a loss of power by the use of the crank ? they had heard e/ his reasons for asserting that their was no loss; and he wished that those who entertained a contrary opinion should declare it. A Member having spoken of Beale's Rotary Engine, the President stated, that he had been concerned in having a trial made of that engine, in a steam boat intended to carry passengers a short distance of only half a mile, to Yarmouth; but when the engine was put to work, he could not get the boat to move forward, and so the experiment failed. He managed to get the boat to sea, and it cost him and his party £40 to bring her back again. As to Lord Dundonald's engine, he was invited, on one occasion, to see it tried on the Liverpool and Manchester railway; but he refused to go, because he was convinced that a failure would be the result; and so it was,—for the engine could not be made to draw a train of empty carriages. Mr. ROBINSON :—But I think you will agree with me, sir, that there is no loss of power consequent upon the principle of the Rotary Engine. The PRESIDENT :—Not if you make it tight. Mr. ROBINSON :—The object of the Rotary Engine is to economise space and power; and if we cannot attain that end, there is something wrong in the mechanical means which are made use of. Mr. BENJAMIN GIBBONS said, that the only difficulty with the Rotary Engine was to keep it tight; but after trying many experiments to overcome that objection, he had never succeeded. The PRESIDENT then called on Mr. Buckle, who read the following Paper; having first premised, that he had chosen the subject in order to give variety to the proceedings. 10

This page intentionally left blank

John Scott Russell (1808-1882) Born at Parkhead near Glasgow, Russell followed many of his fellow Scots into shipbuilding, although by an unusual route. A good scholar, Russell graduated with an MA from Glasgow University and was made temporary Professor of Natural Philosophy on the death of John Leslie in 1832. He commenced a series of researches into wave theory which resulted in his discovery of the "wave of translation". In practical terms, this allowed Russell to devise a wave-line system for the design and construction of ships which would revolutionise naval architecture. Russell's wave researches (summarised in a succeeding paper by William Thomson in this volume, q.v.) earned him the Gold Medal of the Royal Society of Edinburgh. More importantly, vessels were built to his specifications: these were initially small iron steamers such as the Wave, but the years 1839-1840 saw the launch of larger ocean-going ships, the Flambeau and Fire-King. Eventually, Russell's ideas were used in the design of a fleet of transatlantic Royal Mail ships, plying the West Indies route. Russell's reputation is indelibly associated with two projects from the 1850s - his preliminary work with the Great Exhibition in 1851, and his involvement in the building of I K Brunei's Great Eastern. As Secretary to the Society of Arts in the years 1845-1850, Russell energetically promoted the idea of an international trade fair which was triumphantly realised at the Crystal Palace. The Leviathan, as Brunei's ship was originally called, provides a more controversial legacy. Some writers, particularly L T C Rolt, have rather one-sidedly held Russell to account for construction delays. In 1860 a Russell-designed vessel as revolutionary as Great Eastern took to the water - this was the Warrior, the first armoured, ocean-going fighting ship of the Royal Navy. Non-maritime projects also provided scope for Russell's originality of mind. The rotunda used at the Vienna Exhibition of 1873, then the largest dome ever built, was Russell's swansong and a fine rival to Joseph Paxton's better-known feat of structural engineering at Crvstal Palace.

13

PROCEEDINGS. THE usual GENERAL MEETING of the Members was held in tho Theatre of the Philosophical Institution, Cannon Street, Birmingham, on Wednesday, the 25th of October, 1848; J. E. M'CONNELL, Esq., V. P., in the Chair. The minutes of the former meeting having been read by Mr. Kintrea, the Secretary, were confirmed. The CHAIRMAN, in opening the business of the meeting, said they were met for the first time since the death of their lato lamented President; an event which, he was sure, all of them deplored most nnfeiffnedlv. He was a man who raised himself by his own talents to a distinguished position in society, and whose nanie, so long as railways existed, would endure. A member of Council, Mr. Scott Russell, had kindly undertaken the task of writing a memoir of their late President, and had intended to have been present, that evening, to read it. The Secretary had, however, received a letter from Mr. Russell stating, that illness prevented him coming; it was therefore left to him, (Mr. M'Connell) to read it.

NOTICE OF THE LIFE AND CHARACTER OF THE LATE

GEORGE STEPHENSON, FIRST

PRESIDENT

OF THE

INSTITUTION OF MECHANICAL

ENGINEERS ;

Prepared, by desire of Ike Council,

BY J. SC'OTT R U S S E L L , ONE OF THE M E M B E R S OF THE COUNCIIL.

" I wish I could address myself to the business of this evening with a feeling that the duty which you have devolved upon me were less inevitable, or more worthily performed. We have met to deplore the loss, not merely of one of the Founders of our Society, but, also, of a personal friend, whom we have long regarded with reverence and affection. Had these feelings of affection been alone regarded, perhaps our mournful silence would have formed the most expressive exponent of our grief; but the expression of our grief is the least of our duties. In our late President, England has lost one of her most distinguished men,—the world one of its great Benefactors.

15

4 " It is not as our President merely, standing as such at the head of the Mechanical Engineers of Britain, that the name of Stephenson will be known to posterity ; he will he known to posterity as the presiding Genius of our times; for of this we may be firmly assured, that the times in which we live will be known to posterity as the era in which Railways and the Locomotive Engine were first introduced as elements of social progress. It will be recorded, that about the middle of the nineteenth century Locomotives first began to run upon Railways, and that George Stephenson, the President of the Institution of Mechanical Engineers, was the Man to whose original genius chieiiy the world was indebted for the discovery. " It is difficult for us, to whom the words Railway and Locomotive are household words, to us who live, move, and have our being among Railways and their manifold social results, to go back again, even in imagination, to the beginning of the twenty years ago when were without them. So fast, indeed, we may be said to have lived through those twenty years ; so much we have been able to travel over, and see, and learn, and do, that it seems longer to go back over these twenty years, than over centuries of the slower times that went before. We, who have each of us this day come our hundreds of miles to this meeting, and may still have to return hundreds of miles to our homes this night, will find it hard to believe in the records of perils, privations, and delays, which but a few years ago made a journey from Newcastle to Birmingham one of those serious undertakings of life which were anticipated with apprehension, and recollected with congratulation. We now do more work and see more society, acquire more knowledge, by personal observation, in one day of railway life, than we were wont to do in weeks of ' the good old time.' It will be necessary, however, to task our imaginations, and go back to the times before Stephenson, in order duly to appreciate the full value of the benefits which his labors have conferred upon us. " It is not, however, alone with what George Stephenson did, that we are concerned ; still more important it is for us to consider that George Stephenson was. His title to our gratitude is no doubt great; but his claim to our admiration, as a man, is still greater. As a plain labouring workman we first find him distinguished by his untiring industry, by his zeal for the interests of his employers, and by his steadiness, sobriety, and honesty. We next find him, after having mastered all the details and drudgery of his business, continually on the watch for improvements, cultivating habits of accurate observation, and spending every leisure moment in classifying and comparing the results of his own observation, and in deducing from them hints for future improvement. Did an accident occur in his mine, his whole thoughts were immediately directed to the means of preventing its recurrence. His business, in the humble capacity of a breaksman, took him casually to the vicinity of a condensing steam engine, where the property of his master, through ignorance and mismanagement, was in danger of suffering serious damage. The young breaksman had already carefully studied the nature of its parts, and thought over the principles of its construction; the regular engineer had been baffled in his remedies, and despaired of a cure; but the youthful breaksman confided in the strength of his convictions and boldly undertook the task of refitting the machine ; the stubborn engine became at once, in his hands, obedient and useful; he had discovered for himself the secrets of the steam engine ; and at five and twenty the young coal-worker had become a Mechanical Engineer. " Thus early were the results of his self-education manifest. He had mastered the discoveries of Watt. It is true, indeed, his whole life had been one of discovery;

16

5 but as yet he had discovered no more than those who had gone before him. His had been the best of all education,—the education which a truth-loving mind, working its way among dead matter, and wrestling with the laws of nature, receives directly from nature herself,—an education far more profound and prolific than words, books, or lectures can ever impart. He had learned the laws of nature at first hand, and by experience; he knew partially what the true properties of matter were ; he felt that what they were, was exactly what they ought to have been; and however indefinitively he might be able to give reasons to others for his belief, yet one of the most valuable results of his practical self-education was to give him that implicit confidence in his own right understanding of nature, which carried him so boldly through the herculean undertakings of his future life. The whole first years of his early life were, in this way, one continued chain of discovery. Who can tell the pleasure, or weigh the profit, which such an education bestows on the simple and correct student,—compared to the formality of written dissertations, and the dryness of second-hand knowledge. " As yet, we have had said he discovered nothing new; but he was now on the eve of making a discovery, the reputation of which has enobled the name of one of our greatest chemical philosophers. A mechanic, James Watt, had already anticipated the philosophers Cavendish and Lavoisier in the analysis of water; and another was now about to anticipate Sir Humphrey Davy in the invention of the Safety Lamp. " That Stephenson was the original inventor of the Safety Lamp is now happily beyond doubt. Like most other inventions which seem to make their appearance in several places simultaneously, at the moment when the want of them has come to be deeply and generaly felt, the Safety-Lamp seems to have started into being at the same moment, nearly, in London and Newcastle. Stephenson and Davy had both discovered the principle 011 which they proposed to proceed, before either had made the lamp; but Stephenson's was made and used the first. That Stephenson first invented the lamp admits of no doubt, however much the question may remain as to how far Davy may not also be entitled to the merit of equal originality : priority to Stephenson no one can justly lay claim. " It is as a professional Engineer and a practical Mechanic that we here have chiefly entrusted to us to do justice to the memory of our distingished president. But we should do violent injustice to our own feelings if we were to pass altogether without notice his social character and private life. It is well for us all to recollect, that mere eminence as mechanics, or mechanical inventors, is not enough in the social world to make us either command the love or respect of our fellows. It is as men, chiefly, that we respect one another; it is moral character and social virtue for which we chiefly love each other. It has, indeed, been remarked, by some, on the character of our profession, that the continual struggle with tough, hard, and refractory substances, which form the business of the engineer, has the effect of communicating a hardness of character, an obstinacy of disposition, and a rigidity of temper, to men of our craft, which does not add to their excellence as members of society. It must be remembered, however, as a palliative for such faults, where they exist, that every Inventor is at first in a minority of one ; all the rest of the world is, for the time, against him; and it is often only by a long and hard fight that he at last succeeds in converting his minority into a majority.

17

6 "Invention is, therefore, a battle with the world ; and it is not easy, always, for the inventor to again consider with complacency his enemies in the field, and to adopt them as his companions in the closet. The antagonism "between the inventive man and the sceptical world is apt to extend itself to the social state. But Stephenson was, happily for himself and the world, a man endowed with no common share of the endowments which make the intercourse of life useful to himself and delightful to his friends. His energies had been sufficient to cany him through much opposition without cooling the ardour of his affections, originally warm and genial, and, above all, without chilling the enthusiasm, or closing the openness of disposition, which characterized the sanguine youth. In his latter clays he was distinguished for the childlike simplicity of his character, for the transparency of his intentions, for the singleness of his purposes, and for the straightforward manly honesty of his conversation and dealings. If he could hate an enemy, he never masked his antipathy by hypocrisy; but he was a warm and earnest friend. "Greatly, however, as Stepheiison's name will continue to "be distinguished among us as the inventor of the safety-lamp, and as a youthful mechanic of wonderful shrewdness and sagacity, it is as the first constructor and chief inventor of Locomotives and Railways that he will be known to posterity. It is in this capacity that he has conferred on society blessings which are rapidly extending to the widest limits of civilization, and which already cover Europe and one half of America. The introduction of railways is the great distinguishing event of the thirty years peace; and to them must principally be attributed the strong bonds of amity which are continually drawing nations closer and closer together; it is to railways, and the unity of international interests arising from them, that we are indebted for maintenance of that peace, unbroken for thirty years, and for the very remarkable events we are now witnessing in the existence of a casus belli in the heart of Europe, and yet of the invincible reluctance of the great powers to supply the fuel for a general war. The peace of Europe will now, we may trust, by the progress of railways and the consequent multiplication.of intercourse, be rendered as substantial as the peace of the nations of the heptarchy of England;—for we have nearly reached that period of railway intercourse, when the capitals of different nations of Europe are not separated so far from one another, either in the length of time, or in the rarity and peril of intercourse, as were the five capitals of the Anglo-Saxon kingdoms of our ancestors: Canterbury, York, and Gloucester were then more distant than are now London, Berlin, and Vienna. " How all this was early brought about, how much George Stephenson had to do with it, is now too familiar to every mind to need repetition. You all know how he early got permission from Lord Ravensworth and the proprietors of the Killingworth collieries to make an iron substitute for the horses which drew his coal waggons; how he succeeded in driving teams of waggons some six miles an. hour; but all of you who recollect these huge unweildy-looking monsters of that early time, and especially those who, like myself, then had to do with them, must remember how little we dreamed of seeing these clumsy affairs go 10 or 20, much less 50 or 60 miles an hour. Indeed, whether we look at the railway or the machine, both would have immediately been smashed to pieces had any force accelerated their speed to 10 miles an hour. It was never dreamed of, except by one dreamer, who believed in 10, 20,50, and 100 miles an hour, and who had recently determined to do it.

18

7 " The two inventions which have been combined to produce the modem Railway system may be said to be, the malleable iron rails and the locomotive engine. These were the two elements of high velocity,—each of which formed the absolute condition of the existence of the other. Without the system of laying a continuous wrought-iron rail, the notion of a velocity of 50 miles an hour could not have been entertained; and without the locomotive engine, such an expensive line could never have proved remunerative. " Most of us can remember when the idea of laying wrought-iroii bars of 50, 70, or 90 Ibs. weight per yard, for continuous miles, was an expense so utterly beyond the conception of the time, as not to be entertained for a moment; and this for an obvious reason, that no particular amount of traffic would have paid for it. I think I am warranted in saying, that no amount of traffic which horses, merely, could convey along a line of modem railway could yield a remunerative return, unless, perhaps, under peculiar circumstances, which are exceptional; I am therefore, I think, safe in saying, that the wrought-iron railroad was essentially dependent on the locomotive engine. " But that the modern locomotive engine could not subsist without the wrought-iron rail and its multifarious appendages of chains, keys, locks, sleepers, switches, crossings, sidings, and turntables, is too evident to need proof. Without the smoothness of the rail, the engine would be jolted to pieces, and without the easy motion which it gives, the engine could not be made to draw a sufficient profitable load to pay; and further, unless made of wrought-iron, it would be impossible to attain the high speed of the locomotive without iminent danger. It therefore appears that the continuous wrought-iron railway and the locomotive engine were inventions intimately related to each other, and each a condition of the other's success. To Stephenson we are indebted for the chief features of improvement in both. It was the joint perfection of the road and the engine which created the Liverpool and Manchester line, and all the progeny of that wonderful and gigantic experiment; an experiment whose complete success now bears incontrovertible testimony to the genius of the man. " There are several lessons which the life of Stephenson should enforce upon us, the members of a profession which he advanced, and of a society which he so materially assisted in founding, and in the promotion of which he took a constant and deep interest. Indeed, we cannot cast even a hasty glance back over the events of his life, without perceiving that the foundation of our Society was an act most appropriate to the termination of a career so arduous and successful. Let us endeavour to define some of those objects, and then consider how we can best accomplish them. " In the first place, then, one of the great objects of our Society is the encouragement of mechanical invention and the promotion of scientific improvement. Thus it becomes our duty to supply to this generation a great want, chiefly felt by Stephenson in his early career. The unhappy moments of his youth were those in which his inventions encountered the opposition of prejudice and interest, and when his propositions were decried because of their very originality,—because they were new, strange, unheard of, and, therefore, contrary to verified opinion. What he wanted and could not find in his youth, this Society presents to the youthful genius of this generation,—an enlightended, unprejudiced, and first ordeal, where every youthful inventor, every mechanic of original talent, every proposer of that which is new and promises to be useful, will find a body of experienced practical

19

8 men, to whom the country looks up as her wisest men, ready and willing to listen to the plans, to test the proposals, to weigh the value, and to award the praise and approbation to which the rising Stepheiisons of this generation may aspire; but which the old Stephenson could no where find, and in the want of which he was compelled to expend many years of vigour and energy in obscurity and penury. Let us see that in our hands no youthful genius, however little known, shall find his genius obscured, or his energies discouraged, eclipsed, or extinguished. If I rightly interpret the feelings of this Society, they would hail with welcome any discovery, and cooperate heartily and disinterestedly in giving to the world its benefit, and to genius its honors and rewards. " Another circumstance must have greatly impaired the means of usefulness of Stephenson in his early life, and one that he most deeply felt,—viz., the want of knowing that which other men were doing, and had done before him, in subjects allied to those in which he was occupying his mind. Thus much we know with certainty, that no man was more happy to communicate, in after life, to others the abundant stores of practical knowledge he had accumulated, and that no one felt a more kindly interest in the inventions and plans of younger men, or was more disposed to promote their interest and forward their views. Let us regard it as a part of his legacy so to impart, liberally, to all younger members of the profession, what more knowledge or greater experience may have enabled us to acquire. After all, there is no tribute more gratifying to the members of our profession than the due appreciation by each other of that which each of us may have done to advance the interests, and increase the resources of mechanical science. "It would not be fair to the character of our late President to omit from our recollection the very large and original views which he entertained on general science. It has been too common in our profession to place science and practice in opposition to one another; as if true science and hard practice could possibly be opposed. If science mean that which is carefully ascertained, and accurately defined, and truly demonstrated,—then it is impossible that any sound practice can possibly stand in opposition to, or independent of, it. If practice mean the knowledge which is founded on the actual facts and experience of intelligent men, it is impossible to see how the largest amount of that knowledge possessed by any one man can differ from the extensive and generalised facts in which science embodies the experience of all mankind. Stephenson is a remarkable freedom from this prejudice. He was eminently a practical man. He wrought early, and much, with his own hands. He had wrestled with matter, and knew all its qualities by feeling it and pushing it and pulling it, by cutting and filing and chipping it. He had hammered it hot and hammered it cold, he had melted it and moulded it, planed it and sawn it, broken it across, pulled it asunder, and twisted it round. He knew its action and its reaction, its inertia and its momentum, its vis mortua and its vis una. His was a supremely practical and personal acquaintance with the laws and property and phenomena natural to matter, whether solid or liquid, fluid or gaseous, mineral or aerial, more than any man who has ever risen to eminence. Stephenson was entitled to rank as a consummately practical man. But was he not equally, or more, a scientific than a practical Engineer ? Was there ever a bolder theorist than he was ? Were there ever more daring scientific speculations than those wild flights in which his genius delighted to break forth ? In chemistry' in vegetable physiology, in vital mechanism, in electricity, in galvanism, in the theories of the gases, on the inert constitution of matter, and of heat, and even

20

9 on the mechanism of the mind itself, he had deeply thought, profoundly read, and boldly and fearlessly speculated. Every step in his life was the realisation of what had before been a theory. It is true he was not educated early in the rudiments of science, at school or at college; but what of that ? what is life but a great school ? Is not the press our school, and necessity our school of invention ? Stephenson read and studied science ;—he was not ignorant, but he was self-taught. Before he became a great man he had studied profoundly, and he does not appear to have ventured on any construction or invention, before having accurately, and generally truly, calculated by the principles of science, its probable and actual results. In all his works, Stephenson exhibits to us embodiments, eminently practical, of the profoundest principles of mechanical science. Let the men among us who desire to emulate him most, endeavour to combine, in the greatest degree, the truest science with the soundest practical sense. These are not times in which any of us can afford to dispense with any science, or any practice, that it may be in his power to obtain. " I will now venture upon an illustration of the advantage of uniting high science with extensive practice, which has often occurred to me as an excellent illustration of Mr. Stephenson's scientific knowledge, and also as an illustration of the advantage he would have derived, as a practical man, from having been still more profoundly scientific than he was. Stephenson, we know, invented the fishbellied rail, and a great invention it was thought in its day. The Liverpool and Manchester Railway was opened with it. It was an invention to give, with a small addition of metal to the under middle side of the rail, nearly double the strength, and this it successfully accomplished. But here he stopped short: he had not science enough to see, that by making the wrought-iron bar in long lengths, stretching over a number of blocks, or sleepers, he had brought it into a new condition, to which a much higher rule was applicable ;—he neglected the difference between a rail having a joint at every chair, and one having only a joint at every fifth or sixth chair; had he perceived that, he would have invented the parellel rail, and would have learned that the joint chairs require to be nearer together than those removed from the joint by a fixed proportion. The fish-bellied rail was a failure. It was the result of science ; but of science of which there was not enough. It was also the result of practice ; but of practice under different conditions. It was reserved for Mr. Buck, a profoundly scientific pupil of Stephenson.'s, to develope the true science of the wrought-iron rail. Where not a little science had failed, a little more made the invention perfect. Let us learn from this to be always trying to obtain a little more science, as well as a little more practice, than we have got,—remembering that Stephenson continued his education of himself to his dying day. " The best testimony, however, which Mr. Stephenson has borne to the value of scientific education for a practical man, is to be found in the course he adopted for the training of his son to our profession. The assiduity with which he laboured at clock-making, the cleaning of watches, or any other industry, in the intervals of his regular business, in order that he might be able to afford to him those blessings of education of which he himself so deeply felt the want, is one of the most charming features in his character. His most earnest desire, in early difficulty, was to give Robert all those precious thoughts and truths which he himself only acquired late and too labouriously. And how admirably his plan succeeded, his son's unclouded successes, both as a Mechanical and a Civil Engineer, are the B

21

10 evidence to us, as indeed they were the subject of just pride to himself, who never spoke of his son without strong emotions of joy and pride. There are none of us who will question either the justice of his pride or the soundness of his plan of education. " It is one of the peculiarities of genius to inspire those within its influence with some of its own fire. This was peculiarly the case with Stepheiison. Nearly all the present ornaments of our profession have been his pupils. He was the founder of a school of eminent engineers, who in England, Europe, and India, are now extending, amongst all portions of the human race, the blessings of those great bonds of civilization and social intercourse which he first fabricated. It is to his labours, and theirs, that this country owes the addition of £200,000,000 to its productive wealth, the opening up of a host of new branches of industry, the quickening and invigorating influence of rapid and cheap intercourse ; and to him that the poor everywhere owe the blessings of cheapened coal, and the facilities of social enjoyment and healthful recreation. " In this brief notice of the chief features and character of our late president, which I have thus imperfectly, although most earnestly, sketched amid the bustle of business, I have dwelt mainly upon such features and characteristics as were peculiarly interesting or instructive to us, as members of an institution founded, in a great measure by himself, for professional purposes. I have regarded, therefore, chiefly his professional character; but I cannot conclude without expressing an earnest wish, that his life as a man, exhibiting the beauty and excellence of his character in all its cheering aspects, as a boy, as a workman, as an engine-man, as a viewer, as an engine builder, as an improver of mineral railways, as the engineer of the Liverpool and Manchester Railway, should be written by some one who has leisure to collect from his many friends all their recollections of him, while they remain fresh and accessible. I should desire also to see a detailed account given of his progress, his difficulties and his means of success in any one of his labors. This would be a most valuable and instructive work; and I do not know on whom it should devolve more properly to see such a work executed faithfully and judiciously than on this society, whom he made the favoured recipients of his knowledge and experience, and who ought to consider themselves as his literary and scientific executors, to whom the world may naturally look to see justice done to the memory of one of England's greatest men, the founder of our railway system and of the Institution of Mechanical Engineers."

The memoir, which, both at the close and during the time of reading, elicited expressions of admiration, having been read, the CHAIRMAN said, he presumed it would be unnecessary to put it to the vote, that the members return their best thanks to Mr. Scott Russell, for his very able memoir. The vote was carried with acclammation. Mr. GEACH rose and said, it was with melancholy satisfaction he begged to move, that they place on the minutes of their proceedings an expression of the regret they all felt at the loss of so excellent a man as their late friend, Mr. George Stephenson,— apart from his having been the President of this Institution. He

22

11 well recollected Mr. Stephenson, on the last occasion of their meeting, filling the place which Mr. M'Connell now so worthily occupied, in high spirits and in good health. The recollection of the circumstance cast a gloom over his feelings, and he was sure it would have the same effect on every member present. He had known Mr. Stephenson a shorter time, perhaps, than many of them; but he had known him well enough not only to entertain respect for him, but affection also. There was something so endearing about his manners, so open and kind, and so encouraging to all those less experienced than himself,—there was so much of kindheartedness about him, that no one could help entertaining for him a high respect. He would quite allow that his manners were sometimes rough,—he would quite allow that there were peculiarities in his character, which had to be considered as peculiarities; but he was quite sure those who knew him best considered that these very peculiarities gave him a greater claim on their regard. He was willing to allow that he had seen in Mr. Stephenson what in other men might subject them to criticism; but when it came from Mr. Stephenson, it came from a privileged person. Mr. Stephenson was proud of his own early life, and he never lost any opportunity of expressing it,— he never attempted to conceal that he came from the lowest grade of society, and had raised himself to his high station; and he ever evinced the same pleasure in meeting an associate of his early life, in humble circumstances, as he did to meet the peers of the realm, with many of whom he associated in later life. He had the same gratification on meeting one whom he had known in early life, or the son, or connection, of such a one, and in referring back to the time when they had struggled together through difficulties, as he did in referring to the occasion when he was taken by the hand by the highest in the land. Although oppressed with these recollections, he could not content himself without making the few remarks he had; and he now begged to move,— " That the members of this Institution desire to express their deep regret at the decease of their late President, George Stephenson, whose early support of the Institution so greatly contributed to bring it to its present state of prosperity and success." Mr. FOTHERGILL, in seconding the resolution, expressed his great sorrow that such a duty should devolve on him. After the 23

12

observations of Mr. Geach he should content himself with merely seconding the resolution, for he was sure that every one participated in the same feelings of deep regret. The resolution was put and carried unanimously. The CHAIRMAN then rose and said, that immediately on hearing of the death of their late President, the Council met, at Manchester, and, after forwarding a letter of condolence to his widow, for the irreparable loss she had sustained, they resolved, that the best tribute they could pay to the late Mr. Stephenson's memory, and the best way in which they could testify their appreciation of his merits, besides at the same time the best selection of a future president they could make from among the eminent men of the dav, would be to invite Mr. Robert Stephenson to become his father's successor, as president of this institution. The Council did so feeling assured that the members would entertain the same opinion. Accordingly, two of the Council were appointed a deputation to wait on Mr. Stephenson. Owing to an accident they were prevented from seeing Mr. Stephenson, but a most satisfactory and pleasing correspondence ensued, which would be read by one of the deputation. Afterwards it would be his duty to nominate Mr. Robert Stephenson, as the future president. Mr. FOTHERGILL said, that Mr. Buckle and himself were the deputation appointed to wait on Mr. Stephenson; but for the reason stated by the Chairman, they had not been able to see him. They had, however, a correspondence, and Mr. Stephenson's reply did equal credit to his character as a man, and to his feelings as a dutiful son. He had not the letter with him; but the substance of it was, that, if elected, nothing would be wanting on his part to discharge the duties of the office in a manner satisfactory to the members, and that he would endeavour to watch over the interests of the Institution as earnestly as his lamented father had done. The CHAIRMAN then begged to propose Robert Stephenson, Esquire, as President of the Institution. He felt certain that every member would agree with him, that a better choice could not be made. Mr. Robert Stephenson was a worthy son of a worthy father; and the Institution would gain additional lustre by having that gentleman as its President. The resolution was

24

13

seconded by Mr. Fbthergill, and carried unanimously, amid every demonstration of satisfaction.

25

This page intentionally left blank

Richard Peacock (1820-1889) From Swaledale, West Yorkshire, Peacock was a true child of the railway age - it is said that with his father, he witnessed the opening of Stephenson's Stockton to Darlington line. Thereafter, Peacock served his apprenticeship with Fenton, Murray, and Jackson of Leeds, who were then engaged in locomotive building for the Liverpool to Manchester Railway. Under the management of Peter Rothwell Jackson, the firm also manufactured engines, pumps, and hydraulic machinery of various types. At the age of eighteen, he became locomotive superintendent of the LeedsSelby Railway, beginning a long and distinguished career in the railway industry. Following some experience at I K Brunei's Great Western Works, Peacock took up the position of superintendent on the Manchester and Sheffield Railway, expanding and developing the new concern in the years 1841-1854. Like his mentor P R Jackson, Peacock became one of the founding members of the Institution of Mechanical Engineers during this period. His earliest contribution to the Proceedings describes his other great endeavour, the design and laying out of the Gorton locomotive workshops. Peacock's association with Gorton in Greater Manchester shaped the remainder of his career. He partnered Charles Beyer, another IMechE founder, in creating the Beyer-Peacock company in 1854, so founding one of the great locomotive-building firms. Peacock's extrovert nature found him involved with local affairs: he was in turn a banker, magistrate, and Member of Parliament. Peacock served as President of various organisations, notably the Manchester Steam Users' Association, the Royal Manchester Institution, and the Locomotive Manufacturers' Association. His belief in the engineering future of his region prompted him to serve as an original director of the Manchester Ship Canal Company - a project he did not live to see completed.

27

ON THE WOKKSHOPS FOR THE LOCOMOTIVE CARRIAGE AND WAGGON DEPARTMENTS OF THE MANCHESTER, SHEFFIELD, AND LINCOLNSHIRE RAILWAY. These works are erected at the first point from Manchester where the Railway and the land take the same level, viz.: at Gorton, about two miles from Manchester, this being considered the best position from its being near the principal terminus of the Company's Lines, and from the facility with which materials can be procured and workmen engaged ; and though it is a terminal establishment, with the advantage of being situated near enough to a First-class Mechanical Town to secure any benefit that may be had therefrom, it is sufficiently far out of it to be clear of the heavy local taxes with which such establishments in all large Towns are burdened. The site was fixed upon, and land purchased to construct Workshops for the Sheffield and Manchester Railway only; but subsequent to the amalgamation of that Company with the net-work of Lincolnshire lines, and which now form the Manchester, Sheffield, and Lincolnshire Railways, more land was purchased and the Workshops increased in size to meet the wants of the joint Companies. The total quantity of land purchased is nearly Twenty Acres, about Nine of which is occupied by the Workshops and Storeyard, and the remainder is being used for the construction of Reservoirs for supplying the works with water, and for erecting Cottages upon for the workpeople in the Company's service. The block plan, Plate 21, shows the general arrangement and relative positions of the Shops, Cottages, Reservoirs, &c., the plot of land being bounded on the south by the Railway, on the east by the Peak Forest and Macclesfield Canal (also belonging to this

29

LOCOMOTIVE WORKSHOPS.

28

Company), and on the north adjacent to the Manchester and Ashtonunder-Lyne highroad. The Reservoirs are calculated to hold a month's consumption of water, and are supplied from the adjoining Canal, the water passing through filter beds in its course from the Canal to the Reservoirs. These Reservoirs from their elevated position supply the water directly into the Tenders upon the Railway and throughout the Workshops, their position being sufficiently high to do this, and the Canal high enough to supply the Reservoirs. The Cottages shown on the plan are 140 in number, arranged in four blocks ; and between the Cottages and Reservoirs, and the Workshops, is a plot of vacant land that may be used for increasing the number of Cottages, or for any other purpose that may be hereafter required. The plan of the works is nearly that of a square, the Watchhouse or entrance being situated towards the Cottages on the east side of the works, as also is the rail entrance, and adjoining are the Offices and General Stores. The Engine-house or shed for working Engines, Plate 22, is a rotunda of 150 feet in diameter inside, and will hold seventeen Engines with their Tenders, leaving the entrance and exit lines clear. The advantage in this description of building over the ordinary polygon is in the absence of pillars for supporting the roof, there being but one in the rotunda, while in the polygon, say of twelve sides, there would be twelve, and the number of pillars would determine the number of lines and consequently the number of Engines it will hold, while in the rotunda the number of lines is with the number of Engines, influenced only by the clearance required for each other; thus, the polygon would hold eleven engines with the entrance clear, while the rotunda will hold seventeen. To the left of the entrance is a furnace for lighting up the Engines from, and the points for the two lines to the table are set so that the Engines will (on entering) go upon the right hand side of the pillar ; and thus, supposing them to enter Engine first, they must be backed into each line, which will cause the smoke box, or chimney end of the Engines, to be always nearest the table, and consequently in a right position for the tubes &c. being cleaned. The turntable in the centre is 40 feet diameter, with two lines of rails upon it, one upon each side the Centre Pillar around which it moves The centre pillar is of cast iron, the base forming the bed

30

24

LOCOMOTIVE WORKSHOPS.

for the inner rollers of the turntable to revolve upon ; the top of the pillar is sufficiently large to receive the shoes for carrying the principals of the roof, and to which they are secured by bolts, each principal radiating from the centre of the pillar, and its opposite end resting upon the outer wall of the building. A collar is cast upon the pillar about 8 feet from the top, which was intended to carry one end of a circular travelling frame; the frame being intended to revolve round the pillar, and the opposite end having a carriage running upon a circular rail beam, which was to have been supported by the pilasters built on the inside of the walls, the frame being surmounted by a travelling crane in the usual way; this however has not yet been carried out. The roof is of wrought iron, surmounted by a louvre, the top of which is glazed; the whole forming a beautifully ventilated and well-lighted building. To the left of the rotunda are the workshops, with Engine-house, boiler, &c. The Fitting and Tool Shop is 120 feet by 60 feet, and contains the whole of the Tools, with the exception of the Punching and Shearing Machines. Two rows of Fitter's Benches are erected near the far end; the Lathes, Drills, &c., are placed down each side, and have their counter shafts carried by wall plates, built into the side walls, and the Planing Machines are placed in the centre, the whole being driven from two lines of Main Shafting passing longitudinally down the shop, one over the vertical shaft from the Engine, and the other equidistant from the opposite wall, this shaft being continued over the Shop Stores and passing over the Travelling Platform into the Carriage Shed for- driving the Hoist therein. The Smith's Shop is next to the Fitting Shop, and is of the same dimensions, 120 feet by 60 feet; it contains a Fan and sixteen Smiths' Fires, eight of which are placed upon each side of the Shop, and if necessary three more can be placed at the ends. Next to this is the Boiler Shop, the same size as the Smithy, in which are erected eight smiths' fires, on the side next to the Smiths' Shop; four boiler fires are placed upon the opposite side, and the punching and shearing machines at the entrance end, these and the fan being driven by a shaft passing from the Engine transversly across the ends of the shops. Adjoining to the left and at right angles with these is the Erecting Shop, which is 150 feet by 60 feet, in this are nine transverse lines of rails, each line holding two engines, down each side and the centre are pillars supporting longitudinal beams for carrying the Travelling Cranes one

31

LOCOMOTIVE WORKSHOPS.

25

upon each side ; both these Cranes traverse the full length of the shop, and are each calculated to lift an Engine and move it to any part of the shop, if necessary. To the left, and bounding the west side of the works, are the Waggon and Carriage Shops, the Waggons being on the ground-floor and the Carriages above; the carriages are lifted up and down by a self-acting worm-hoist, worked by the shop Engine. These rooms are 320 feet by 70 feet, the Carriage shop will hold thirtyeight carriages, and the ground-floor about fifty waggons ; at the end of these are the lifting-room below, and the Trimming and Saddlery room above, each 60 feet by 70 feet. The lines in the lifting and waggon shops are served in common with the erecting shop by a travelling platform, SO feet by 12 feet, running upon three rails at right angles with the lines in the shops. Opposite the lifting-shop, and formingpartof the south boundaiy, is a paint shop, 60 feet by 40 feet, and in continuation of this is a shed for working stock not required for present use; this shed is 165 feet by 40 feet, and may be used for working engines if necessary. In a line with this and at the south east corner of the works is the Coke Shed, 100 feet by 40 feet; this is so constructed that the Coke Waggons are on one side and the Engine on the other, the Coke being filled into baskets upon a platform between the Engines and Waggons, and transferred from thence to the Engines, the waggon line side of the shed is closed, as also the ends, but the Engine line inside of the shed is open, the roof merely projecting over the Engines, where they are being coked. The arrangement of the lines into and in the works, with the four sidings running parallel with the railway outside, is shewn by the block plan, Plate 21.

Mr. PEACOCK observed, they had never experienced the slightest difficulty with the turn-table in the rotunda, or the two lines, during the two years that they had been at work. There was no danger arising from a want of balance on the turn-table when only one line was loaded with an engine, because each line of rails was carried by an independent pair, of girders, supported by rollers, and joined together in the centre. They could turn an engine upon the table in about a minute, with three men, D 32

26

LOCOMOTIVE WORKSHOPS.

and it was sometimes done by two cleaners. The object in the arrangement had been to get as many engines in as small a space as possible, and they could find no other shape so well adapted for the purpose, or into which so many engines could be got in proportion to the area, with the same convenience and room for getting about them. The total area of the floor of the building is a little over 17,000 square feet, which is equal to 1,000 square feet of shed surface per Engine accommodated, with ample room to get conveniently around each, and leaving the entrance and lines clear. Mr. DOCKRAY did not see what was gained by the oblique arrangement. At Camden-town the rotunda was 160 feet in diameter, and held 24 engines on the old arrangement, the space allowed for an engine and tender being 50 feet in the centre to turn. If the columns in that arrangement were placed sufficiently far back to get a clearance between the engines, he considered they would not lose any space. Mr. PEACOCK observed, that although the columns might be put so far back as to clear the lines, columns were always very objectionable and inconvenient at the side of the engines, and he thought the central column much preferable. Mr. COWPER suggested, that with a roof of only 150 feet span, the columns might be entirely done away with, and the cost not be increased more than £1 per square. Mr. PEACOCK observed, with respect to lifting the carriages into the upper shop, that it was effected in two minutes by the worm hoist; the time was not an object of importance, as there were only about two carriages raised per day. Mr. GIBBONS suggested, that the air-lift might be very advantageously employed for the purpose. By a. very small abstraction of power continuously going on they would procure a reservoir of power of large amount ready to be applied when requisite, and be enabled to lift the carriages in a quarter of a minute. There was one advantage in the employment of compressed air, that it was more under command than any other power, and more easily regulated. Mr. PEACOCK did not think that plan could be applied economically in the present case, as it would involve the expense 33

IMPROVED VACUUM GAUGE.

27

of a large reservoir; the object had been to get something to answer the purpose as simple and cheap as possible, and the only apparatus employed was a 1-foot worm working into a 4-feet worm-wheel. The CHAIRMAN proposed a vote of thanks to Mr. Peacock, which was passed ; and the following paper, by Mr, F. Bramwell, of London, was then read :—

34

LOCOMOTIVE

WORKS.

Plate

21

LOCOMOTIVE WOKKS Plate22. Manchester, Sheffield £ Lincolnshire Railway

Fig. 1.

Fig. 2. Plan,

Section of Engine Shed

William Menelaus (1818-1882) Although forever associated with Welsh engineering, in particular with the iron and steel industry of the Dowlais Iron Works of Merthr Tydfil, Menelaus originally hailed from Edinburgh. He worked as an engineer and millwright in Scotland and in London before undertaking a successful consulting exercise on a corn-mill belonging to Rowland Fothergill of Cowbridge. This led to a permanent appointment at the iron works of Messrs Fothergill and Scales at Abernant. By 1851, Menelaus was engineer at Dowlais, taking over management of the whole plant in 1856. The works were already famous, having produced rails for the Stockton to Darlington Railway, and by this date occupied a large complex, employing over 7,000 people. Menelaus was the first British engineer to take up the challenge of large-scale manufacture of Bessemer steel, shortly after Henry Bessemer's preliminary description of the (as yet unperfected) process. Early results were unpromising, but Menelaus had sufficient belief in the idea to ensure its eventual success. By 1870, Dowlais had six Bessemer converters each with a capacity of five tons. On other occasions, the engineer's venturesome qualities were far ahead of his contemporaries. His trial of steel girders was rebuffed in the marketplace. The reason given was that "English architects would not adopt girders...in their buildings". Menelaus turned the Dowlais Works from a loss-making operation, £52,907 on the debit side in 1856, to a hugely successful business, earning £229,836 profit in 1872. His contributions to the understanding of his profession were of like proportions. Menelaus founded the South Wales Institute of Engineers at Merthr in 1857, and was similarly instrumental in the establishment in 1869 of the Iron and Steel Institute, becoming its President in 1875-6.

37

112

DOWLAIS IRONWORKS ENGINES.

DESCRIPTION OF THE LARGE BLOWING ENGINE AND NEW ROLLING MILL AT DOWLAIS IRON WORKS. The large Blowing Engine and Rolling Mill forming the subjects of the present paper are remarkable particularly for their great size, the blowing engine being the largest of its class hitherto erected either in this country or abroad; and were designed with a view to turning out a large quantity of work with the greatest possible security from risk of failure or deficiency of the blast or of breakage of the machinery. The Blowing Engine was erected in 1851, and is shown in Figs. 1, 2, and 3, Plates 119 and 120. Fig. 1 is a side elevation of the engine, and Fig. 2 an end elevation. Fig. 3 is an enlarged vertical section of the blowing cylinder. The blowing cylinder A is 144 inches diameter with a stroke of 12 feet, making 20 double strokes per minute, the pressure of the blast being31/4Ibs. per square inch. The discharge pipe B is 5 feet diameter and about 140 yards long, thus answering the purpose of a regulator. The area of the entrance air valves is 56 square feet, and of the delivery air valves 16 square feet. The quantity of air discharged at the above pressure is about 44,000 cubic feet per minute. The steam cylinder C is 55 inches diameter and has a stroke of 13 feet, with a steam pressure of 60 Ibs. per square inch, and working up to 650 horse power. The steam is cut off when the piston has made about one-third of its stroke, by means of a common gridiron valve D near the back of the slide valve E, as shown enlarged in Figs. 4 and 5, Plate 120 ; there is also on one side of the nozzle a small separate slide valve F for moving the engine by hand when starting. The cylinder ports are 24 inches wide by 5 inches long, and the slide valve E has a stroke of 11 inches with 1/2 inch lap. The engine is non-condensing, and the steam is discharged into a cylindrical heating tank, 7 feet diameter and 36 feet long, containing the feed water from which the boilers are supplied. Under the steam cylinder C there are about 75 tons of cast iron framing G, and 10,000 cubic feet of limestone walling in large blocks, some of them weighing several tons each.

39

DOWLAIS IRONWORKS ENGINES.

113

The beam H is cast in two parts of about161/2tons each, the total weight upon the beam gudgeons being 44 tons; it is 40 feet 1 inch long from outside centre to outside centre, and is connected to the crank on the flywheel shaft I by an oaken connecting rod K strengthened from end to end by wrought iron straps. The beam is supported by a wall L across the house, 7 feet thick, built of dressed limestone blocks, to which the pedestals M are fastened down by twelve screw bolts of 3 inches diameter. The flywheel I is 22 feet diameter, and weighs about 35 tons. Eight Cornish boilers are employed to supply the steam, each 42 feet long and 7 feet diameter, made of 9/16 inch best Staffordshire plates, and having from end to end a single 4 feet tube in which is the firegrate 9 feet long. For some time this engine supplied blast to 8 furnaces of large size, varying from 16 to 18 feet across the boshes; it is now blowing, with three other engines of small dimensions, 12 furnaces, some of which make upwards of 235 tons of good forge pig per week, the weekly make of the 12 furnaces being about 2000 tons of forge pig iron. With the exception of the cylinders, made and fitted at the Penan Foundry, Truro, this engine and boilers were made at the Dowlais Iron Works, and erected according to the design and under the superintendence of Mr. Samuel Truran, the Company's engineer. The engines for driving the new Boiling Mill now in course of erection at the same works are a pair of high pressure engines coupled at right angles, shown in Figs. 6 and 7, Plates 121 and 122. Fig. 6 is a side elevation, and Fig. 7 an end elevation of the engines. Fig. 11, Plate 123, is a general plan of the rolling mill to a smaller scale. The steam cylinder C is 45 inches diameter with a stroke of 10 feet, making 24 double strokes per minute. Each cylinder has a common slide valve of brass worked by an eccentric on the main shaft. The expansion valves are of the gridiron sort, worked by a cam on the main shaft, the steam being cut off at about one-third of the stroke; an arrangement is made for throwing these valves out of gear when the engines are doing heavy work. Each engine is furnished with a small slide valve to be worked by hand for the purpose of starting and reversing. The steam is supplied by six Cornish boilers, 44 feet long and 7 feet diameter, having a 4 feet tube in each; the whole of the plates are best Staffordshire 9/16 inch thick, and the total weight is 120 tons.

40

114

DOWLAIS IRONWORKS ENGINES,

The framing G under the engines and machinery is of cast iron, and consists of four lines, each 75 feet long, 12 feet high, and 21 inches wide; the whole weighing about 850 tons. Each beam H is in two parts, the sides weighing about 17 tons, making the total weight of each beam when complete about 37 tons. The two beams are supported upon eight columns L, 24 feet long and 21/2 feet diameter, securely fastened at the bottom in deep jaws cast upon the framing. Upon the top of each group of four columns is a large and heavy entablature plate N, which carries the plummer blocks M under the main gudgeons. Each column passes through the entablature, the bosses at the junction being 24 inches deep ; these are bored and the tops of the columns turned so as to ensure a perfect fit. The plummer blocks M are secured by wrought iron keys in jaws cast on the entablature N in the usual manner. The connecting rods K are of oak with wrought iron straps. The driving wheel shaft I is of cast iron with bearings 24 inches diameter; the flywheel shaft 0 is also of cast iron with bearings 21 inches diameter. The diameter of the driving wheel is 25 feet to the pitch line, width on the face 27 inches, and pitch 7 inches. The diameter of the spur wheel or pinion on the flywheel shaft is 6 feet, and the teeth are strengthened by a flange running up to their points on each side. The flywheel 0 on the mill shaft is 21 feet diameter and weighs about 30 tons, making upwards of 100 revolutions per minute. The whole of the fastenings both of the wheels and framing are of dry oak and iron wedges, as shown enlarged in Figs. 8, 9, and 10, Plate 121. These engines will drive one rail mill capable of turning out 1000 tons of rails per week, another mill capable of making 700 tons of rails or rougheddown per week, and one bar or roughing-down mill capable of making 200 tons per week; they will thus readily turn out 2000 tons of iron per week. Two blooming mills with three-high rolls and two hammers will also be worked by the same engines. The saws and small machinery will be driven by separate engines, as will also the punching and straightening machines. The roofs cover a space of 240 feet by 210 feet, and are to be covered with corrugated black plates of No. 14 wire gauge thickness. The span is 50 feet, the roofs being supported upon lattice girders of an average length of 45 feet. The position of the columns is shown on the ground plan, Fig. 11, Plate 123; and it will be observed that the entire mill floor is free from obstruction. The flooring will be of cast iron plates 1 inch thick.

41

DOWLAIS IRONWORKS ENGINES.

115

It has long been felt that the power of rolling wrought iron of large section and great lengths has not kept pace with the requirements of engineers, who are hampered in their designs by the impossibility of obtaining iron of sufficient dimensions. For engineering works of any magnitude, bars of great length, considerable width, and moderate thickness, are frequently required. In the ordinary mode of rolling, the length and width of the bar are measured by the power of the engine and the time occupied in rolling. It is obvious that to finish a bar quickly it is necessary that it should be rolled in two directions to prevent delay; and long and heavy bars can be thus rolled only by an engine of enormous power. This object is designed to be attained by the large combined engines now described. A simple arrangement of rolls for working in two directions is shown in Figs, 12, 13, and 14, Plate 123 ; the lower pair of rolls P is driven from the flywheel shaft, and under ordinary circumstances will be worked in the usual manner, rolling the bars in one direction and lifting them over the top roll in coming back. When it is necessary to make extra sized bars the rolls RR will be put in the standards, and driven from the flywheel shaft by the pair of wheels SS, Fig. 6, Plate 121, thus giving the means of working the iron in both directions, as shown by the arrows. By this arrangement the mill is expected to be able to roll iron of such sections and lengths as have been hitherto unattainable.

Mr. E. JONES said he had had an opportunity of seeing the large blast engine at work at the Dowlais Iron Works, and was much struck with it. It was a very remarkable and interesting sight to see 40 tons of iron in the main beam moving at the high speed of 520 feet per minute at the extremities, a much greater speed than had been contemplated when the engine was first erected. The whole working of the engine was so smooth and quiet, that its great magnitude was not apparent at the first glance. Mr. McCoNNELL thought the engine was one of great interest, and the excellent drawings that were exhibited of it would form a valuable addition to the collection in the Institution; he hoped that some indicator diagrams from the engine would be also supplied, as they would be of particular value from an engine so remarkable for size and speed.

42

116

DOWLAIS IRONWORKS ENGINES.

Mr. MENELAUS said he would be happy to supply diagrams of the working of the engine ; it had now continued regularly working six years without any failure, and had proved very satisfactory. The CHAIRMAN observed thai the engine appeared not to be a condensing one, and enquired the reason for such an engine being made non-condensing. Mr. MENELAUS replied that a high pressure non-condensing engine had been adopted, because of the difficulty in getting a sufficient supply of water for condensing in that situation; it was worked with 60 Ibs. steam, expanding through about 2-3rds of the stroke. The CHAIRMAN asked whether the waste gases from the blast furnaces were used for heating the steam boilers. Mr. MENELAUS replied that they did not make use of the waste gases for that purpose, not having been able as yet to get a completely satisfactory result from their use; raw coal was used in most of their furnaces with cold blast, and was not so suitable as in other cases for taking off the waste gases from the furnace; the make of the furnaces was very large, as much as 235 tons per week being obtained from one furnace working with raw coal and cold blast. Mr. HENRY MAUDSLAY observed that an oak connecting rod was used in the engine, and enquired the reason for its adoption, as he was surprised to see that construction in an ironworks, all the rest being so massive. Mr. MENELAUS said the result of 40 years' experience had led them to prefer oak with wrought iron straps, having found it the best on the whole for the purpose. A cast iron connecting rod was liable to break, and it was of particular importance to guard against risk of breakage, as the consequences of such an accident would be very serious; and wrought iron would be very expensive when of that extreme size. They found the oak connecting rod very durable, and easy to repair, whilst it was a much less expensive construction and considerably lighter than iron. Mr. COCHRANE enquired what was the total weight of metal in the two engines described, as they appeared to be very massive in their proportions and considerably stronger than they would be by the usual calculations. Mr. MENELAUS replied that the blowing engine of 650 nominal horse power weighed about 300 tons, including the bed-plates; and the pair of rolling mill engines of 1000 nominal horse power would have about

43

DOWLAIS IRONWORKS ENGINES.

117

1000 tons of metal in them. These were probably nearly double the ordinary proportions; but it was their practice to make their engines very strong, and they generally put in three times the calculated weight of metal, as it was of the first importance to them to have no stoppages of the works from any break-down of the engines. They expected their engines when once fairly started to go on working week after week without stoppage or delay, and expense of construction was not a consideration compared with this object; if they could arrive at a machine that would never break, it would be considered perfection. The CHAIRMAN asked what was the object of having the set of four rolls that was shown in the new rolling mill. Mr. MENELAUS said that the limit as to length in making very large bars of thin section was the time required for the rolling, as they had to be completed at one heat, in consequence of the impossibility of getting such long bars into the furnace to reheat; in the ordinary rolling so much time was lost in bringing back the bar over the top of the rolls that it was found impossible to make the larger sizes that were wanted in modern work. The plan was therefore contrived of having a second pair of rolls running in the opposite direction, placed at the back of the first rolls, the lower one of the second pair being raised just enough above the upper one of the first rolls to clear the bar in coming through; and the bar was passed back through the second rolls as readily as over the top roll in the ordinary arrangement. By this plan half the time was saved, and the rolls were all as handy for adjustment as the ordinary ones; and they gained the great advantage of being enabled to manufacture bars up to 60 feet length for deck beams and keels of iron ships in one length without a weld. They had rolled already quite successfully bars 46 feet long and 10 inches wide and only1/2inch thick, and expected to go up to 60 and even 70 feet length when the new rolling mill was got to work; the only chance of rolling such extreme lengths was by having a high speed of the rolls, so as to complete the work before the bar got too cold, and this necessitated the great power of the engines described, as the full power required for rolling had to be exerted at this increased speed. Mr. COCHRANE remarked that reversing gear had been very successfully employed at the Patent Shaft Works, Wednesbury, for rolling long links for suspension bridges ; the motion of the ordinary rolls being reversed by a lever shifting an ordinary reversing clutch, the engine shaft running

44

118

DOWLAIS IRONWORKS ENGINES.

always one way, returning the bar through the rolls in the opposite direction without any delay and without requiring it to be lifted at all; this plan had been at work several years with complete success, and was a more simple arrangement than the double set of rolls now described. Mr. MENELAUS replied that the objection to reversing gear was that it limited the speed at which the rolls could be driven, and he believed that it was not found applicable to rolls running above 45 revolutions per minute, on account of the violent shock in reversing the motion at a higher speed. They were obliged to employ a much higher speed in order to roll the lengths contemplated, and were working at nearly three times that speed at the Dowlais Iron Works; the ordinary rail rolls running at 120 revolutions per minute, and the other rolls for large sections at 110 revolutions, the rolls being of the full ordinary size, 21 inches diameter. He considered the speed of 120 revolutions per minute was necessary to obtain the long bars of 1/2 inch or even3/8inch thickness that were now required to be made, and they had now doubled the speed of the old Welsh rail mills which was only 60 revolutions per minute. Mr. COCHRANE asked whether the quality of the bars was interfered with by the increase in speed, and whether the rolls running at such a high speed turned.out as good work. Mr. MENELAUS did not know of any necessary difference in the quality of the bars on account of the speed ; an inferior quality of iron was found to stand the rolling better at the high speed than at the old speed. The CHAIRMAN remarked that several specimens of wrought iron of great section and weight were shown at the Paris Exhibition, of French manufacture, which showed that an important advance had there been made over what had previously been practicable. It was a great desideratum to avoid the necessity for angle iron in the deck beams for iron ships and similar constructions, by getting the iron rolled complete with the required flanges, and also to obtain the beams in one length without a joint; and the new rolling mill that had been now described appeared to be a highly important step in accomplishing that object. He thought the paper was a very interesting and valuable one, and proposed a vote of thanks to Mr. Menelaus, which was passed.

45

Fig.l. Side Eleration.

Fig.2. End

Eleration:

Jig. 4*. Plan.

Fig. 3. Vertical Section of Blowing

cylinder.

Fig. 6. Side Eleration:

Fig.8.

Kg. 9.

Fig.10.

Fig.7. End

Eleration.

Fig.11. GeneralPlanofRolling

Mill.

Eg. 12. Front Elevation.

Fig.13. Transrerse Section.

Fig.14. Plan 0f Standard.

William George Armstrong (1810-1900) William George Armstrong, later to become Baron Armstrong of Cragside, was a yeoman's son who demonstrated his early promise by constructing miniature corn mills from discarded spinning wheel parts. Although he joined a law firm in Newcastle-upon-Tyne as his first profession, Armstrong's continuing interest in hydraulics led him to publish papers on the possible uses of water-powered machinery in the Mechanics'Magazine of the 1830s and 40s. Over the next ten years, Armstrong's interest in water gave rise to hydroelectric power generation, the provision of clean water for the population of his home city, and the invention of the velocity-activated pipe valve. His thinking turned to the use of water in distributing power, which he achieved by patenting a hydraulic crane, the first example of which worked the Tyne quayside. Such was the demand for Armstrong's machinery that he was enabled to open his first manufactory, at Elswick. The date of his preliminary business partnership, January 1847, was a propitious one, coinciding as it did with the creation of the Institution of Mechanical Engineers, where Armstrong would later become President. His new position as engineer allowed his permanent retirement from the legal profession. Armstrong continued to invent. He improved upon the efficiency of hydraulic machines, freeing them from dependency on conventional reservoirs by the invention of the high-pressure accumulator. With the advent of the Crimean War, however, Armstrong turned his attention to munitions and armaments. He virtually reinvented the artillery piece, championing the use of steel-jacketed breech-loading guns with rifled barrels. His Elswick Ordnance Company produced weapons of all types. A related enterprise commenced the building of ships. Later mergers would create the well-known firms of Armstrong, Mitchell & Co. (1882), for shipbuilding and Armstrong, Whitworth & Co. (1897), for weapons manufacture. Armstrong's later years were occupied with electrical researches at his home, Cragside, and by distinguished public service. A contemporary obituary noted, without exaggeration, that "by his death Newcastle lost her greatest citizen".

51

126

ON WATEK PBESSURE MACHINEKY.

The subject of Water Pressure Machinery first attracted the writer's attention about twenty-three years ago, on occasion of his noticing a small stream of water which flowed from a great elevation down a steep declivity and turned a single overshot wheel near the end of its course. Observing that the portion of the fall which was thus utilised by the wheel was not more than a twentieth part of the whole descent, he was forcibly struck with the inadequacy of the wheel as a means of realising the power of such a fall ; and conceiving it practicable to render the entire head available by bringing down the water in a pipe, and causing it to act by pressure upon a suitable machine at the bottom, he applied himself to devising an engine to be worked by such a pressure. Other inventors, guided by the same idea, had been previously led to apply water pressure in various ways to driving machinery, although the employment of that power had never been extensively adopted; but with these applications the writer was at the time wholly unacquainted. After meeting with many difficulties and discouragements, the machine was produced, which after many years of seclusion in a lumber room has now been brought out for inspection upon this occasion. It is, as will be perceived, a species of rotary engine, admitting of a continuous and uniform flow of water through it, and exempt from all contracted passages. This machine was tried first in Newcastle, with the pressure from the street pipes equivalent to a column of about 200 feet; then in Gateshead with a still greater pressure; and in both cases it yielded a very high effect in relation to the theoretical power of the moving column. Up to this time the writer's object had been merely to provide a machine for turning to profitable account the power which great altitude gives to mountain streams; believing then, as now, that when roads and railways have rendered accessible the valleys where such

53

WATER PRESSURE MACHINERY.

127

streams abound, the water which descends from the heights will ultimately prove an efficient source of motive power. But having then had practical experience of the power residing in street water pipes, he conceived the idea of applying such power to the cranes at that time so slowly and expensively worked by hand upon the quay of this town. Accordingly after duly considering the means of effecting this purpose, a working model of a hydraulic crane was constructed, and exhibited at a meeting of the Literary and Philosophical Society of Newcastle, where it was connected with the town water pipes, and went through the several operations of lifting, lowering, and slewing in a very satisfactory manner. The next step was to carry this scheme into actual practice; and in the year 1846 the first hydraulic crane was erected at the upper end of the Newcastle Quay, where it has ever since continued to do good service in discharging ships. The crane model exhibited to the present meeting is nearly identical with the original model, and does not differ materially from the first actual crane erected in this town. The writer was soon enabled to introduce the hydraulic crane at Liverpool, and shortly afterwards at the New Dock at Grimsby, where, at the instance of his friend, the late Mr. Rendel, the engineer who constructed that dock, he also applied the same kind of machinery for opening and closing the lock gates and sluices. An extensive system of water pressure machinery was accordingly carried out at that dock ; and .the result afforded the first practical demonstration that the pressure of a column of water could be advantageously applied as a substitute for manual labour, not merely for the cranage of goods, but also for various mechanical operations in connexion with the entrances to docks. At all these places the effective column of water was about 200 feet head. At Newcastle and Liverpool the supply was derived from the pipes communicating with the town reservoirs; but at Grimsby a tower was built for supporting a tank, into which water was pumped by a steam engine. In the former cases the irregularity of pressure, consequent upon the variable draught from the pipes for the ordinary purposes of consumption, proved a serious disadvantage ; but this objection was not experienced at Grimsby, where the tank upon the tower furnished a separate source of power, undisturbed by

54

128

WATER PRESSURE MACHINERY.

any interferences. But the erection of a tower was a formidable undertaking, and so long as it remained a necessity, a great extension of the principle could not be anticipated. The writer therefore resorted to another form of artificial head, which possessed the advantage of being applicable at a moderate cost in all situations, and of lessening the size of the pipes and cylinders by affording a pressure of greatly increased intensity. The apparatus by which this is effected is named the " accumulator," from the circumstance of its accumulating the power exerted by the engine in charging it. The accumulator is in fact a reservoir giving pressure by load instead of by elevation; and its use, like that of every provision of this kind, is to equalise the duty of the engine in cases where the quantity of power to be supplied is subject to great and sudden fluctuations. The construction of the accumulator is shown in the vertical section, Fig. 1, Plate 30. It consists of a large cast iron cylinder A fitted with a plunger B, from which a loaded weight case C is suspended to give pressure to the water injected by the engine. The load upon the plunger B is usually such as to produce a pressure in the cylinder equal to a column of 1500 feet head; and the cylinder is made sufficiently capacious to contain the largest quantity of water which can be drawn from it at once by the simultaneous action of all the hydraulic machines connected with it. Whenever the engine pumps more water into the accumulator than passes to the hydraulic machines, the plunger rises and makes room in the cylinder for the surplus; but when, on the other hand, the supply from the engine is less for the moment than the quantity used, the plunger with its load descends and makes up the deficiency out of store. The accumulator serves also as a regulator to the engine; for when the plunger rises to a certain height, it begins to close the throttle valve in the steam pipe so as gradually to reduce the speed of the engine, until the descent of the plunger again calls for an increased production of power. With regard to the pumping engine used for charging the accumulator, the most approved form is that of two high pressure cylinders fixed horizontally, with double-acting pumps directly connected with the piston rods. At first a simple plunger pump at each

55

WATER PRESSURE MACHINERY.

129

end of the cylinder was used; then, with a view to",greater compactness, the pump behind the cylinder was discontinued, and the other made a double-acting one upon the combined bucket and plunger system. Finally a modification of this form of pump was adopted, by dispensing with the clack in the bucket and substituting an external delivery valve D, as shown in Fig. 2, Plate 30. In this arrangement the out-stroke of the pump causes the water contained in the annular space surrounding the plunger E to be forced into the accumulator, while a further supply of water enters behind the piston F through the suction valve G; in the in-stroke the water behind the piston is discharged through the delivery valve D, and half of it passes round into the annular space on the other side of the piston, while the remaining half is forced into the accumulator, the area of the plunger E being exactly half that of the piston F. Each stroke of the pump thus delivers the same quantity of water into the accumulator. This neat modification of the bucket and plunger principle, by which equally easy access is given both to the suction and the delivery valves, was the suggestion of Mr. Henry Thompson, the writer's late intelligent foreman. The introduction of the accumulator in the year 1851 removed all obstacles to the extension of water pressure machinery, which has since been applied in nearly all the principal docks, and in many of the government establishments in this country. Nearly 1200 hydraulic cranes, hoists, and other machines of that description have been applied; and 125 steam engines, collectively of more than 3000 horse power, are now daily at work to supply the pressure for working them. The system has also been adopted in many of the principal railway stations, not only for cranage, but also for working turntables, traversing machines, wagon lifts, hauling machines, &c. It is also extensively used for raising and tipping wagons in the shipment of coal, for opening and closing swing bridges, and for many other purposes. New forms of application are continually being developed, and no doubt can be entertained of its capability of further extension. The form of mechanism which prevails to the greatest extent in these various applications of water pressure consists of a hydraulic press

56

130

WATER PRESSURE MACHINERY.

with a set of sheaves used in the inverted order of blocks and pulleys, the object being to obtain an extended motion in the chain from a comparatively short stroke of piston. The general arrangement of the machinery for working a hydraulic crane is shown in Figs. 4 and 5, Plate 31; Fig. 4 being an elevation and Fig. 5 a plan. The pressure cylinder A for lifting the load is fixed horizontally below the surface of the ground in a chamber at the foot of the crane, and is fitted with the ram B carrying the pulleys C at its outer extremity. The lifting chain is made fast at one end to the cylinder A, and passes alternately round the moveable pulleys C and the fixed pulleys D at the inner end of the cylinder A; and is then led round the guide pulley E up the crane post F and along the jib to the load. The motion of the lifting chain is controlled by means of the handle G, acting upon the inlet and outlet valves which are kept closed by the weights H and I: by opening the inlet valve H the water is let into the cylinder A from the pressure pipe J, and raises the load; and by opening the outlet valve I the water escapes from the cylinder into the exhaust pipe K, allowing the load to descend. The travel of the ram B in the outward stroke is prevented from exceeding the proper limit by the pulley block C coming in contact with a stop connected with the handle G, which closes the inlet valve H and prevents the load from being raised too high. The return stroke of the ram is effected by the load suspended from the chain; and in the absence of any load a small supplementary ram L is employed to force the main ram B back, the slack chain being made to run out by the weight M. To meet variation of load it was formerly the practice to combine three of the pressure cylinders so as to act either separately or collectively upon the lifting chain; but the writer has more recently introduced a method by which a variation of power is obtained with a single bored cylinder fitted with a combined piston and ram. Fig. 7, Plate 32, is a longitudinal section of the cylinder and valves of a double power hydraulic crane; Fig. 8 is a section of the piston enlarged; and Fig. 9 shows the valves in detail. A is the cylinder fitted with the piston E and ram B; the water from the accumulator enters the valve chest F through the pressure pipe J and the inlet valve H. For the lower power the water is admitted

57

WATER PRESSURE MACHINERY.

131

to both sides of the piston E by opening the valve L, as shown in the drawing, in which case the power exerted and the water expended are each proportionato to the area of the ram B. For the higher power the valve L is closed and the valve M opened, so that the front side of the piston E is thrown open to the exhaust K, and the result both as regards power and expenditure is then proportionate to the full area of the cylinder A. It is seldom necessary to have more than a double power ; but where a lower or third power is required a smaller ram may be inserted within the other, but in this case it is necessary to make fast the larger ram while the smaller one is at work. For lowering the load the valves H and M are closed and the outlet valve I opened, allowing the water to escape from the cylinder A into the exhaust pipe K; at the same time the valve L is also opened to allow the water to follow up the piston in the inward stroke. In hydraulic cranes the power is applied not only for lifting the load, but also for swinging the jib, which latter object is effected by means of a rack or chain acting on the base of the moveable part of the crane, and connected either with a cylinder and piston, or with two single-acting cylinders applied to produce the same effect by alternate action. Figs. 4 and 5, Plate 31, show the arrangement of the two cylinders N and 0 for turning the crane, fitted with rams, acting by a chain passing round the base of the crane post F. The motion is controlled by means of a slide valve worked by a handle situated beside the handle G, so that while the water is admitted to one cylinder the other is open to the exhaust. The travel of the rams is limited by means of a tappet rod connected with the handle of the slide valve, whereby the crane is prevented from being turned round too far. The absence of any sensible elasticity in water renders the motions resulting from its pressure capable of the most perfect control by means of the valves which regulate the inlet and outlet passages; but this very property which gives so much certainty of action tends to cause shocks and strains to the machinery by suddenly resisting the momentum acquired by the moving parts. Taking for example the case of a hydraulic crane swinging round with a load suspended from the jib, the motion being produced by the water entering into one cylinder and

58

132

WATER PRESSURE MACHINERY.

escaping from the other, it is obvious that if the water passages be suddenly closed, the ram of the latter, impelled forwards by the momentum of the loaded jib but met by an unyielding body of water deprived of outlet, would be brought to rest so abruptly as to cause in all probability a fracture of the machine. So also in lowering a heavy weight, if the escape passage be too suddenly closed, a similar risk of injury would arise from the sudden stoppage of the weight, if a remedy were not provided. But these liabilities are effectually removed, in the case of a single-acting cylinder, by employing a relief valve in connexion with the water passages, consisting of a small clack valve N, Fig. 9, opening upwards against the effective pressure, so as to permit the pent up water in the cylinder to be forced back into the pressure pipe whenever it becomes subject to a compressive force exceeding the pressure given by the accumulator. In the case of a double-acting cylinder fitted with a piston and slide valve, or where two single-acting cylinders with rams working alternately are controlled by a slide valve, as in the instance of the cylinders N and 0, Fig. 5, for turning the crane, relief valves are employed in connexion with the slide valve, consisting of four small leather flap valves arranged as shown in Fig. 6, Plate 31. The passages PP communicate with the pressure pipe J,and he passages EE with the exhaust K. When the slide valve is moved in the direction of the arrow, the pressure is first cut off from the port R by the lap of the valve, the port S being still open to the exhaust K; at the same instant the flap valve T opens upwards and allows a small quantity of water to pass from the exhaust K into the port B to follow up the ram until brought to rest. When the slide valve arrives at the centre position, as shown in the drawing, the port S is closed to the exhaust; and the pressure in the port S being increased by the further motion of the ram before it is completely stopped, the second flap valve U is raised and a small quantity of water forced back into the passage P communicating with the pressure pipe J. When the slide valve is moved in the opposite direction, the two remaining relief valves are called into action in the same manner. By these means all jerks and concussions are avoided, and perfect control over the machine is combined with great softness of action.

59

WATER PRESSURE MACHINERY.

133

The method which the writer has most generally adopted for opening and closing lock gates by means of hydraulic pressure consists in applying to each gate a pair of cylinders with rams and multiplying sheaves similar to those constituting the hoisting apparatus in hydraulic cranes. One of these cylinders opens the gate and the other closes it, the whole of the machinery being placed in chambers beneath the ground, as shown in Figs. 10 and 11, Plate 33. The water is admitted from the pressure pipe J to the cylinder A through the inlet valve H by means of the handle G; the same motion of the handle also opens the outlet valve of the other cylinder B. The opposite motion of the handle G opens the outlet valve I, allowing the water to escape from the cylinder A into the exhaust pipe K, and at the same time admits the pressure to the cylinder B. A stop M connected with the handle G prevents the ram from travelling too far in the out-stroke, by closing the inlet valve ; and the return stroke of the ram is effected by means of the weight L. This arrangement is in use at the Sunderland Docks, the London Docks, the Victoria Docks, and the West India Docks, and is now being applied to the new 100 feet entrance to the Liverpool Docks. Another method has also been adopted for this purpose in some instances, which has been found to answer extremely well; Fig. 12, Plate 34, is a general plan of a dock entrance showing the arrangement of the lock gate machinery. Instead of connecting hauling cylinders with each gate, a line of shafting A, driven by a small water pressure engine B, is laid beneath the surface of the ground parallel with the coping on each side of the entrance, and by means of clutches is thrown into or out of gear with each gate crab. A wire from the engine extends the whole length of the shafting, so that the motion of the engine can be governed from the point where the work is being done. This method is already in use at several of the ports in South Wales, and is in course of being applied at the Jarrow Docks and the new dock at Silloth. The rapidity with which dock gates can be opened and closed by these appliances is limited only by considerations of safety to the gates, and the operation is in practice usually performed in about two minutes. The water pressure engines most recently constructed for driving the shafting consist of a combination of three oscillating cylinders u

60

134

WATER PRESSURE MACHINERY.

working cranks inclined at 180° to one another, with slide valves simply arranged, as shown in Figs. 13 and 14, Plate 35 ; and a small engine of this description is exhibited to the meeting. The cylinders A are fitted with plungers B instead of pistons, and are therefore only single-acting. The slide valves V are worked by the oscillation of the cylinders, communicated through the levers L. When the back end of the cylinder is depressed, the slide valve is lowered and allows the water to enter from the pressure pipe P through the pipe C to the cylinder, where it acts upon the plunger in the out-stroke. In the return stroke, the back end of the cylinder being raised, the cylinder port C is closed to the pressure pipe P and open to the exhaust E. A small relief valve is provided in connection with the cylinder pipe C, opening against the pressure, which prevents any shock when the communication with the exhaust is closed at the end of the return stroke. These engines have occasionally been made with pistons, so as to be double-acting; but for the heavy pressures employed where accumulators are used the single-acting arrangement with plungers is preferred. In nearly all cases in which hydraulic pressure has been applied for the moving of dock gates, it is also used for opening and closing the levelling shuttles, and in many cases also for working the capstans. The former purpose is effected by the direct application of a cylinder and piston fixed above the shuttle ; and the latter is accomplished by throwing the capstan C, Fig. 12, Plate 34, into gear with the shafting A, or by applying to it a separate engine similar to that which has been described. A good example of the application of hydraulic pressure to the opening and closing of scouring shuttles may be seen at Sunderland Docks, where by means of this power the enormous area of nearly 500 feet is opened in a few seconds and closed with equal rapidity. At the same place an example may also be seen of a heavy drawbridge worked by water pressure. The bridge is supported upon wheels, and is first lifted from its bearings to such a level as to enable the wheels to roll back upon a suitable railway, both operations being performed with great steadiness by the action of the water in little more than a minute.

61

WATER PRESSURE MACHINERY.

135

In working swing bridges by means of water pressure, a central press is applied to lift the entire bridge clear of its supports, and it is then turned by an application similar to that used for slewing a crane. The most remarkable specimen of a bridge of this description is one erected by the writer at Wisbech under the direction of the late Mr. Rendel. In this case an opening of 85 feet is spanned by a double roadway bridge in one leaf weighing about 450 tons. At present it is but little used as a moveable bridge; but notwithstanding its great length and weight it can be lifted and turned in less that two minutes. The power is derived from an accumulator charged by hand labour, and it is expected that one man by continued working will be able to store up sufficient power for opening and closing the bridge as often as necessary. This system of using an accumulator in connection with a hand force pump has also been successfully applied to a drawbridge erected by the writer near Carmarthen on the main line of the South Wales Railway ; and it might be advantageously adopted for many other purposes requiring a concentrated exertion of power with intervening periods of inaction. Amongst the numerous applications that have been made of water pressure for the purpose of rapidly lifting and lowering heavy loads, there is one which calls for special remark on account of its growing importance: namely, its application in connection with an accumulator for working vertical hoists at the landing stations of steam ferries, in cases where a railway traffic is required to be passed over a river or estuary not easily crossed by a bridge. The traffic of the Aix-laChapelle, Dusseldorf, and Rhurort Railway is by this means shipped and unshipped at the ferry across the Rhine ; and such is the rapidity and facility of the operation that a train of 12 coal wagons, weighing collectively 133 tons, can be transferred from the deck of the steamer to the railway, a height of about 20 feet, in 12 minutes. Each hoist lifts two wagons at a time, and raises its load in 10 or 12 seconds. These hoists are so arranged as always to accommodate themselves to the level of the boat, and always to stop at the exact level of the railway. It is not necessary to say more upon that branch of the subject which embraces the two important principles of accumulation and

62

136

WATER PRESSURE MACHINERY.

transmission of power, by means of which a steam engine can be economically applied to give safe and rapid motion to a multiplicity of machines intermittent in their action and distant from each other. But it remains to notice the applications made by the writer of the pressure derived from natural falls. When the moving power consists of a natural column of water, the pressure rarely exceeds 250 or 300 feet; and in such cases he has employed to produce rotary motion, in preference to the original scheme of a rotary engine, a pair of cylinders and pistons, with slide valves resembling in some degree those of a high pressure engine, but having relief valves to prevent shock at the return of the stroke, as shown in Fig. 6, Plate 31, already described. Where the engine is single-acting, with plungers instead of pistons, as in the water pressure engine already described, the relief valves are greatly simplified, and in fact are reduced to a single clack in connexion with each cylinder, opening against the pressure, which is the same as the relief valve N, Fig. 9, in the valve chest of the hydraulic crane. The water pressure engines erected by the writer at Mr. Beaumont's lead mines at Allenheads in Northumberland present examples of such engines applied to natural falls. They were there introduced under the advice of Mr. Sopwith, and are now used for the various purposes of crushing ore, raising materials from the mines, pumping water, giving motion to machinery for washing and separating ore, and driving a saw mill and the machinery of a workshop. In all these cases nature, assisted by art, has provided the power. Small streams of water, which flowed down the steep slopes of the adjoining hills, have been collected into reservoirs at elevations of about 200 feet, and pipes have been laid from these to the engines. Another application of hydraulic machinery at the same mines is new being made in situations where falls of sufficient altitude for working such engines cannot be obtained, which from its novelty deserves special notice. For the purpose of draining an extensive mining district and searching for new veins, a drift or level nearly six miles in length is now being executed. This drift runs beneath the valley of the Allen nearly in the line of that river, and upon its course three mining establishments are being formed, At each of these

63

WATER PRESSURE MACHINERY.

137

power is required for the various purposes above mentioned, and it was desired to obtain this power without resorting to steam engines. The river Allen was the only resource, but its descent was not sufficiently rapid to permit of its being advantageously applied to water pressure engines. On the other hand it abounded with falls suitable for overshot wheels, but these could not be applied to the purposes required without provision for conveying the power to many separate places. Under these circumstances it was determined to employ the stream through the medium of overshot wheels in forcing water into accumulators, and thus generating a power capable of being transmitted by pipes to the numerous points where its agency was required. In this arrangement intensity of pressure takes the place of magnitude of volume, and the power derived from the stream assumes a form susceptible of unlimited distribution and division, and capable of being utilised by small and compact machines. A somewhat similar plan is also adopted at Portland Harbour, in connexion with the coaling establishment there forming for the use of the navy. The object in that case is to provide power for working hydraulic cranes and hauling machines, and more particularly for giving motion to machinery arranged by Mr. Coode, the present engineer of the work, for putting coal into war steamers. A reservoir on the adjoining height affords an available head of upwards of 300 feet; but in order to diminish the size of the pipes, cylinders, and valves connected with the hydraulic machinery, and also with a view of obtaining greater rapidity of action, a hydraulic pumping engine and accumulator are interposed, for the purpose of intensifying the pressure and diminishing the volume of water acting as the medium of transmission. In concluding this survey of the system of water pressure machinery which has grown to such large proportions, chiefly during the last seven years, the writer feels that the present occasion has afforded the fittest opportunity he could have taken of communicating the particulars contained in this paper; and he trusts that the details of the progress of this system may not have proved uninteresting.

64

138

WATER PRESSURE MACHINERY.

Mr. W. G. ARMSTRONG exhibited a working model of a hydrauliccrane, and a specimen of the most recently constructed water pressure engines of 8 horse power, together with the original model engine that was first contrived for the purpose. He observed that the application of water pressure to produce rotary motion in an engine was not successful until the relief valves were designed; the great difficulty experienced being the stubbornness of the water and its want of elasticity. The relief valves were almost equally important in hydraulic cranes and other machines of that class: if in lowering a weight by the crane, the outlet valve for the escape of the water were suddenly closed, the piston would be still driven onwards by the momentum of the weight in motion; and as the water was prevented from escaping in front of the piston by the valve being shut, something would inevitably give way or burst to relieve the sudden pressure, on account of the water jammed up in the end of the cylinder being incompressible. This was however entirely obviated by the adoption of the relief valves, and the effect of the momentum of the moving weight acting upon the piston was simply to force open the relief valve at the instant of the pressure in the cylinder becoming any greater than that in the pressure main, allowing a little of the water to be squeezed back into the main through the relief valve. In consequence a heavy weight could be lowered rapidly, and the outlet valve be then suddenly closed; but the weight still proceeded a little further, and was stopped softly, without any shock or jerk, the motion being gradually retarded by the resistance of the uniform pressure of the water column acting upon the front of the piston through the opening of the relief valve into the pressure main ; this caused the non-elastic water to act as if elastic, and with the advantage that the pressure could never exceed that of the original water column. In slewing the crane round in either direction the same cushioning action was obtained; the crane could be swung round quickly and the valve suddenly closed shortly before reaching the required spot, when the whole moving weight was safely brought to rest without any shock, by the effect of the resisting pressure steadily opposing its motion. The action of the relief valves gave the water in fact a sort of fictitious elasticity, which was just what was wanted in stopping the motion of

65

WATER PRESSURE MACHINERY.

139

heavy bodies; and there was no waste attending this action, as the water was simply forced back into the store of the pressure main. In the case of slewing the crane round, the moving weight dragged the piston forwards a little after the water was shut off, leaving a vacancy behind the piston which was filled by the entrance of a little water from the escape pipe through the second relief valve, so as to allow the weight to come to rest gradually and without any recoil. Mr. T. SOPWITH considered that very important advantages were gained by this application of water pressure to drive machinery, as it admitted of the power being collected in reservoirs where it could be obtained without cost, and then conveyed to any distance where it might be wanted. The first of these engines had been applied under his own advice for drainage of water, and for raising, crushing, and washing the ore, at lead mines in a hilly district at Allenheads, Northumberland, belonging to Mr. Beaumont, where steam power would have been impracticable from the great cost of conveying coals to the works. Four of these water pressure engines had now been working in that neighbourhood for six years with complete success. He could bear testimony to the admirable working and the efficiency of these engines; and the results were so satisfactory that he was having three others erected at the newr works now in progress in the same district. Mr. W. FAIRBAIRN considered the country generally was greatly indebted to Mr. Armstrong for having worked out so completely this very useful and efficient engine, which would prove of great advantage in a number of cases where water power was supplied without cost by mountain streams, but had been hitherto lost from the want of suitable means for making the power available at places where it could be usefully applied for driving machinery. The system of relief valves that had been introduced was a very ingenious and beautiful plan for preventing the shock of the water at the stoppage of the piston, and removing the serious defect of non-elasticity in the water. He remembered having, about 20 years since, constructed an engine for forcing water under heavy pressure, and had experienced much difficulty from that cause; and was obliged to admit a little air under the piston in order to obtain a certain extent of elasticity, or the

66

140

WATER PRESSURE MACHINERY.

engine would otherwise have knocked itself to pieces by the severe shock produced at each stroke. These water pressure engines were certainly of great value for situations where coals could not be obtained for steam power, inasmuch as they were calculated to work economically; he would however like to ask their comparative value in cases where steam was available, and under what range of circumstances water could be applied direct as the driving power. He remembered in one of the early applications of Newcomen's steam engine, at Hebburn Colliery in that neighbourhood, the engine was employed to pump water on to a water wheel for driving the winding drum, before any attempt was made to apply the steam power direct to winding. The plan of a hydraulic hoist for a ferry was a valuable application for crossing railway trains over a river that could not be spanned by a bridge, from its great safety and certainty of action; and for working cranes the water pressure machinery was highly advantageous on account of the complete control that it afforded, and its soft action which greatly facilitated the lifting of heavy loads. Mr. J. ANDERSON had several of the water pressure cranes in constant employment at Woolwich Arsenal, and great success and advantage had attended their working. He noticed one point that had not been referred to in the paper, the objection likely to be felt at first of liability to stoppage or accident of the machinery from the water freezing in winter time; but he could state that during two winters' work no difficulty had been experienced from this cause at Woolwich, and he had found that a single jet of gas kept burning under each crane in the frosty weather was sufficient to prevent any risk of its occurrence. Another point to be noticed was the construction of the joints of the pipes, as the success of such an apparatus mainly depended upon the perfection of the joints for standing under the heavy pressure to which they were exposed; but the very simple and perfect make of joint had proved completely successful, and they had no trouble from leaky joints. There was also an important point of advantage in the great economy of working that was obtained by the use of water power cranes instead of hand labour; and a great saving was effected in

67

WATER PRESSURE MACHINERY.

141

the cost of the work, as well as an important advantage in the expedition with which it was done. There had been £30,000 expended at Woolwich Arsenal for fitting up 10 hydraulic cranes with their pumping engines and boilers, during the Crimean war, for the purpose of loading vessels direct from the quay side by powerful self-acting cranes; instead of the previous plan of loading most of the vessels by means of barges, which involved two transfers of all the load and incurred very heavy expense and serious delay. The saving effected in the demurrage charge per day for the vessels whilst loading was found alone sufficient to clear off the whole of this outlay upon the machinery in a short time. Mr. T. SOPWITH observed in reference to the comparative economy of water power and steam, that it would depend entirely upon the local circumstances of the situation as to the cost of coals, and the extent of water supply. Wherever the carriage of coal became expensive, and water pressure could be obtained even at a considerable distance, there would be a great economy in employing water power, as the current expenses of working would be only the interest of the outlay in pipes for bringing the water to the place required. The water engines at Allenheads, already referred to, were situated at a height of 1300 feet above the sea, where a supply of coal could be obtained only with great difficulty and at a heavy cost; but there was an abundant and constant supply of water from the mountains, from a great height above. In such a case the relative economy of water power was of course very great, and it was in fact the only power practically attainable in some of the situations where hydraulic engines were employed in the Allendale mines. Mr. T. E. HARRISON had used the water pressure cranes extensively for several years, and could speak very highly as to their value and efficiency ; at the Newcastle station this power had been constantly used for seven years in lifting all the coke for charging the locomotive tenders, the chaldron wagons being each raised by a hydraulic lift to an upper platform for discharging by shoots into the baskets on the coke stage so as to save all hand labour. The same power was also used for working the large engine turntable and the cranes in the goods warehouse. In the case of these cranes the pressure of water from v

68

142

WATER PRESSURE MACHINERY.

the town waterworks reservoir was the only pressure required, and the water used was afterwards supplied to the locomotive engines, having spent its pressure in working the cranes and been then discharged into the engine water tank; the power for working the cranes was thus obtained without additional cost, excepting a nominal annual payment to the Water Works Company for the use of the water in transit. This water power was also more economical than any other for such cases as the extensive goods warehouses now constructing by the North Eastern Railway at Hull, with water pressure hoists and cranes; and they had in consequence come to the conclusion of doing away with all the cranes worked by steam power, and employing the water power throughout. Besides the greater economy in working, the water pressure machinery had also important advantages in convenience of application over a large extent of premises, and simplicity in construction and keeping in order. In the construction of this machinery there was not only great novelty in the details, but great mechanical skill was shown in perfecting the several portions so as to make the whole thoroughly satisfactory in working and durability ; it certainly formed one of the greatest steps in the application of water as a moving power for machinery, and they could not too highly estimate the ingenuity displayed in working it out so completely, and the great value of this contribution to engineering science. Mr. W. G. ARMSTRONG observed that in the application of the water pressure machinery where an artificial head of water had to be obtained, as in the case of working dock cranes, the real source of the power was the steam engine employed in pumping water into the accumulator, and the water acted simply as a convenient means of storing up the power and applying it whenever wanted at the distant points where the work had to be done. In the early application of a steam engine to pump water for driving a water wheel, it would of course have been more economical to drive the machinery direct by the steam power, avoiding the loss of power attending the intervening water wheel; but that was not the present question, which in the majority of cases was simply as to the mode of transmitting power where the direct application of steam was not available, the power

69

WATER PRESSURE MACHINERY.

143

having to be applied at numerous points, many of them a long way removed from the source of the power, and in different directions. At the Victoria Docks, in London, the area over which the power was extended was so great as to require four miles' length of mains to convey the water to the several cranes and hoists and the lock gates; and any system of shafting for the purpose was of course quite out of the question long before that extent of length was reached. Then it was not practicable to have steam engines in the warehouses, nor to extend steam pipes over the premises : so that there was no other plan than water pressure available for working the whole by power, excepting the use of compressed or exhausted air ; but both these latter involved serious practical objections, such as the extent of power lost in the compression of an elastic medium by the liberation of heat and the increased defect of pumps, and the circumstance that the leakages could not be detected by sight as in the case of water or steam. With the water pressure system there was of course some loss of power from the friction of the water in the mains, &c., but in transmitting power to a distance some sacrifice was unavoidable ; and this plan appeared to involve the least proportion of loss. But it was not so much a question of comparative economy of power as of safety and convenience; and in these respects no other means of transmitting power was to bo compared with water pressure. In the application to natural heads of water, it was an important consideration that these were generally met with in situations inaccessible for the erection and working of machinery, and where the power was not wanted; but by obtaining the power there free of cost, they could well afford to convey it to a distance where it could be employed for work, at the expense simply of the interest on the outlay for pipes and machinery. And as the means of communication by railways extended, no doubt this application of water power would become greatly extended, by enabling the work to be brought within reach of the source of power. In reference to the construction of the joints in the hydraulic machinery, that point had been the subject of much attention, as the pressures employed kept increasing progressively; and the result was that he found no plan stood so well as a small ring of gutta percha

70

144

WATER PRESSURE MACHINERY.

compressed into the joint, which was turned with a projection on one side fitting accurately into a corresponding recess on the other side of the joint, like a spigot and faucet, as shown in Fig. 3, Plate 30. The gutta percha ring was put into the bottom of the recess, which was in a dovetail form expanding outwards to confine the gutta percha in its place, and the ring was then compressed by screwing up the joint, which compelled the gutta percha to fill up the recess completely ; and it was found that the water was unable to pass the joint at the greatest pressure that was employed in working. The gutta percha, so long as it was protected from the air and exposed only on the edge to the water, appeared to be imperishable; and this was found to be the best joint for the hydraulic machinery, and gave no trouble under any pressure that they worked to, amounting now to as much as 700 Ibs. per square inch; the joints also admitted of being readily opened and re-made if desired. As to the liability of accident from freezing of the water in the engines or mains, he had found that when they were buried in the ground the hardest winters had never affected them; and consequently no precaution was required in the ordinary cases, as the whole of the portion of the apparatus containing water was generally sunk below the surface of the ground. In the cases where any portion was raised above the ground, the simple precaution of a gas burner as had been described was quite sufficient to prevent any risk of freezing. A still more simple and equally effective method was to employ a cock for emptying the cylinder of water after each time of using in frosty weather; so that there was no water exposed to freezing, the pipes underground being sufficiently protected by their position. Mr. W. FAIRBAIRN observed that they had now in Manchester a pressure of about 200 feet head of water in the street mains, and the introduction of water power for driving certain descriptions of machinery was well worthy of consideration, to avoid the inconvenience and risk of accident attending the use of steam power in many situations, such for instance as hoists in warehouses, even where from the low cost of coals the water power might not be less expensive than steam power. From the improvements already effected in the application of water power he thought a further extension of the system might confidently be anticipated.

71

WATER PRESSURE MACHINERY.

145

The CHAIRMAN observed they were greatly indebted to Mr. Armstrong for the very valuable and interesting paper he had given them, and the excellent models and specimens he had exhibited to the meeting; and he proposed a vote of thanks to Mr. Armstrong for his paper, which was passed.

72

WATER

PRESSURE MACHINERY.

Fig.1

Plate30.

Vertical' Section of Accumulater.

73

WATER

PRESSURE

Hydraulic Crane. General Arrangement of Machinary.. Fig.4. Elevation.

74

MACHINERY.

Plate31.

Fig.8. Section of Piston,

WATER

enlarged.

PRESSURE

Fig 7. Double .power Hydraulic Longitudinal of Cylinder

section

and

Valves.

MACHINERY.

fig.9. Detail.of Valves enlarged.

Fig.10. Eleration of

Fig 11. Plan,.

Prassurecylindersfor

cpening and closing Lock Gates.

WATER

PRESSURE

MACHINERY.

Plate.34.

Fig.12. Arrangement

of

Lock

General

Gate

Machiney.

Plan.

77

WATER

PRESSURE

Fig. 15. Vertical Section "through, Cylinder and Valve Chest.

Fig.14. Plan,.

78

MACHINERY,

plate35

Charles William Siemens (1823-1883) Carl Wilhelm Siemens was born near Hanover into a family which would become famous for its scientific attainments. One of several brothers, William was destined to become the engineer of the group, commencing his technical training at the age of fifteen at Magdeburg, and then Gottingen University. To further his knowledge of electroplating, William travelled to England in 1843, where he settled and prospered. After selling an improved electroplating process to Elkingtons of Birmingham, Siemens entered a phase of productive invention. The Society of Arts awarded a prize for his steam engine governor in 1850. The inventor's struggles to gain acceptance for his regenerative steam engine and condenser (1847) met with little success. However, Siemens' idea for a water-meter in 1851 published in Proceedings', secured his financial independence. Experiments in furnace design, made in association with his brother Frederick, resulted in the use of regenerative heat exchange and gaseous fuel. What became the Siemens-Martin process for steel manufacture was a tremendous innovation in terms of the temperatures achieved and the fuel economy. Appointed London agent for his brother Werner's telegraph company Siemens & Halske, William Siemens began to contribute to the new field of electrical engineering. Insulating cabling with gutta-percha proved to be an interesting engineering project, which had enormous consequences for communications. International links to India and America were made possible with incalculable benefits. The Siemens Brothers' Company, established in 1858, also produced electrical lighting, famously illuminating the British Library in 1879. With over a hundred patents to his name, Siemens was a prolific and energetic engineer. Once elected to the Institution of Mechanical Engineers, he contributed thirteen papers to the Proceedings, many of which outlined key discoveries or processes for which he was responsible. Siemens served as President in 1872-73.

79

137

DESCRIPTION OF A MACHINE FOR COVERING TELEGRAPH WIRES WITH INDIA-RUBBER. BY MB. C. WILLIAM SIEMENS, OF LONDON.

A submarine telegraph cable is composed of three essential parts:—1st, the conductor, which generally consists of a strand of seven copper wires twisted together, to give it strength and pliability: 2nd, the insulating coating, which consists almost without exception of several coatings of gutta percha put on while hot and in a semifluid state by means of piston and cylinder machines analogous to the presses used for making lead pipes; with intervening coatings of a bituminous compound, called Chatterton's mixture, to establish a more intimate union of the different layers of gutta percha : 3rd, the sheathing, which is added to protect the insulated conductor and to give strength to the cable, and consists generally of a hemp serving and a spiral covering of iron or steel wire. Respecting the conductor, it is important that it should consist of the best conducting material, in which quality pure copper far surpasses all but some of the precious metals and possibly pure aluminium. If the conductivity of silver is expressed by 100, that of pure precipitated copper may be taken at 90. The conductivity of the copper of commerce varies however between extraordinary limits; and it may be accepted as a rule that all foreign matter contained in it, whether metallic or otherwise, diminishes its conductivity. Thus 2 per cent, of alloy is known to reduce the conductivity of copper from 90 to 13, and even the best selected copper used for telegraph conductors varies in practice as much as 20 per cent, in conducting power. The foreign substance which it is most difficult to remove from the copper is oxygen; and a process to effect this would be of considerable value. The insulating covering of the conductor is the most delicate and essential part of the telegraph cable. It has to form an effectual barrier against escape of the current throughout the whole length, for v 81

138

INDIA-RUBBER COVERING MACHINE.

a single flaw in this coating causes the failure of an entire cable. Nor does a flaw show itself always in testing cables however thoroughly previous to their submersion ; for experience has proved that flaws are produced gradually by the chemical action of the galvanic current itself in any places where the thickness of insulating coating has been considerably below the average, either owing to an air bubble forced open by the pressure of the water, or owing to an eccentric position of the conductor. The latter defect may be produced either in the covering machine, or afterwards by exposure of the cable to the heat of the sun, or to a strain producing a permanent elongation of the copper; in consequence of such elongation the gutta percha endeavours to return to its original length and causes the copper core by degrees to assume a serpentine position in the covering. Gutta percha was till lately thought almost a perfect non-conductor of electricity; but in dealing with long lines of submarine electric telegraph its conductivity has become well established and is often a source of painful anxiety to the electrical engineer, obliging him to search for other insulating materials. Glass and other vitreous substances, which possess the highest insulating properties, are of course inapplicable; and amongst the resinous insulators there is none that combines insulating quality with tenacity and other desirable mechanical properties in so high a degree as india-rubber. The accompanying table shows the respective non-conducting or insulating power of gutta percha, india-rubber, and Wray's mixture, which last is a compound of india-rubber with shellac and pounded flint; and of the two latter substances combined :— Specific Non-conducting and Inductive Power of Gutta percha, India-rubber, &c. Specific Non-conducting Power. Temperature Fahrenheit

82

Specific Inductive Power.

52°

72°

92°

52°

72°

92°

Gutta percha

3.01

1.20

0.38

1.00

1.00

1.00

India-rubber

50.70

45.10

27.60

0.68

0.62

0.70

Wray's mixture

23.60

26.00

38.40

0.77

0.63

0.96

Combination of India-rubber and Wray's mixture

38.40

40.55

38.40

0.77

0.78

...

INDIA-RUBBER COVERING MACHINE.

139

The great superiority of india-rubber and its compounds over gutta percha in insulating power is at once apparent, india-rubber itself being 16 times better than, gutta percha as a non-conductor at a temperature of 52°, and 70 times better at 92°; and the combination of india-rubber and Wray's mixture is on the average as good a nonconductor as india-rubber, while its inductive power, which causes retardation of the electric current in its passage along the wire, is only three quarters that of gutta percha. To these advantages the greater tenacity of india-rubber and its greater power to resist heat have to be added. India-rubber has been tried for the purpose of insulating telegraph conductors more than -twenty years ago, when it was employed by Jacobi of St. Petersburg for underground telegraphic lines. In. 1846 Dr. Werner Siemens employed it for the same purpose, previous to his application of gutta percha. About the same time india-rubber was put to the same use in this country, and it is said remains still in good condition in Portsmouth harbour. There is nothing new therefore in substituting india-rubber and its compounds for gutta percha in insulating submarine or other telegraph conductors : the present paper has special reference to a new method of effecting the covering. The method hitherto adopted consists in cutting the indiarubber into strips, and winding these strips spirally upon the wire to be insulated : a tedious and expensive operation, which has to be repeated several times to afford any security that the water is entirely excluded from the wire. The insulation of the wire depends in fact upon a perfect joint being formed throughout between the strips; for it is evident that where the strips overlap a spiral channel is formed, which, if in any one place penetrated will allow the water to spread till it may chance to find a transverse passage into the spiral channel of the next lower coating, and so forth until it reaches the wire. Formerly the layers of india-rubber simply touched one another, and could readily be displaced; but lately a process of soldering the spiral layers has been introduced by Messrs. Silver, which greatly increases the security of the coating, although it does not remove the objections to the spiral channels which must always be formed in lapping. This

83

140

INDIA-RUBBER COVERING MACHINE,

process of soldering consists in exposing the covered wire to boiling water for about half an hour, when a most perfect cohesion between adjoining surfaces is produced. The india-rubber so treated adheres to the fingers, or feels sticky; it also loses part of its elasticity and strength. It may therefore be inferred that -the heat produces some chemical alteration in the material, changing the gum into an oil. It has been observed that india-rubber so heated has gradually changed bodily into a viscid liquid, where it is in contact with the metal conductor, so as to render it unsafe to be used. The method of covering which it is proposed to substitute for the above combines the advantages of comparative cheapness and certainty of result with that of rendering the application of heat unnecessary. The operation is based on the well known adhering property of india-rubber, when two fresh-cut surfaces are joined together under considerable pressure. The mechanical problem consisted in the construction of a machine which would draw the india-rubber tight upon the wire, so as completely to exclude air; and would then cut the india-rubber at the proper inclination, and join the fresh-cut edges together at the same instant under a sufficient pressure to make the joint perfect. The machine finally arranged for this purpose is shown in Figs. 1 to 5, Plates 27, 28, and 29, one quarter full size. Fig. 1, Plate 27, is a side elevation; Fig. 3, Plate 28, an end elevation partly section ; and Fig. 5, Plate 29, a plan partly sectional; The machine consists of two grooved pressing rollers A and B, Figs. 1, 2, and 3, Plates 27 and 28, and of two cutting or shearing rollers CC, all of which are of hardened steel, and are shown enlarged to half full size in the section, Fig. 4, Plate 28. On each side of the groove in the pressing rollers A and B is a small cylindrical portion, as shown enlarged to double full size in Figs. 9 and 10, Plate 30, of a breadth equal or nearly so to the thickness of the intended coating to be applied; but these cylindrical sides must be slightly rounded off towards the groove and sharp on the outer edge, as shown at DD. The cutting rollers C are so placed on each side of the grooved rollers that in turning round their cutting edge crosses the edges of the grooved rollers a little before the centre line of the machine, as shown

84

1NDIA-RUBBER COVERING MACHINE.

141

double full size in Figs. 7 and 9, Plate 30, at a point where the distance between the edges of the grooved rollers is about equal to half the thickness of one of the strips of india-rubber used. The axis of the cutting rollers is slightly inclined to the axis of the grooved rollers, as shown in the end elevation, Fig. 3, and plan, Fig. 5; so that being pressed against the latter by means of set screws they only touch hard at the shearing point, as seen in Figs. 9 and 10, Plate 30 The wire to be covered and the two strips of india-rubber for covering it are guided into the machine by suitable guides E, Figs. 1 and 5. The two strips in closing upon the wire are drawn tight over it by the inner edges of the grooved rollers A and B; and being caught between the closing cylindrical portions of the grooved rollers are compressed to one fourth their original thickness, the material being forced outwards from the middle; the cutting rollers C then suddenly intersect them, as in Fig. 9, Plate 30, cutting off the superfluous breadth of strips and at the same time preventing further escape of the material towards the sides. As the edges of the grooved rollers continue to close upon one another, the material remaining between them can only escape inwards, by which means the two fresh-cut edges are brought one upon the other under a heavy rolling pressure, from which they glide inwards towards the groove, as in Fig. 10, and in so doing form a complete and permanent joint, Fig. 11. In order to effect several successive coatings, a train of machines is provided, as shown in Fig. 6, Plate 29, so placed that the wire to be coated passes in a straight line through them all, receiving in each successive machine an additional coating, with the longitudinal seams at right angles to those of the previous and succeeding coatings, as seen in Fig. 11, Plate 30, which is effected by the different angular positions in which the machines are placed. The last machine in the train is supplied with strips of cloth or felt covered with india-rubber, which is also capable of being joined by compression of the fresh-cut edges, and is extremely useful in adding firmness and protection to the insulated conductor. This machine is also applicable, with certain modifications of details, for covering wire with the compound of india-rubber, shellac, and pounded flint, known by the name of Wray's mixture, which

85

142

INDIA-RUBBER COVERING MACHINE.

possesses in common with india-rubber very remarkable insulating properties. The machine is also applicable, with great apparent advantage, for the manufacture of india-rubber tubes, and for several other similar purposes. In producing tubes by this process, a spiral or tube of wires is first prepared, which is coated with india-rubber in one or several layers, with or without intermediate layers of canvas previously coated with india-rubber. The spiral wire is then either withdrawn or left to support the tube, which is finally subjected to the vulcanising process. In order to produce a submarine cable, an outer covering is required for protection and strength. Instead of the ordinary hemp serving and iron sheathing, the author proposes to saturate hemp yarn with a cement consisting of ordinary marine glue mixed with a certain proportion of pitch and shellac, applied to the yarn in a fluid state and under pressure so as to penetrate the fibre completely. Two or more layers of this yarn are put upon the insulated conductor by means of a train of machines, which cause each strand to be drawn tight uniformly, and to pass separately through a heated chamber, so as to soften the cement and unite the yarn in complete layers upon the core, winding alternately right and left. The covering thus produced combines great tensile strength and lightness with the power to exclude the sea water from the core. It thus adds very considerably to the insulating coating, whereby the retarding effect of induction is greatly diminished; and forms a thorough protection to the more tender coating of highly insulating material. The necessity for a metallic sheathing is however not entirely avoided, in order to afford protection against abrasion and against marine animals; and this sheathing is proposed to consist of very thin brass or iron wire wound on in the form of a tight lapping while the cement is still soft, so as to be imbedded completely in it. The cable is then drawn through a hot die, which causes the superfluous cement to cover the wires completely and to preserve them from rusting. A cable so prepared combines the qualities essential for crossing deep and broad oceans. Its specific gravity will not exceed 1.5, which experience has proved to be the most desirable weight for submersion,

86

INDIA-RUBBER COVERING MACHINE.

143

and its tensile strength is such that it will support 15 miles of its own length in sea water, instead of only 3 miles which is the length an ordinary iron-sheathed cable will support. The sheathing of this cable will not be acted upon by sea water, and will retain its full strength therefore in case it should have to be taken up for repairs : it will not be liable to form kinks, which are fraught with danger to the insulation. The chief advantage however is supposed to reside in the insulating coating, which consisting of a succession of perfect tubes of the most highly insulating and tenacious material known, unaltered by heat or solvents and thoroughly protected against external injury, offers the greatest chances for permanent efficiency that could well be realised. For shore ends this cable should receive an additional external covering of strong wires to resist the effects of anchors and violent abrasion; and these wires in their turn should be covered with saturated fibre to render them durable. The experience with long submarine cables has hitherto been anything but satisfactory; but there is in the writer's opinion no reason to prevent their being made very permanent and valuable property, if only the experience now gained is turned to good account.

Mr. SIEMENS exhibited the machine in action, covering pieces of wire with india-rubber, showing that the joint made by rolling the two fresh-cut edges together under a heavy pressure was so strong that the india-rubber covering would tear at any other part as readily as at the joint. He showed also a number of specimens of the different descriptions of telegraph cable now in use. The process of joining the strips of india-rubber by the machine depended on the well known property of india-rubber, that when two perfectly clean fresh-cut surfaces were pressed together with great force they would unite as completely as two pieces of iron welded together. After many trials for effecting this by machinery, he had now succeeded perfectly with the machine exhibited, in which the two cut edges made by the cutting wheels on each side were instantly pressed together between

87

144

INDIA-RUBBER COVERING MACHINE.

the pressing rollers and joined without having been ever exposed to the atmosphere. This was the essential point in the machine, as any exposure of the cut surfaces however momentary interfered with the perfection of the joint. In putting on a series of coats of india-rubber for making telegraph cables, a train of machines was employed through which the wire was passed in a continuous line, the joints in each successive covering being in a line at right angles to those in the previous covering, which gave a greater security against failure at the joint. This insulating covering had been subjected to severe tests, and proved highly satisfactory and superior to any other mode of insulation. Gutta percha, which had hitherto been the material used for covering telegraph wires, was a good non-conductor; but its resistance to the passage of an electric current was only relative, like that of all other insulating materials, and it would conduct to a certain extent, the conducting power being about 3 trillion times less perfect than that of mercury, which was adopted as the standard of comparison. But india-rubber had much less conducting power than gutta percha, being 16 times better as a non-conductor at a temperature of 52°, and 70 times better at 92°. The insulating power of india-rubber was moreover less affected by difference of temperature than was the case with gutta percha; and in the combination of india-rubber and Wray's mixture, which he had produced, the average insulating power was not less than that of india-rubber, while it was to a less extent affected by change of temperature. Before however a current of electricity could pass along the wire, it had to induce a statical charge in the insulating material throughout the whole length of the wire, as in a Leyden jar; and the delay or retardation thus produced depended on the inductive power of the insulating material, which was independent of its insulating or non-conducting power, but affected by its thickness, the inductive power diminishing as the thickness was increased. A thicker coating of the insulating covering therefore offered less resistance by induction to the passage of an electric current, and allowed of more rapid speaking. In this respect also india-rubber and its compounds had an advantage over gutta percha, its inductive power being about three quarters that of the latter.

88

INDIA-RUBBER COVERING MACHINE.

145

In the use of gutta percha as the insulating material, a great amount of care was necessary in the process of coating the wire, and there was great risk of imperfection in the covering. In the submarine telegraph between Rangoon and Singapore the cable was very good for many miles, but a point was then found to exist where the insulation failed from a defect in the original construction of the gutta percha coating ; and such defects were liable to arise in the manufacture from various causes. In covering the wire the gutta percha was squeezed forwards in a semifluid state through the die, by means of a piston in a cylinder; and air bubbles were liable to get enclosed within its substance, which were so minute as not to be detected at the time of manufacture, though the cable was tried under a pressure of 600 or 1000 Ibs. per square inch; but they were sufficient to impair the insulation at the part where they occurred, and ultimately cause the failure of the cable. Moreover the manufacture was a hot process, as the gutta percha had to be kept soft in coating the wire; and if a slight delay took place in the operation, the gutta percha was too much softened at that part, and the weight of the wire cable itself made the coating thinner on one side than the other, so that the insulation was defective; the electric current afterwards sent through the wire was constantly leaking out more or less at the imperfectly protected parts, and caused a chemical action on the gutta percha, gradually decomposing it at the leak and increasing the amount of leakage. If the finished cable were allowed to lie for only a quarter of an hour exposed to a hot sun, it would be completely spoiled, as the heat would soften the covering and the core would take an eccentric position by sinking through the gutta percha by its weight \ and in the event of a strain coming on the cable in laying it, the copper core being non-elastic would remain permanently stretched, while the gutta percha would be constantly endeavouring to regain its original length, forcing the copper by degrees into a serpentine curve. These difficulties had at present caused failures to a greater or less extent in all submarine cables constructed with gutta percha. But in the process now described it was expected that the chances of failure through defects of manufacture would be much diminished, as there was less liability to accidental imperfections in the work, and the durability of the cable was not affected by the temperature to which it was exposed.

w 89

146

INDIA-RUBBER COVERING MACHINE.

Mr. E. A. COWPER observed that the Rangoon and Singapore telegraph cable would support only 3 miles of its own length in water, and its specific gravity was 3 times that of water; but the new indiarubber cable had a specific gravity of only 11/2 times that of water, and would consequently support 12 miles of its length in water, if of only the same tensile strength ; it appeared really to support 15 miles of its length, which was a most important advantage in point of safety, as submarine cables were liable to have great lengths suspended between the summits of mountains at the bottom of the sea. The CHAIRMAN asked what would be the difference in cost per mile between the new cable and one covered with gutta percha. Mr. SIEMENS replied that for equal efficiency, or the same speed of speaking, the new cable would be the cheapest, because a thinner coating of india-rubber would be sufficient to produce an equal insulating effect: but if estimated by weight, a gutta percha cable would be the cheapest 011 account of the greater cost of india-rubber. The first cost of the cable was however a secondary question, the great object being to obtain a cable that could be depended upon for a number of years. In a gutta percha cable, if the covering were thin at any one place, then each successive current passing along the wire produced an alteration, since the gutta percha conducted at the leak by decomposition of the water contained in its substance ; and this action gradually disintegrated its substance and destroyed its insulating power at that part, so that the electric current soon made its escape there. The CHAIRMAN asked how long the new cable would last at work. Mr. SIEMENS replied that there was not one of the new cables laid at present, and it required that several hundred miles should have been down for some years in order to show practically its durability in work; but some miles had been made and tested with very satisfactory results, and there was good reason for expecting this construction of cable would prove far more durable than those hitherto laid. The CHAIRMAN moved a vote of thanks to Mr. Siemens for his paper, which was passed.

90

Plate 27 Fig 2. Side Elevation of Rollers. Fig 1 Side Eleration.

Fig 3. End Elevatio partly Section Scale, 1/4.th

Fig 4. Vertical Section enlarged. Scale 1/2 full size..

Fig 5. Plan Partly Sectional,

Fig 6. General; Arrartgement

of

Train

of

Machines.

Fig 7. Longitudinal Sections through, Pressing Rollers

Fig 9. Section at Fig 8. Section, at

C. C.

Fig

10. Section at H. H.

F. F. Fig 11 Section at II.

Henry Bessemer (1813-1898) Bessemer's well-known facility for metalworking was apparently inherited. His father had escaped the French Revolution, settling in Charlton, Hertfordshire, and eventually entered the chain-making and type-founding business. The young Bessemer was trained at his father's works. His early inventions included a forgery-proof deed stamp, a method of compressing plumbago for pencil-making and a process for the liquid rolling of glass. Henry Bessemer contributed to meetings of the fledgling Institution of Mechanical Engineers, showing a model railway axle in only the second official meeting on 28 April 1847. His paper on this axle pre-dated the creation of the Proceedings, and was therefore never published. Since Bessemer did not become a full member of the Institution until 1861, this lost connection must rank as one of the great missed opportunities of early IMechE history. The inventor continued to produce a stream of practical and profitable ideas, but his most important work on steel arose from investigations into guns during the Crimean War. The steelmaking process which bore Bessemer's name was the subject of a series of patents in the years 1854-69, including in 1855 a description of the Bessemer converter. The first announcement of his discoveries was given before the British Association in the following year. Bessemer steel was produced under license immediately, and at Bessemer's own works, built in Sheffield. It was in Sheffield, during the Institution of Mechanical Engineers' Summer meeting of 1861, that he read the Proceedings paper on that topic, one of the very few technical papers he produced. The widespread adoption of Bessemer steel, in railways particularly, ensured world-wide recognition for the author of the process. Many Institution members - William Menelaus, John Ramsbottom and Daniel Adamson particularly - were instrumental in perfecting Bessemer's ideas and ensuring their practical application. Towards the end of his life in 1897, over 10,500,000 tons of Bessemer steel were being produced, on an annual basis, internationally. Bessemer himself preferred more simple statistics: "picture...a giant armour plate, 100 miles long, 5 feet thick and 20 feet high as representing the output of steel every year".

95

133

ON THE MANUFACTURE OF CAST STEEL AND ITS APPLICATION TO CONSTRUCTIVE PURPOSES. BY MR, HENRY BESSEMER, OF LONDON. The mode of manufacturing Cast Steel, which now forms so important a branch of the Sheffield trade, was discovered in the year 1740 by Mr. Benjamin Huntsman of Handsworth near Sheffield; who subsequently established steel works at Attercliffe, where his most valuable invention has ever since been successfully carried on. In its early stages many difficulties had doubtless to be overcome : materials for lining the furnaces and for making the crucibles had to be sought for and tested ; the peculiar marks of iron most suitable for melting had to be determined on by numerous experimental trials; and such was the difficulty at that time of making crucibles which would stand the excessive heat of melted steel that for a long period only very highly carbonised or "double converted" steel, which required the lowest temperature, could be successfully melted. The first products of a new manufacture, even while the invention still remains in a partially developed state, but too frequently stamp its subsequent character. Thus Huntsman's cast steel, although it was acknowledged to be a pure homogeneous metal of great value for certain purposes, was still looked upon as a hard and brittle material of very limited use, not bearing a high temperature without falling to pieces, and quite incapable of being welded : even within the last few years this has been the popular idea of cast steel. Improvements in its manufacture have however from time to time been introduced; and steel of a milder and less brittle character has long been made, capable of welding with facility and working at a high temperature without falling to pieces. Its uses have consequently been greatly extended, and the employment of cast steel for the best cutlery and edge-tools has now become universal; indeed the excellent quality of the cast steel at present made in Sheffield for these purposes is u

97

134

BESSEMER STEEL.

scarcely to be surpassed. Of late years several of the most enterprising manufacturers have sought to introduce cast steel for a variety of other purposes besides those for which it was originally employed, and it is now used in some form or other in almost every first class machine. Its employment as a material for founding bells and various other articles in clay moulds, as carried out by Messrs. Naylor and Vickers, and the introduction of a valuable material by Messrs. Howell and Shortridge, under the name of homogeneous metal, are prominent examples of the successful adaptation of cast steel to engineering purposes. The manufacture of cast steel by Huntsman's process is so extensively practised and is so well known that it is unnecessary to do more than recall to mind that crude pig iron has first to go through all the stages of melting, refining, puddling, hammering, and rolling, in order to produce a bar of malleable iron as nearly pure as the most careful manipulation in charcoal fires can make it. Bar iron, on which so much labour, fuel, and engine power have been expended, thus becomes the raw material of this most expensive manufacture. In order to convert the wrought iron bars into blister steel, they are packed with powdered charcoal in large firebrick chests, and are exposed to a white heat for several days ; the time required for heating and cooling them extending over a period of 15 to 20 days. When thus converted into blister steel they are broken into small pieces and sorted according to the quality of the steel, which sometimes differs even in the same bar. For melting this material powerful air furnaces are employed, containing two crucibles, into each of which are put about 40 Ibs. of the broken blistered steel. In about 3 hours the pots are removed from the furnaces, and the melted steel is poured into iron moulds and formed into ingots of cast steel, from31/2to 4 tons of hard coke being consumed for each ton of metal thus melted. When large masses of steel are required, a great many crucibles must be got ready all at the same moment, and a continuous stream of the melted metal from the several crucibles must be kept up until the ingot is completed, since any cessation of the pouring would entirely spoil it: hence in proportion to the size of the ingot are the cost and risk of its production increased. The ordinary manufacture of cast steel is

98

BESSEMER

STEEL.

135

therefore obviously conducted at a great disadvantage. If cast steel is to supersede wrought iron for engineering purposes, it will be necessary to cease employing wrought iron as a raw material for this otherwise most expensive mode of manufacture. The extremely high temperature requisite to maintain malleable iron in a state of fusion has from the earliest period of the history of iron down almost to the present day rendered its purification in a fluid state practically and commercially impossible. Hence arise all those imperfections to which bar iron is subject, every small piece consisting of numerous granules partially separated from each other by scoria, and every large mass being produced only by piling together small bars, with the inevitable result of increasing the former imperfections ; for no two pieces of iron can be brought to a welding heat without becoming coated with oxide, and when this coating is rendered fluid by welding sand a fluid silicate of the oxide of iron is formed, covering the entire surface to be united. The heavy blows of the hammer or the pressure of the rolls may and do extrude the greater portion of this fluid extraneous matter, but it is never wholly removed from between the welded surfaces, and hence a portion of the cohesive force of the metal is lost at every such junction. When a bar of iron is nicked on one side and bent, the rending open of the pile clearly shows this want of perfect cohesion. Nor is this the only difficulty to be encountered; for in the production of large masses of wrought iron it is necessary to raise the temperature nearly to the fusing point, in order to render each additional piece sufficiently soft and plastic to become united to the bloom : this softening of the iron induces a molecular change in the structure of the metal; its natural tendency to crystallise is so powerfully assisted by the long continuance of the high temperature that its whole structure undergoes a change ; large and well defined crystals are formed almost independent of each other, and cohering so feebly to the other contiguous crystals as in some cases to separate with as little force as would overcome the cohesion of ordinary cast iron. In the substitution of cast steel for malleable iron, both these sources of difficulty are escaped : for the mass whether of 1 ton or 20 tons weight may be formed in a fluid state into a single block, wholly free from admixture of scoria, while

99

136

BESSEMER STEEL.

it is perfectly and equally coherent at every part; and the forging of such a solid block of metal into shape is only the work of a few hours, and as there is no welding of separate pieces it may be worked under the hammer at a temperature at which no molecular change will take place, the metal being far below its fusing point and much too solid to undergo that destructive crystallisation so common in large masses of wrought iron. Thus the difficulties and uncertainty attending the production of all large masses of wrought iron are wholly avoided in producing equally large masses of cast steel. But however desirable in the abstract it may be to employ cast steel as a substitute for malleable iron for engineering purposes, it must not be forgotten that there are several important conditions indispensable to its general use. Firstly, the steel must be able to bear a good white heat without falling to pieces under the hammer ; otherwise the process of shaping it will not only be expensive, but the partly finished forging may be spoiled at any moment by being overheated. Secondly, the steel should be of such a tough character as to admit of being twisted or bent into almost any form in its cold state before fracture takes place, whether the force be applied as a gradual strain or by a sudden impact. Thirdly, it should have a tensile strength at least 50 per cent, greater than that of the best marks of English iron. Fourthly, it must especially be soft enough to turn well in the lathe, to bore easily, and to yield readily to the file and chisel, so as not to enhance its original cost by the difficulty of working it into the requisite forms. The last is both commercially and practically an important condition, and one which will in future greatly determine the extent of its use. Steel to the engineer has hitherto stood in much the same relation as granite to the builder : the superior hardness, beauty of polish, and durability of granite as compared with other building stone are universally acknowledged, nature has provided it in great profusion, and it has only to be lifted from the earth and made use of; but the practical man has found that to drill a hole in granite for blasting takes days of labour to accomplish, that the stone blunts all the chisels, defies the saw, and is faced only at a great cost 5 hence the builder goes on using an inferior soft stone over which the tools have perfect command. The

100

BESSEMER STEEL.

137

problem to be solved therefore is how to produce cast steel that will take any form in the mould or under the hammer, that will yield quickly and readily to all the present cutting and shaping machines, and will retain all the toughness of the best iron with a much greater tensile strength, and all the clearness of surface, beauty of finish, and durability that so eminently distinguish the harder and more refractory qualities of the steel in common use. These desirable objects are believed by the author to be fully accomplished by his process of converting crude pig iron into cast steel at a single operation, forming the subject of the present paper. This process has now been in daily operation in Sheffield for the last two years. The apparatus by which it is effected is shown in Plates 30 to 33, which represent the arrangement at Messrs. John Brown and Co.'s, Atlas Steel Works, Sheffield : Fig. 1, Plate 30, is a side elevation, and Figs. 3 and 4, Plate 31, a front elevation and plan. The crude pig iron chiefly used in this process has been the hot-blast haematite pig smelted with coke, which is melted in a reverberatory furnace adjoining, and is then run into the converting vessel A, Figs. 1 and 3, Plates 30 and 31, in which its conversion into steel is to be effected. The converting vessel is shown enlarged in section in Fig. 5, Plate 32, which represents its position in filling, the melted pig iron being run into it by the spout B direct from the furnace. It is made of stout boiler plate and lined with a powdered silicious stone found in the neighbourhood of Sheffield below the coal and known as " ganister." The rapid destruction of the lining of the converting vessel was one of the great difficulties met with in the early stages of the invention : the excessive temperature generated in the vessel together with the solvent action of the fluid slags was found to dissolve the best firebrick so rapidly that sometimes as much as 2 inches thickness would be lost from the lining of the vessel during the 30 minutes required to convert a single charge of iron into steel The ganister now used however is not only much cheaper than firebricks, costing only about 11s. per ton in the powdered state, but it is also very durable : a portion of the lining of the vessel is shown which has stood 96 consecutive conversions before its removal. The converting

101

138

BESSEMER

STEEL.

vessel A is mounted on bearings which rest on stout iron standards CC, Figs. 3 and 4, and by means of the gearing and handle D it may be turned into any required position. There is an opening at the top for filling and pouring out the metal; and at the bottom of the vessel are inserted seven fireclay tuyeres, Fig. 9, Plate 33, each having seven holes, as shown enlarged in the longitudinal section and plan, Figs. 10 and 11. The blast from the engine is conveyed through one of the bearings E of the vessel, Fig. 3, Plate 31, into the tuyere box F, and enters the tuyeres at a pressure of about 14 Ibs. per square inch, which is more than sufficient to prevent the fluid metal from entering the tuyeres. Before commencing with the first charge of metal, the interior of the converting vessel is thoroughly heated by coke, with a blast through the tuyeres to urge the fire ; when sufficiently heated it is turned upside down and all the unburnt coke falls out. The vessel is then turned into the position shown in Fig. 5, Plate 32, and the melted pig iron is run in from the furnace by the spout B, the vessel beingkept in such a position during the time it is being filled that the holes of the tuyeres are above the surface of the metal. When the proper charge of iron has been run in, the blast is turned on and the vessel quickly moved up into the position shown in Fig. 6. The blast now rushes upwards into the fluid metal from each of the 49 holes of the tuyeres, producing a most violent agitation of the whole mass. The silicium always present in greater or less quantities in pig iron is first attacked, and unites readily with the oxygen of the air. producing silicic acid : at the same time a small portion of the iron undergoes oxidation, and hence a fluid silicate of the oxide of iron is formed,, a little carbon being simultaneously burnt off. The heat is thus gradually increased until nearly the whole of the silicium is oxidised, which generally takes place in about 12 minutes from the commencement of the process. The carbon of the pig iron now begins to unite more freely with the oxygen of the air, producing at first a small flame, which rapidly increases, and in about 3 minutes from its first appearance a most intense combustion is going on : the metal rises higher and higher in the vessel, sometimes occupying more than double its former space, and in this frothy fluid state it presents an enormous surface to the action

102

BESSEMER STEEL.

139

of the air, which unites rapidly with the carbon contained in the crude iron and produces a most intense combustion, the whole mass being in fact a perfect mixture of metal and fire. The carbon is now burnt off so rapidly as to produce a series of harmless explosions, throwing out the fluid slag in great quantities ; while the combustion of the gases is so perfect that a voluminous white flame rushes from the mouth of the vessel, illuminating the whole building and indicating to the practised eye the precise condition of the metal inside. The blowing may thus be left off whenever the number of minutes from the commencement and the appearance of the flame indicate the required quality of metal. This is the mode preferred in working the process in Sweden. But at the works in Sheffield it is preferred to continue blowing the metal beyond this stage, until the flame suddenly drops, which it does just on the approach of the metal to the condition of malleable iron : a small measured quantity of charcoal pig iron containing a known proportion of carbon is then added, and thus steel is produced of any desired degree of carburation, the process having occupied about 28 minutes altogether from the commencement. The converting vessel is tipped forwards and the blast shut off for adding this small charge of pig iron, after which the blast is turned on again for a few seconds. The vessel is then turned into the position shown in Fig. 7, Plate 33, and the fluid steel run into the casting ladle G, which is carried by the hydraulic crane H, being counterbalanced by the weight I on the opposite end of the jib. When all the metal is poured out of the converting vessel, the crane is raised by water pressure and turned round, as shown in Fig. 2, Plate 30, for the purpose of running the steel into the ingot moulds K. Instead of tilting the casting ladle for pouring into the moulds, it is made with a hole in the bottom fitted with a fireclay seating L, Fig. 8, Plate 33, and closed by a conical plug of fireclay M, forming a conical valve. The valve rod N is coated with loam and bent over at the top, and works in guides on the outside of the ladle, as shown in Fig. 7, with a handle 0 for opening and closing the valve. By thus tapping the metal from below, no scoria or other floating impurities are allowed to run into the mould, and the stream of fluid steel is dropped straight

103

140

BESSEMER STEEL.

down the centre of the mould right to the bottom, without coming in contact with the sides of the mould. The moulds are made of a slightly tapered form,' as shown in Fig. 8, so that as the ingot contracts in cooling it liberates itself from the mould completely on all sides; and the mould is removed by being lifted off the ingot when sufficiently set. The moulds are arranged in the moulding pit in an arc of the circle described by the casting ladle, as shown in the plan, Fig. 4, Plate 81. By this process from 1 to 10 tons of crude iron may be converted into cast steel in 30 minutes, without employing any fuel except that required for melting the pig iron and for the preliminary heating of the converting vessel, the process being effected entirely without manipulation. The loss on the weight of crude iron is from 14 to 18 per cent, with English iron worked in small quantities ; but the result of working with a purer iron in Sweden has been carefully noted for two consecutive weeks, and the loss on the weight of fluid iron tapped from the blast furnace was ascertained to be only 8| per cent. The largest sized apparatus at present erected is that in use at the Atlas Steel Works, Sheffield, as shown in the drawings already described, the converting vessel being capable of converting 4 tons at a time, which it converts into cast steel in 28 minutes. In consequence of the increased size of the converting vessel in this case no metal is thrown out during conversion ; and the loss of weight has fallen as low as 10 per cent., including the loss in melting the pig iron in the reverberatory furnace. Specimens of this manufacture as carried on at the author's works in Sheffield are exhibited, consisting of a piece of the pig iron employed, which is No. 1 hot-blast haematite made with coke ; also a portion of an ingot of very mild cast steel, broken under the hammer to show the purity and soundness of the metal in its cast unhammered state ; and an ingot partly forged to show how little work with the hammer will produce a forging from these solid blooms of steel. There are also two pieces of steel of the quality employed for making piston rods, which have been bent cold under a heavy steam hammer to show the toughness of the metal: it requires very

104

141

BESSEMER STEEL.

much more force to bend it than would be required to bend wrought iron, but notwithstanding this additional rigidity it yields to any extent without snapping. The tensile strength of this soft and easily wrought metal is as much as 40 tons per square inch, or from 15 to 18 tons greater than that of best Yorkshire iron. In turning, planing, boring, and tapping, it will be found that the uniformity of its quality will be less trying to the cutting tools than the hard reeds and sand cracks met with in the common qualities of malleable iron. The above tensile strength of the piston-rod steel however is by no means the maximum, but on the contrary is nearly the minimum strength of the steel converted by this process ; but at the same time it possesses nearly a maximum degree of toughness, for every additional ton in tensile strength obtained by the addition of carbon hardens the steel for working, renders it more difficult to forge, and brings it nearer to that undesirable state when a sudden blow snaps it like a piece of cast iron. The extreme limits of tensile strength of the converted metal are shown in the following tables, which give the results of many trials made at different times at the Royal Arsenal at Woolwich under the superintendence of Colonel Wilmot:—

BESSEMER STEEL. Tensile Strength per square inch. Bessemer Steel.

Various trials.

In the cast unhammered state.

42,780 48,892 57,295 61,667 64,015 72,503 77,808 79,223

63,023 lbs. = 28.13 tons per square inch.

After hammering or rolling.

136,490 145,512 146,676 156,862 158,899 162,970 162,974

152,912 Ibs. = 68.26 tons per square inch.

Mean Tensile Strength.

Lbs.

V

105

142

BESSEMER STEEL.

BESSEMER IRON. Tensile Strength per square inch, Bessemer Iron.

Various trials.

Mean Tensile Strength.

Lbs.

In the cast unhammered state.

38,197 40,234 41,584 42,908 43,290

41,243 lbs. = 18.41 tons per square inch.

After hammering or rolling.

64,059 65,253 75,598 76,195 82,110

72,643 lbs. = 32.43 tons per square inch.

Flat Ingot rolled into Boiler Plate without piling.

63,591 63,688 72,896 73,103

68,319 lbs. = 30.50 tons per square inch.

From these tables it is seen that, after hammering or rolling, the steel or highly carbonised metal exhibits a mean tensile strength of 68 tons per square inch, but from its hardness and unyielding nature it is totally unfit for many purposes ; while the iron or entirely decarbonised metal is so soft and copper-like in its texture as to yield to a mean tensile strain of 32 tons per square inch, a point unnecessarily low except in cases where a metal approaching copper in softness is required. The soft easy-working tough metal of the quality used for piston rods is therefore believed by the author to be the most appropriate material for general purposes, while the hard steels that range up to a tensile strain of 50 or 60 tons per square inch should be avoided as altogether too expensive to work and too dangerous to be employed in any case where sudden strains may be brought upon them. With reference to the employment of the mild cast steel for constructive purposes, there are few applications of more importance than that which has recently and successfully been made to the construction of steam boilers. The Cornish boiler, as improved by Mr. Adamson of Hyde near Manchester, has a large flue tube constructed

106

BESSEMER STEEL.

143

with narrow plates more than 12 feet long, extending round the flue in one length, and flanged at each edge in a manner which, while it adds greatly to the stability of the flue, demands such qualities in the material employed for its manufacture as are completely found only in metal that has undergone fusion and has become perfectly homogeneous throughout. A practical illustration of the excellence of this mode of constructing boilers and the powerful strains which the new steel is capable of sustaining safely is afforded by the steam boilers employed for some time past at Messrs. Platt's works at Oldham, where six of these boilers are in daily use ; they are 30 feet long and 61/2 feet diameter, and the flue is 4 feet diameter; the plates are5/16inch thick, and the working pressure 100 Ibs. per square inch. The advantages of cast steel are still more marked in the construction of the fireboxes of locomotive engines. The difficulty of flanging and shaping this work in plate iron without splitting the metal at some part is so great as to have rendered the employment of copper necessary hitherto for this purpose ; but the shape required can now be obtained with ease and certainty by hammering up a sheet of metal rolled from one of the cast ingots, such as that now exhibited. One of these firebox plates flanged by Mr. Adamson is also shown, and clearly illustrates the facility with which the new metal may under skilful hands be wrought into any required form. The perfect continuity of the material and its entire freedom from joinings or weldings also obviously render it specially suitable for the tube plates of locomotive engines ; for however near the holes are made to one another, there is no danger of their having a flaw or other weak place between them. This is exemplified in the piece of plate now exhibited, in which rivet holes have been punched so close as to remove almost all the metal, without splitting the narrow piece still left between the holes. Nor is it in the construction of the boiler alone that the cast steel may be employed with advantage in locomotives : the axles whether plain or cranked, the piston rods and guide bars, and last but not least the wheel tyres, are all exposed to so much abrasion and to such sudden and powerful strains that a tough strong material capable of withstanding this destructive wear and tear is imperatively

107

144

BESSEMER

STEEL.

demanded for the satisfactory construction and economical working of the engine. The special aim of the author during the first year of his labours, which throughout the last six years has never been lost sight of, was the production of a malleable metal peculiarly suitable for the manufacture of ordnance. By means of the process that has been described solid blocks of malleable cast steel may be made of any required size from 1 to 20 or 30 tons weight, with a degree of rapidity and cheapness previously unknown. The metal can also with the utmost facility be made of any amount of carburation and tensile strength that may be found most desirable : commencing at the top of the scale with a quality of steel that is too hard to bore and too brittle to use for ordnance, it can with ease and certainty be made to pass from that degree of hardness by almost imperceptible gradations downwards towards malleable iron, becoming at every stage of decarbnration more easy to work and more and more tough and pliable, until it becomes at last pure decarbonised iron, possessing a copper-like degree of toughness not found in any iron produced by puddling. Between these extremes of temper the metal most suitable for ordnance must be found ; and all qualities are equally cheap and easy of production. From the practice now acquired in forging cast steel ordnance at the author's works in Sheffield it has been found that the most satisfactory results are obtained with metal of the same soft description as that employed for making piston rods. With this degree of toughness the bursting of the gun becomes almost impossible, its power of resisting a tensile strain being at least 15 tons per square inch greater than that of the best English bar iron. Every gun before leaving the works has a piece cut off the end, which is roughly forged into a bar of 2 inches by 3 inches section, and bent cold under the hammer in order to show the state of the metal after forging. Several test bars cut from the ends of guns recently forged are exhibited. The power of this metal to resist a sudden and powerful strain is well illustrated by the piece of gun muzzle now shown, which is one of several tubular pieces that were subjected to a sudden crushing

108

BESSEMER STEEL.

145

force at the Royal Arsenal, Woolwich, under the direction of Colonel Wilmot; the pieces were laid on the anvil block in a perfectly cold state, and were crushed flat by the falling of the steam hammer, but none of them exhibited any signs of fracture when so tested. Probably the best proof of the power of the metal to resist a sudden violent strain was afforded by some experiments made at Liege by order of the Belgian government, who had one of these guns bored for a12 Ibs. spherical shot of 4f inches diameter, and made so thin as to weigh only 91/4 cwts. This gun was fired with increasing charges of powder and an additional shot after each three discharges, until it reached a maximum of 63/4 Ibs. of powder and eight shots of 12 Ibs. each or 96 Ibs. of shot, the shots being thus equal to about one tenth of the weight of the gun. It stood this heavy charge twice and then gave way at about 40 inches from the muzzle, probably owing to the jamming of the shots. The employment of guns so excessively light and charges so extremely heavy would of course never be attempted in practice. Some idea of the facility of this mode of making cast steel ordnance is afforded by the time occupied in the fabrication of the 18 pounder gun now exhibited, which was made in the author's presence for his experiments on gunnery. The melted pig iron was tapped from the reverberatory furnace at 11.20 A.M., and converted into cast steel in 30 minutes ; the ingot was cast in an iron mould 16 inches square by 4 feet long, and was forged while still hot from the casting operation. By this mode of treating the ingots their central parts are sufficiently soft to receive the full effect of the hammer. At 7 P.M. the forging was completed and the gun ready for the boring mill. The erection of the necessary apparatus for the production of steel by this process, on a scale capable of converting from crude iron enough steel to make forty of such gun blocks per day, will not exceed a cost of £5000, including the blast engine; hence the author cannot but feel that his labours in this direction have been crowned with entire success : the great rapidity of production, the cheapness of the material, and its strength and durability, all adapt it for the construction of every species of ordnance.

109

146

BESSEMER STEEL.

For the practical engineer enough has already been said to show how important is the application of cast steel to constructive purposes, and how this valuable material may be both cast and forged with such facility and at a cost so moderate as to produce by its superior durability and extreme lightness an economy in its use as compared with iron. The construction of cast steel girders and bridges, and of marine engine shafts, cranks, screw propellers, anchors, and railway wheels, are all deserving of careful attention. The manufacturer of cast steel has only to produce at a moderate cost the various qualities of steel required for constructive purposes to ensure its rapid introduction ; for as certainly as the age of iron superseded that of bronze, so will the age of steel succeed that of iron.

110

Figl. Side Elevation of

Cenverting

Casting Ladle, and, Crane.

Vessel,

Fig.2.

Casting the Steel into Ingots.

Fig 5. Front Elevation of Converting Vessel-

fig. 4. Plan of

Converting Vessel

and Moulding Pit.

112

Fig. 6. Blowing. Fig. 5. Fitting.

Pig. 9. Plans of

Tuyeres.

Fig 10. L ongitudinal Section, of a Tuyere.

Fig 11. Plans.

Fig. 8. Vertical Section of Casttrbg Ladle and Ingot Mould. Fig. 7. Pouring out the Steel into Casting Ladle.

Francis William Crossley (1839-1897) Francis (Frank) Crossley was born in Dunmurry, Co Antrim. He came to England at the age of eighteen and became an apprentice in Robert Stephenson's works in Newcastle-upon-Tyne. His younger brother, William (1844-1911), became an apprentice with Sir W G Armstrong and Company at Elswick. Both brothers worked closely together, and formed Crossley Brothers in Manchester in 1867. The Crossley brothers originally manufactured a range of equipment associated with the Manchester cotton industry. They also developed and patented machinery for the rubber and flax industries. However, the emphasis of the company changed irrevocably when William, on his honeymoon, saw a demonstration of Otto's engine in Cologne. Frank saw an Otto-Langen engine in 1866, and was convinced of its potential. The engine was demonstrated in the Paris Exhibition of 1867, where it was a centre of attraction. Crossley Brothers acquired the patents for OttoLangen engines world-wide, with the exception of Germany, and entered into agreements to manufacture the engines in Britain in 1869. The company built 1400 engines, and made significant improvements to the design. In 1876 the Crossley Brothers company was offered the sole rights for the Otto "silent" four stroke cycle engine, which they took up. Frank made many improvements to the Otto engine, particularly in the areas of valves and ignition systems. The engines were a tremendous success for low power applications and in this size range quickly supplanted the steam engine. Frank was the technical expert of the company and had the duty of defending the patents in court against any infringements. This patent was defended very robustly and successfully, and was applied not only to the four stroke cycle but to any compression engine. A measure of the strength of the patent was that the two stroke engine of Dugald Clerk was also a casualty of an infringement threat. In 1882 the company moved from Great Marlborough Street, Manchester, to new premises in Pottery Lane, Openshaw, which it occupies to this day. In 1897, on the death of Frank, the company became a public limited company with William as the head. The Crossley brothers introduced the internal combustion engine into Britain and made a major impact on its development. The company went on to manufacture both cars and heavier vehicles, and later built engines for aeroplanes.

115

Plate from Crossley Bros, advertisement.

191

ON OTTO AND LANGEN'S ATMOSPHERIC GAS ENGINE, AND SOME OTHER GAS ENGINES. BY MR. FRANCIS W. CROSSLEY, OF MANCHESTER. The Otto and Langen Atmospheric Gas Engine is not only an exceedingly ingenious, but also a well tried and now largely used machine. The attempt to obtain a steady rotary motion from a series of either regularly or irregularly fired explosions—irregular in point of time—and to do this also in accordance with scientific principle, is no easy task ; but here it has been accomplished. The main characteristic of the engine is the "free piston"; the piston when impelled by the explosion rises freely upwards in the vertical cylinder, without at that moment actuating the machine, the motive power being obtained indirectly during the descent by atmospheric pressure acting on the upper side of the piston, in consequence of there being a partial vacuum below it, following the explosion. In the other type of gas engines, on the contrary, the force of the explosion is employed as the motive power, acting direct on the piston. Upon this principle were the two principal gas engines introduced in this country a couple of years previously to 1868, when the Otto and Langen engine appeared. They were both partially successful at the outset, but owing to numerous defects they never came into much use in this country. They were both very similar in appearance to the ordinary horizontal steam engine, and the principal difference between them was that the explosive charge was fired in the one engine by electricity and in the other by an arrangement of gas lights. A further difference was added when a jet or spray of water was admitted into the cylinder of the latter, which was evaporated there by the heat generated by the explosion, and somewhat aided in propelling the piston.

117

192

ATMOSPHERIC GAS ENGINE.

The principle of both these gas engines however was entirely wrong in the writer's opinion, for the following reason. In both of them the explosion delivers its force upon a piston connected to a crank and flywheel, exactly as is done by steam in a steam engine; but therein lies the evil. Steam gives a steady and sustained thrust against the piston, and the gradual motion of the piston is suited to this; but it is quite otherwise with an explosion, as in that case the stress is intense but instantaneous only. The effect of delivering this sudden blow against a piston, connected rigidly with a heavy flywheel, is simply that, instead of the heat, set at liberty by the union of the oxygen and hydrogen in the explosion, being converted into mechanical motion, it remains in the form of heat, and has to be got rid of by a very large external supply of cold water, lest it should destroy the surfaces of the cylinder and piston, and even lead—as it has often done—to the buckling of the piston-rod when it has grown red-hot. In consequence the common steam-engine pistons of these engines, with their connecting-rods and cranks, will not, under any circumstances conceivable by the writer, enable them economically to utilise the suddenly generated and suddenly expiring force of an explosion. The blow given to the piston by the explosion is received by the heavy mass of the necessarily heavy flywheel, which cannot rapidly yield to it; and just as when a cannon ball strikes a massive target which it cannot carry along with it, a flash of fire is the result in which the energy of the shot disappears, so in these engines heat instead of motion is the result of releasing the stored energies of the gases, and in this case heat is not what is wanted. The flame of carburetted hydrogen, when the combustion is perfect, is intensely hot; and when repeated discharges take place— say 150 a minute, as in the case of these horizontal engines— it is easy to see how much cold water must be circulated through the jacket of the cylinder in order to keep the temperature down to something below that at which oil oxidises, so as to prevent the destruction of the piston. It is possible however to keep the temperature sufficiently low even in these engines by supplying

118

ATMOSPHERIC GAS ENGINE.

193

water enough; but it takes a great deal, and where constant working is required it often adds much to the cost of running, and all the heat taken up by the water is carried off without doing work. There is yet another element of difficulty connected with these engines, and that is the deposit left by the gas after explosion. Their cylinders are very much smaller, say less than one fourth the volume of the Otto and Langen engine for the same power. This is in their favour in cost of construction, but it is a drawback in working. The difference in consumption of gas for the same power is found to be about as 1 to 6 in favour of the Otto and Langen engine, and as the cylinder in which this is burnt is also about 4 to 1 in volume, there is twentyfour times the space per unit of gas consumed in the Otto and Langen engine relatively to the others. There is consequently less liability to clog from deposit; and in order to become equally dirty the engine should require twenty-four times as many hours' work. There have been however some favourable reports of one of the horizontal engines in this particular, and perhaps the quality of the gas may often be such as to cause no difficulty from deposit even in them: though when the gas is dirty it is obvious the Otto and Langen engine has the advantage. Some of the disadvantages accruing from using a common steam engine as a gas engine have now been shown ; and this is practically what is done in the horizontal engines, excepting as regards the arrangements made for the firing of the charge and for the supply of the gas and air, in place of the valve for the passage of the steam. It has been pointed out that great waste of fuel, great generation of heat as a necessary consequence, a large supply of cold water as a further necessary consequence, and sometimes if not often a heavy deposit of carbon and tar in the cylinder and passages, are the results obtained from this principle of gas engine. Now in the Otto and Langen engine, the idea of a "free piston" involves great constructive difficulty. The engine is really a gun, which stands vertically, with open mouth pointing upwards; the

119

194

ATMOSPHERIC GAS ENGINE.

explosive compound of gas and air takes the place of the powder, and the piston represents the shot. The charge measured off is however not sufficient to drive the shot or piston out of the gun, and only to within an inch or two of its mouth. The engine is shown in Figs. 1 to 3, Plates 23 to 25, which represent a1/2H. P. engine. It is single-acting, the upper side of the piston B being continuously exposed to the pressure of the atmosphere through the open mouth of the cylinder C; and this arrangement greatly aids in keeping the cylinder cool. The pistonrod A is a rack, and it gears into a toothed wheel D on the main shaft E of the engine, which is mounted on the top of the cylinder. The length of the rack is about equal to twice the circumference of the wheel, so that the single-acting effect of the engine is very different from what it would be in a steam engine. The toothed wheel D however is not keyed fast upon the shaft E, but is attached to it by a friction clutch, which permits the rack to rise without moving the shaft at all, and connects them in the downstroke only. Thus the shock of the explosion is not sustained by the shaft, and the piston is able to move freely in the upstroke independently of it, being arrested only by the resistance of the atmosphere at the end of the upstroke. The following is the series of operations in each stroke, commencing with the piston at the bottom, as shown in Fig. 1. First the piston is lifted through a space of about 1-11th of the length of stroke, as shown in Fig. 2, in order to draw in the charge of gas and air; and the power to effect this movement is obtained from the momentum of the flywheel. The charge is then fired by contact with a gas light, and the piston flies up freely to the top. As it ascends, the plenum caused by the explosion is changed to a partial vacuum, which reaches about 22 in. of mercury at the top of the stroke, and thus the motion of the piston is quickly reversed, and the downstroke is performed under a pressure of about 11 Ib. per sq. in. derived from the atmosphere; this driving power is communicated through the rack and toothed wheel to the shaft. When the piston has reached within a short distance of the bottom, the vacuum, which has been gradually decreasing, is

120

ATMOSPHERIC GAS ENGINE.

195

again changed to a plenum, and the weight of the piston and rack expels the burnt gases during the last few inches of the stroke, thus completing the cycle of operations. The friction clutch, used to connect and disconnect the piston-rod rack with the driving shaft, is shown in Figs. 6 and 7, Plate 26. It consists of a pulley G keyed on the shaft E and surrounded by a ring D, on the interior of which are cut three inclined surfaces. On each of these inclined surfaces a set of live rollers is free to travel, and to press against a corresponding curved wedge I while the piston is descending; and at the same time the opposite side of the wedge, which is faced with leather, presses against the pulley G. While the piston is ascending, no pressure is put on these wedges by the ring, and hence in the upstroke the ring can freely revolve backwards upon its bearings on the shaft; but as soon as the downstroke commences, a firm hold of the shaft is immediately gained by the ring, and the shaft thereby becomes connected direct with the toothed wheel which gears with the piston-rod rack. As by this means the piston is free to take any length of stroke within the limit of the cylinder, it is impossible to determine exactly at what moment it shall reach the bottom ; and as the apparatus for lifting it to draw in the next charge is not required to move until the return of the piston, an intermittent motion is provided for lifting the piston, which is started at the right moment by a tappet fixed to the rack. The mechanism consists of a pair of eccentrics H and K, Figs. 1 and 2, one of which H moves a lever L, which lifts the piston by means of a tappet M projecting on the side of the rack A; and the other eccentric K moves the slide-valve N of the engine. These eccentrics, which are made fast to each other, are carried loose upon an independent shaft T driven by spur wheels from the main shaft E of the engine; and they are started and stopped by an arrangement of ratchet-wheel R and catch P. The catch or paul P is carried by the eccentric H, and is made with a projecting tail opposite to the hook; and a stop S is arranged to strike the tail of

121

196

ATMOSPHERIC GAS ENGINE.

the paul at a fixed point of its revolution, thus arresting it by throwing it out of gear with the ratchet-wheel B, which is keyed upon the shaft T. When the paul is in gear with the ratchet-wheel, it carries the eccentrics round with it; and when out of gear, all stop together. The stop S for disconnecting the paul is held up in position by a spring; but when the rack A descends to the bottom, the tappet M on the rack depresses the stop out of the way, and allows the paul to fall into gear with the ratchet-wheel. A revolution of the eccentrics now takes place, and when this is completed the stop arrests them until the next descent of the piston again effects their release. But while the piston is lifted to draw in the charge, it is necessary to give the valve the power both of admitting the explosive mixture of gas and air, and also of applying the light to fire it when received by the cylinder. The valve for this purpose consists of a flat plate N, which moves between two faces attached to the cylinder base, the outer face being a moveable plate kept up to its place by springs. The valve is provided with ports, adjusted for proportioning the gas to the air so as to form the required compound; and it is further made with a small chamber, Fig. 2, having an opening both to the inside and to the outside; and an independent gas-pipe is carried into this chamber, so that a light may burn in it. Whilst the valve is at rest during the downstroke of the piston, the opening on the outside of the valve is exposed to the atmosphere, and close to it burns constantly a small gas-light, by which the gas fed into the chamber in the valve is ignited ; but on motion being given to the valve, the communication of the chamber with the outside is cut off and is opened with the inside of the cylinder, and the flame that remains in the chamber explodes the charge. This is not effected until the gas and air supply are also cut off by the same movement of the valve. The release of the exhaust or burnt gases is effected by the valve being provided with an independent exhaust port, which is open while the valve is at rest awaiting the descent of the piston, and so permits the escape to take place as soon as the vacuum has changed to a plenum.

122

ATMOSPHERIC GAS ENGINE.

197

It is necessary however to provide a clack valve V in the exhaust pipe, closing inwards, as shown in Fig. 3, Plate 25, otherwise the atmosphere would enter the cylinder and destroy the vacuum, the exhaust port being open during the entire downstroke; and this leads to an interesting point in the action of the engine. It is clear that another explosion cannot take place until the piston has arrived at the bottom and given motion to the eccentrics, and so shifted the valve; but the piston cannot get to the bottom until the escape of the burnt gases is effected, and therefore by simply preventing this escape the interval between the explosions may be indefinitely prolonged, and thus the power and speed of the engine is controlled. This is done by arranging a common governor to press upon the clack valve V in the exhaust pipe, as shown in Fig. 3, and so delay the escape of the gases. This is one of the most important features of the engine, and is the invention of Mr. Otto. To take an example of its effect, suppose an engine is employed in hoisting, and that the load demands the exertion of the full power to accomplish the work; in this case the governor will not press upon the exhaust valve at all, and explosions will take place as rapidly as possible—say at the rate of 30 per minute—whilst the load is being raised. Now suppose the load gets to the top, and all the work is suddenly thrown off, the effect of this will be to increase the speed of the engine sufficiently for causing the governor to close the exhaust valve; and until this is re-opened by the speed dropping, the engine will not make another stroke or explosion. In some of the best of these engines there will be a pause of one minute before this takes place; or in other words only l-30th of the power of the engine is required to move itself whilst doing no work. Perfect as the above method of governing the engine is in principle, it was found in practice to present a drawback. When the piston became leaky through wear, the exhaust was not obliged to escape only through the clack valve V; and thus by going out another way, evaded the action of the governor, and so " running away" of the engine was possible. The necessity for something G2

123

198

ATMOSPHERIC GAS ENGINE.

better has now led to a great improvement in this detail, as shown in Figs. 4 and 5, Plate 26. Instead of preventing the piston from descending by closing the clack valve V, it is now allowed to descend, but its descent is not allowed to release the paul that puts the eccentrics in gear and moves the slide-valve N, unless the governor also permits it. The governor now controls a second stop, which keeps the paul from falling into gear with the ratchetwheel and moving the eccentrics, not only until after the piston has descended to the bottom, but until the speed of the engine has dropped to the desired limit. There are points of comparison between this gas engine and a steam engine in which the latter has the advantage. A steam engine does not necessarily make a noise in working; and the noise has hitherto proved to some extent an insurmountable difficulty with these gas engines, on account of the rapid and intermittent character of their movements. But on the other hand it is a very beautiful and advantageous feature in these gas engines that the governor is able, as described, to stop all motion of the parts, except the flywheel and shaft, as soon as the work is thrown off or less than full work is required. Instead of the piston continuing to rush to and fro while no useful duty is being done, as is the case with steam, it is here at perfect rest; and there is consequently economy both in fuel and wear and tear, as compared with steam. The gas engine has other advantages, in the power of starting at a moment's notice, and starting too at full power: in the fact that while the engine is standing no fuel at all is burnt: and in the very trifling attendance required, which is very much less than with steam. There is no trouble with coals and ashes, nor in many cases is any water consumed. The engine not having any boiler, no boiler explosion can take place; and thus insurances are not affected by its use, or are only very rarely affected. Lastly, the economy of fuel is perhaps the most important difference. Where gas can be had at low rates, the 1 II. P. engine will often run for 2s. 6d. or 3,9. per week; it has not been known to exceed an average cost of Id. per H. P. per hour, with gas at 4s.

124

ATMOSPHERIC GAS ENGINE.

199

per 1000 cub. ft., and it is generally much below this consumption, though no doubt not exerting its full power continuously. Though 1d. cost of gas per H. P. per hour will not appear an economical consumption when compared with the value of say 2 Ib. of coal per H. P. per hour, it has to be borne in mind that this comparison is made between the best and largest steam engines and the gas engine; whereas its practical opponents are the smallest and least perfect ones, in which the fuel burnt will probably be 15 or even 30 Ib. of coal per indicated H. P. per hour. Also gas compared with coal is of course greatly more expensive in cost per heat unit. Taking all into consideration however, the commercial side of the question is entirely in favour of the gas engine in the matter of economy of fuel. When the fuel lost in raising steam and while a steam engine is standing, and the cost while running, are all taken into account, it is found that the gas engines have sometimes saved upwards of 15s. or even 20s. per week in fuel alone where they have replaced small steam engines, besides the saving in attendance, which is often nearly as much more. But the most interesting question is the consideration of the engine from a purely scientific point of view. How many foot-lbs. are obtained on the break per heat unit supplied by the fuel? Here the best steam engines are surpassed by this gas engine. Were pure hydrogen the fuel, instead of adulterated coal gas, no less than 2-5ths of the theoretical efficiency of the fuel might be realised on the break; coal gas however is less productive. Taking the proportion of coal gas to air for complete combustion as 1 vol. of gas to61/2vols. of air, and the theoretical efficiency of the gas as equal to 24,000 heat units per Ib. weight consumed, one heat unit being equal to 772 ft.-lbs.,—and taking the density of the gas as 40 per cent. that of air, or 1 cub. ft. of gas = 0.03 lb. weight,—then the heat units supplied to the engine per minute are equivalent to 584,000 ft.-lbs., the consumption of gas being 1.05 cub. ft. per min.; and the return for this on the break is about 70,000 ft.-lbs., or 12 per cent. Now to compare this with the best steam engine, allowing for air pumps and feed pumps and friction, is it not even too much to

125

200

ATMOSPHERIC GAS ENGINE.

say that 2 Ib. of coal will give under the most favourable conditions of a trial trip 1 H. P. per hour on the break ? But taking it at that figure, and taking 1 Ib. of coal to supply 15,224 heat units or 11,752,928 ft.-lbs., then = 990,000ft.-lbs., or only81/2per cent, of the theoretical efficiency of the fuel: while the gas engine realises 12 per cent., or nearly 1| times the amount. Some careful indicator diagrams lately taken from the gas engine, one of which is shown in Fig. 8, Plate 27, fully illustrate the fact that it is impossible to get a good result from the sudden force of an explosion except by some such means as a free piston. The very sudden and extreme rise in pressure at the moment of explosion is due simply to the expansion of the gases under the temperature of the flame. If this temperature be taken at 5000° Fahr., and divided by 520 for the rate of expansion from an initial temperature of about 60°, it gives an expansion of about 10 times; and as the gas compound occupied 1-11th of the cylinder at the moment of ignition, if it expands 10 times it gives very nearly the stroke actually taken by the piston. The 5000° is an assumption only, but seems to be confirmed by the amount of expansion which follows it. After the explosion the temperature falls almost instantaneously, as shown by the sudden drop of pressure in the diagram. The driving pressure in the return or working stroke varies from 11 Ib. per sq. in. at first, to nothing at about 4-5ths of the stroke, averaging 9 Ib. during the time of action, or a mean of about 7 Ib. per sq. in. effective pressure throughout the stroke. The steady and long sustained pressure during the return stroke, as shown by the diagram, is a proof of the continued fall in temperature even during the contraction of the gases, and is a testimony to the efficiency of this remarkable form of engine. Its spasmodic and intermittent movements are not what can at first sight prove palatable to engineering taste; but experience of its working shows that its faults lie almost wholly on the surface, consisting chiefly of noise, while its advantages are deeper seated, and make it in its very limited sphere a formidable rival of steam.

126

ATMOSPHERIC GAS ENGINE.

201

The second indicator diagram, shown in Fig. 9, Plate 27, is taken from one of the other type of gas engines, using the explosive force of the gas as the direct driving power, and shows the inefficiency of that mode of employing gas. Nearly half the length of stroke is occupied in this engine by drawing in the charge of mixed gas and air, which is then exploded, and the pressure caused by' the explosion propels the piston through the remainder of the stroke; the return stroke is an open exhaust into the atmosphere. The driving pressure indicated varies from 57 Ib. per sq. in. at first, to nothing at the end of the stroke, averaging131/2Ib. per sq. in. during the time of action, or a mean of about 6 Ib. per sq. in. effective pressure throughout the stroke, which singularly coincides with the mean effective pressure in the Otto and Langen engine. In the horizontal engine 146 explosions, each burning about as much gas as one explosion in the 1 H. P. Otto and Langen engine, result in no better effect in driving power on the shaft than about 30 explosions in the latter engine.

Mr. E. A. COWPER said he had had an opportunity of seeing this gas engine at work; and in reference to the application of the explosion of gas for the production of power, he remembered an attempt was made many years ago to produce power, or rather to raise water, by the explosion of gas in an experimental engine, known as Brown's gas vacuum engine, in which the gas and air were mixed in a vessel having a loose cover; the cover was blown up by the explosion and then fell back into its place, and the vacuum produced was made to suck up water from a reservoir below, into the vessel, from which it was afterwards run out for use. That was an attempt to do away with the evils of the tar and dirt which resulted from the explosions of the gas in a closed vessel. A few explosions did not dirty a cylinder very much, but after working for some time there was an accumulation of condensed sulphurous acid and other substances, including tar, and these after some time

127

202

ATMOSPHERIC GAS ENGINE.

generally caused the leakage of the piston, which was so exceedingly objectionable in an engine upon that principle. In the engine now described this difficulty seemed to have been got over to a considerable extent; but when the cylinder became worn, there would be an escape past the piston at the moment of the explosion. He had hoped that this leakage past the piston had been entirely overcome; he understood it had been overcome to a certain extent, and he should be glad to know how far. This plan of gas engine had been spoken of as applicable only to produce a small amount of power. Now it was much more easy to work a small engine than a large one on such a plan, as the quantity of heat developed by the explosion of the gas was much greater in a large engine than in a small one. He should be glad to hear of further attempts and success with this engine constructed in a larger size; and to know what was the largest size yet made or in contemplation. In reference to the mode of utilising the explosion of the gas, a larger space was evidently required in this engine for obtaining the power by means of the weight lifted by the explosion, than if the exploded gas were confined in a smaller space so as to drive by direct pressure. The exploded mixture of gas and air produced at the moment of explosion a pressure of about 150 Ib. per sq. in. under favourable circumstances, but it was difficult in any engine to take full advantage of a sudden pressure of that amount. If it were possible to cut off the exploded gas neatly in the cylinder, and expand it thoroughly, a good indicator diagram would be obtained; but one of the chief difficulties was to explode the gas neatly and properly, and in order to produce a good diagram the stroke of the piston would have to be inconveniently long; besides which there were certain inconveniences connected with the application of the pressure in an efficient manner as a direct driving pressure. But with the vacuum, principle the arrangement was greatly simplified; the piston made its upstroke freely, like a shot, and the power was obtained by the vacuum in the downstroke. He wished that more power could be obtained by that principle in a given space, so

128

ATMOSPHERIC GAS ENGINE,

203

that the engine might be made more powerful and of larger size. The pressure of 7 Ib. per sq. in. in the downstroke alone was equivalent to only31/2Ib. in the double stroke, which was rather a small amount of pressure for a cylinder of the size shown; and he should be glad if an increased pressure could be obtained, so as to get more power from the same size of engine. Great credit was due for the very ingenious construction of the driving clutch, which he had no doubt was efficient in utilising the whole length of the stroke of the piston of a small engine, as long as it was kept in order; and the engine was altogether a highly ingenious arrangement. Mr. A. PAGET considered the main principle of the engine was admirable for utilising the full effect of an explosion. The earlier gas engines had been as great anomalies as would be the use of a sledge-hammer with a cushion interposed between itself and the work on which it was employed; for in those engines the piston, piston-rod, connecting-rod, crank, and shaft, had each formed a cushion, which altogether absorbed and threw away the greater portion of the force of the explosion. The proper plan for utilising the power of an explosion was to store it up by some spring arrangement and give it out gradually; and in the engine now described this had been well carried out, so far as it went. He remembered seeing this engine in Paris in 1867, and its present appearance was certainly very similar to what it bore then, though some of the details had been modified since it was first brought out; but there were others which he thought still seemed capable of being modified with advantage. He enquired what was the width of the driving clutch, and whether there was not a very considerable amount of wear upon the leather of the clutch; and also whether the leather as an elastic material interposed in the clutch might not cause a large amount of back-lash or slip before the clutch got a complete hold, and thus account for what was spoken of as a great annoyance,—the noise : and he enquired whether the amount of back-lash was not greater in this clutch than in many other forms of clutch in constant use.

129

ATMOSPHERIC Fig. 1.

130

Fartical

Section.

GAS ENGINE.

Plate.23.

Piston at bottom of stroke.

ATMOSPHERIC GAS ENGINE. Fig.

2.

Plate

24.

Vertical Section.

131

ATMOSPHERIC Fig. 3.

132

Side

GAS

Elevation.

ENGINE.

Plate 25.

ATMOSPHERIC

Fig.

4.

GAS ENGINE. Fig.

Fig.

Plate 26. 5.

6. Transverse Section.

Fig. 7. Sectional Plan .

133

ATMOSPHERIC Fig. 8.

Fig. 9.

134

Indicator

GAS

Diagram,

Indicator Diagram acting by direct

Otto

from force of

ENGINE. and

Plate

Langen Engine.

Horizontal Gas Engine explosion.

27.

Thomas Russell Crampton (1816-1888) T R Crampton spent his post-apprenticeship years working with some of the best-known Victorian engineers. He was personal assistant to Marc Brunei, of Thames Tunnel fame, during the period 1839-1844. Later, under the direction of Daniel Gooch, he prepared drawings for the first locomotive built for the Great Western Railway. Crampton was rated as "a clever fellow" by Gooch. However, his locomotive designs, awarded the highest class medal at the Great Exhibition above Gooch, were dismissed as "most faulty in construction". Crampton's earliest papers for the Institution's Proceedings were all on railway matters following his election in 1847. His own designs typically incorporated such features as a long boilers, outside cylinders, and drive wheels set behind the fire-box. Although never accepted on British railways, such engines found favour on the Continent, particularly in France. Crampton's lack of orthodoxy probably meant that he was wellsuited for his eventual working life as a consulting engineer. His reputation was enhanced by a second achievement completed shortly before the closing of the Great Exhibition. This was the laying of the first practical submarine cable between Dover and Calais. He not only supervised the work but raised most of the capital, taking a substantial risk. As a consequence, Crampton has been described as "the father of submarine telegraphy". The remainder of a varied career included constructing the Berlin Water Works and Broadstairs Gasworks. Crampton never lost his interest in railway development, laying out lines as far away as Russia and the Near East, but he often worked in South-East England. Several of the tracks that merged to form the London Chatham and Dover Railway were the result of his labours. As to the subject of this Proceedings paper, one can view his visionary interest in physically linking Britain and France by channel tunnel as the natural outcome of his personal history. From Brunei's tunnelling, to telegraph contact, and his efforts on French and British railways, Crampton's involvement appears to be particularly apt.

135

440

AUGUST 1882.

ON AN AUTOMATIC HYDRAULIC SYSTEM FOR EXCAVATING THE CHANNEL TUNNEL. ![Lecture delivered by Trios. R. CRAMPTOX, Esq., Member of Council, at the Conversazione at Leeds, IGth August, 1882, and published by order of the Council.]

The question of constructing tunnels of considerable length has of late been brought more prominently into public notice by the successful completion of the tunnels through the Mont Cenis and the St. Gothard. In both cases the nature of the rock through which the tunnel had to be cut was such that the rate of advance was comparatively slow, and the quantity of debris which had to be removed daily was inconsiderable; consequently the traffic by trucks or wagons, which the removal of such debris necessitated, offered no serious difficulties, and did not to any great extent interfere with the operation of lining such portions of the tunnel as had already been excavated. But when the material through which the tunnel has to be cut is of so soft a nature that blasting operations can be dispensed with, and ordinary appliances, such as boring or cutting tools, can be employed, then the rate of advance may become practically unlimited —being, in fact, controlled mainly by the greater or lesser facility with which the debris from the face can be removed from the tunnel. So far as I am aware, there is not now in practical use any other mode of dcaling with the debris of tunnel perforation, than that of transport by means of trucks and wagons running on rails, as employed in coal mines. The system of removal which I propose to use, and which will be more particularly treated in this lecture, is applicable only in the case of tunnels through chalk or similar materials; and for the purpose of illustration, I may be allowed to connect it for the moment with one of the most prominent engineering problems of the day, namely the proposed submarine tunnel between England and France.

137

AUGUST 1882.

CHANNEL TUNNEL.

441

This is supposed to pass for its whole length through a uniform stratum of grey chalk without water. I eliminate therefore all questions of meeting with hard rock, or with any large quantities of water during construction. Such questions will have to be dealt with in any case as they arise, and by ordinary means. The tunnel is assumed to be 20 miles long, independent of approaches on either side; to be excavated 36 feet in diameter in one operation, which, with an internal lining of 3 feet all round, will leave a clear tunnel 30 feet in diameter; and the work to be commenced simultaneously at both ends. It follows therefore, since the approaches may be made at the same time as the main tunnel, that we need only consider here a length of 10 miles of excavation worked from one face. Practical trials made with machines in chalk many years since, established the fact that a rate of advance may be easily maintained of one yard per hour, or 24 yards per day; at which rate the work of excavating 10 miles of tunnel would take 21/2years to accomplish, taking the year at 300 working days. With the simple apparatus now exhibited, 12 in. in diameter, as much as 5 yards forward per hour has been cut, The advance of one yard forward per hour in a 36-foot tunnel will necessitate the removal of 113 cubic yards of chalk per hour. In order to ensure the due performance of the necessary work, I will add 50 per cent, to the figures here given, and shall henceforth deal with other items in the same proportion. We have to provide then for the removal of 170 cubic yards of debris per hour, equal in weight to 250 tons—a greater quantity than is lifted in two of our greatest collieries together, in the same time. If we now assume the use of the ordinary system of removal by trucks, we find that the transport of 170 cubic yards of debris per hour would require the passage of 85 trucks of two cubic yards each per hour, or one truck every 42 seconds: thus, if trains are made uo of ten wagons each, there will be one such train every seven minutes passing out, and a similar train of empties coming back continuously. Of course heavier or lighter trains may be used as thought most desirable.

138

442

CHANNEL TUNNEL.

AUGUST 1882,

Arrangements must then be made for raising these wagons to the surface up the vertical shaft, 450 feet in height; and this means the lifting of 255 tons per hour, or 6000 tons per 24 hours, independent of the weight of wagons, men, tools, stores, &c. The wagons may of course be drawn up inclined approaches, but I assume they will be lifted. The lining of the tunnel three feet all round requires materials to the amount of 34 cubic yards per hour, or, with 50 per cent, added for contingencies, 50 cubic yards per hour. These materials will, of course, be introduced by the empties going back to the working face of the tunnel, and will be discharged where required at different parts of the tunnel. It will be easy, without entering further into details, to understand the difficulties to be overcome in accomplishing a task of such magnitude as the lining of the tunnel, at a rate keeping pace with the advance of the boring machinery, while so great a traffic of debris, &c., in trucks is going on in both directions almost uninterruptedly. The system by the use of which I would propose to obviate some of these difficulties eliminates four-fifths of this movement in trucks. It is founded upon the employment of hydraulic power for driving all the machinery required to cut down the chalk at the tunnel-face, and to remove the debris out of the tunnel to the surface, or to any place where it can be disposed of. For the moment the question of different levels in the tunnel will be disregarded, and I will treat the question on the assumption that the whole length of 10 miles is executed on a level 400 feet below the level of the sea ; and further, that the arrangements are such as if the work of excavation and the removal of the debris had all to be carried on at a distance of 10 miles from the starting point. Near the mouth of the upright shaft, A, Fig. 1, Plate 82, powerful machinery will be erected to pump water from the sea, to compress it, and to hold it under pressure, by means of force pumps and accumulators. The water will be compressed at the top to 512 Ibs. per sq. in., and the fall through 400 feet from the sea-level will add another 188 Ibs. per sq. in.—producing thus at the bottom of the shaft 700 Ibs. per sq. in., a pressure commonly employed.

139

.AUGUST 1882.

CHANNEL TUNNEL.

443

The inlet pipe B, which conducts the water under high pressure down the shaft and to the face, will be lengthened as required by the advance of the boring machinery; and this operation will be facilitated by the interposition between the end of the inlet pipe and the boring machinery of an ordinary telescopic joint C, with a free run of say 72 feet, or 24 yards, whereby only one stoppage will become necessary in 24 hours. The cutting machinery at the face, Figs. 2-5, Plates 83 and 84, will be driven by an ordinary hydraulic motor D, direct and without the intervention of gearing. The debris of the chalk cut down will be taken up by a series of cups, and thrown into a shoot E, to the top of which the waste water from the hydraulic motors is conducted. The water flowing down carries with it the chalk debris, and both pass into an ordinary cylindrical revolving drum F, where it is reduced to sludge. The quantity of water used by the hydraulic motors will be so calculated that it will amount to about three times the quantity of chalk debris by weight. When mixed with the water in the revolving drum, the very fine debris almost instantly dissolves, and the result is a cream or sludge, which, passing through small apertures on the face of the drum, is taken up by ordinary pumps G, worked by hydraulic motors H, and forced through the main outlet pipe J, Fig. 1, to the bottom of the shaft A, or direct up the shaft to the sea if required. The pumps G are placed upon the main frame of the boring machine, and the motors H are driven by high-pressure water, taken from the main inlet pipe B. The main outlet pipe J is provided with a telescopic joint similar to the one described for the main inlet pipe. These two telescopic lengths, being attached to one another, are moved forward together. Between the telescope and main pipe two valves UV are placed, one 011 the telescope, the other on the main pipe. When the ram is run out, these valves are closed, and the joint I between the valves broken; the telescope is then moved forwards on the ram by opening a small cock K, connecting the ram with a vacuum chamber. The water in the ram will then flow into the vacuum chamber, and the telescope will be drawn over the ram close up to the machine by the air-pressure behind it. The space thus left between the valves UV is filled up by an extra length of pipe, with a valve attached : the small 2 N

140

444

CHANNEL TUNNEL.

AUGUST 1882,

cock K is shut, the valves U and V opened, and the working resumed. When the machine is required to be brought back, e.g. to change the cutters, the rack and pinion will be thrown out of gear, and an ordinary hydraulic ram (not shown) will be used for the purpose. The cream is forced by the pumps through the excavated portion of the tunnel to the bottom of the shaft, and thence may be raised by pumps or other suitable means to the top, and discharged into the sea. or disposed of in other ways. It will now be perceived that the space lying between the boring machinery and the shaft is left entirely free, excepting the small portion of it occupied by the two pipes—the pressure-water inlet pipe and the cream outlet pipe. The operation of lining the tunnel may therefore be carried on with the greatest facility, there being notraffic upon the rails, and no hoisting up or lowering in the shaft, except that necessary to transport the workmen and the building materials for lining the tunnel, amounting to only one-fifth of that required on the ordinary system: in other words four-fifths of the whole weight to be disposed of is carried through pipes, instead of by locomotives and trucks. A few details regarding the power required for the various operations, the size of inlet and outlet pipes, and other data in connection with the hydraulic machinery, may be interesting. Cutting the Clialk.—A Cutting Machine of the most simple construction will be used, for the purpose of excavating the chalk. As shown in Pigs. 2-6, Plates 83 and 84, designed for an 8-ft. heading, it consists of a number of small discs L, placed at an angle, and attached to a large boring head M, made to revolve at any given speed. Each cutting disc removes at each revolution a concentric ring from the face of the chalk, of 1/16 in. or any other required thickness, and of a width equal to one-fourth of the diameter of the disc ; and the discs are arranged to follow each other in such a manner that while they are all continuously in action, yet none has ever more to remove than its apportioned width and thickness of cut. The disc-spindles, Fig. 6, Plate 84, turn freely in their sockets, and as the discs cut only a width about one-quarter of their diameter,

141

AUGUST 1882.

CHANNEL TUNNEL.

445

they turn in an opposite direction to that in which the large disc or head M revolves, and act by rolling into the chalk. The cutting edge is thus changed continually, whereby the wear and tear of the edges is reduced to a minimum; and at the same time they do not require sharpening, a most material feature. I have cut with a similar machine at the rate of 5 yards forward per hour, without injury to the cutters. By trials I have ascertained that 2 HP. per cubic yard of chalk excavated per hour would be more than ample power ; hence if, in piercing a tunnel 36 feet in diameter, 170 cubic yards of chalk per hour have to be cut down, 340 HP. will be necessary for this part of the work. The pressure of the incoming water upon the area of the telescopic joint on the one hand, and the back pressure of the cream, forced towards the exit, on the other, will always push the machine forward automatically; and it becomes necessary to provide an arrangement to control this speed, and allow the machine to advance only at the desired speed. There are various simple means of effecting this object: that shown is by screw gear and pinion, working into a rack. To cut a clear face 36 feet in diameter will require 72 12-inch cutting discs fixed upon the arms or cross-beams, each cutter in one revolution of the machine taking off a concentric ring 3 inches in width, and1/16of an inch thick. Supposing the head M to turn at the rate of 10 revolutions per minute, this would give the outside cutter a periphery speed of 1130 feet per minute, which has been found to be well within practical limits. It will be understood that the cutters turn at different speeds, those near the outer periphery doing considerably more work than those near the centre. With chalk the revolving cutters are found to be equally effective at all speeds. The sleepers on which the machine slides or runs, and the rack by which its advance is regulated, are made in pieces of convenient length ; and as the last piece of the set is left behind, it is continually shifted forward and put down in front. This is effected (see the cross-section, Pig. 5, Plate 84) by means of two arms N N, Figs. 3 and 5, mounted in a line with the main shaft in rear of the machine. These arms are lowered, and bolted to the cross-sleeper P, which has attached 2 N2

142

446

CHANNEL TUNNEL.

AUGUST 1882.

to it lengths of the rails R and rack S. By means of a small hydraulic engine or other means the arms with the sleeper are made to revolve and lifted into the higher position, where they deposit the sleeper &c. on two small carriages T T, running on rollers upon girders Z. These carriages have hooked brackets, as shown, on which the pieces of rail are caught and rest. The arms are then unbolted, and the carriages, with the sleeper &c. upon them, are pushed by hand, or run by gravitation, along the girders Z to the front of the machine, where the sleeper &c. is picked up by exactly similar arms or levers, and brought down and deposited in its place at the bottom of the heading. The girders, with the carriages upon them, can be traversed endways through a short distance, so as to be out of the way while the sleeper is being raised or lcwered. Reduction of the Chalk debris to Sludge or Cream.—Some years since it became necessary for me to design an apparatus which in a small compass could produce large quantities of chalk-cream, such as is used in all cement works. The apparatus, which has been long in operation, is a plain cylindrical drum, 4 feet in diameter, and 2 feet 8 inches long inside, revolving at the rate of 32 revolutions per minute. One face of this drum is made of a strong wire grating, except in the centre, where a hole of 15 inches diameter is left. Through this central aperture the chalk debris and the water, in whatever quantities may be required, are introduced, and as the drum revolves, the particles of chalk, saturated and softened by contact with the water, are quickly dissolved; a cream or sludge, of more or less consistency, is thus produced, which escapes through the meshes of the wire grating and collects in a reservoir. It is then taken up by a pump, and forced to any place required. In this small apparatus a quantity of chalk debris, amounting to 14 cubic yards, or 21 tons, was reduced to cream within one hour. Fig. 2, Plate 83, shows a drum of the dimensions first given, placed in an 8-foot tunnel, a space considerably larger than is necessary to contain it. It will be a simple question of proportion to determine the size of the cylindrical drums capable of reducing

143

AUGUST 1882.

CHANNEL TUNXEL.

447

the 170 cubic yards of chalk debris per hour, resulting from the cutting of a tunnel 36 feet in diameter. As a matter of fact, two drums 7 feet diameter and 7 feet in length will be amply sufficient for that purpose. It will be quite safe to assume, as my experiments show, that one half of a horse-power per cubic yard of chalk per hour is sufficient for its reduction ; this gives a total under this head of 85 HP. Conveyance of Cream back to the bottom of the Shaft.—Before deciding upon the quantity of water necessary to be mixed with the chalk, for its speedy reduction to cream, I made a series of trials in small pipes, with a view of ascertaining the amount of extra friction which may be caused by the passage of cream through them, as against the passage of water alone. I need not enter into details, but may state at once that while an admixture by weight of equal quantities of chalk and water required nearly 14 per cent, more power than pure water, an admixture of 1 of chalk and 2 of water required 33/4per cent., and an admixture of 1 of chalk to 3 of water only 21/3per cent.; it was therefore decided to use the proportion of 1 of chalk to 3 of water by weight, or 1 of chalk to 6 of water by bulk. It was also ascertained that it would not be safe to pass cream through long lengths of pipes at a less velocity than 1 ft. 6 in. per second, as otherwise there would be a tendency for the solid particles to settle ; I therefore decided upon a minimum velocity of 2 feet per second, at which to pass the cream. It was previously stated that 170 cubic yards of chalk debris would be produced per yard forward in driving a 36-foot tunnel; we require therefore 170 x 6 = 1020 cubic yards of water per hour, or 17 cubic yards per minute, equal to 459 cubic feet. Now a 12-inch main will deliver this quantity, at the end of 10 miles of pipe, at a pressure of 700 Ibs. per square inch; the water passing through it at a velocity of 9 • 5 feet per second. The total horse-power developed by this quantity of water, at 700 Ibs. pressure, amounts to 1377 HP., at our disposal at the face; 337 HP. being given by the pressure of the sea, and 1040 by pumping machinery.

144

448

CHANNEL TUNNEL.

AUGUST 1882.

The sludge is composed of— Chalk, 76 cubic feet per minute, Water, 459 cubic feet per minute, Total of cream, 535 cubic feet per minute. The weight of the cream is 72 Ibs. per cubic foot. A main outlet pipe, 20 inches in diameter, will be required to convey the cream back to the bottom of the shaft through 10 miles of level tunnel, the cream flowing through it at a velocity of 245 feet per minute or 4 feet per second. The total head required to force the cream along a level to the bottom of the shaft is 214 feet or211/2feet per mile. This represents a force of 224 HP., the pressure in the pipe being 96 Ibs. per square inch. Lifting the Cream from the bottom of the Shaft to the Surface.— 535 cubic feet of cream, at 72 Ibs. per cubic foot, lifted 400 feet per minute, will consume about 525 HP., including contingencies. Hence the power required for the several operations is as follows:— (1) Cutting the chalk (page 445) = (2) Keduction of chalk to cream (page 447) = (3) Conveyance of cream to bottom of shaft through 10 miles of pipe = Total required at the face

..

340 HP. 85 „ 224 „ 649 HP.

As we have provided 1377 HP., there will be no deficiency even if the hydraulic motor should only yield 50 per cent, duty, which is a very low estimate. The 525 HP., above shown as required for lifting the cream from bottom of shaft to top, will, of course, have to be provided at the top of the shaft, and will be in addition to the power necessary for the compression of the water. To compress 459 cubic feet of water per minute to a pressure of 512 Ibs. per sq. in. would require a force of 1,040 HP. (the sea pressure giving 337 HP. additional).

145

AUGUST 1882.

CHANNEL TUNNEL.

449

We have therefore to provide on top of shaft, For compression of water . . . „ pumping up the cream . . Total

.

.

1,040 HP. 525 „ 1,565 HP.,

to carry out the entire operation of cutting 170 cubic yards per hour, reducing it to cream, pumping it through 10 miles of pipes on a level into a sump at the bottom of the shaft, and lifting it to the top. It may be remarked that the cream can be pumped from the face along the tunnel up the shaft in one operation, but more power would be required to do this. The above power is independent of that required to transport the material necessary for lining the tunnel, which will be done by locomotive or other means, as in the ordinary system. As an alternative scheme to the one just explained, I would propose, where the tunnel has sufficient fall for drainage, which the Channel Tunnel should have, to do away with the compression of the water on the surface and utilise the pressure of the sea alone; in other words to syphon the water out of the sea and convey it by its own fall 400 feet below the level of the sea, giving a pressure of 188 Ibs. per square inch. In this case the quantity of water, in order to do the work efficiently, must be considerably larger; and instead of a 12-inch main inlet pipe, one of 22 inches will be necessary. If the tunnel is made at an inclination of 1 in 1,000, the cream will flow by gravity to the bottom of the shaft, from which it will be pumped to the top. The larger quantity of water sent down will make the proportion of chalk to water as 1 to 13 ; in that proportion the cream will run as easily as water itself, and an open drain will be sufficient, dispensing with the closed pipe. The power required for cutting down the chalk and for reducing it to cream would be the same as in the former case, namely 425 HP.; but the third item of power for forcing the cream to the bottom of the shaft will not need to be provided for. The 22-inch main inlet pipe, under a pressure of 188 Ibs. per sq. in., will deliver

146

450

CHANNEL TUNNEL.

AUGUST 1882.

974 cub. ft. of water per minute to the face 10 miles off, at a velocity of 368 feet per minute or 6.1 feet per second; representing a total of 800 HP. available, or nearly double that required. It will be necessary in this case to lift a larger quantity of creamup the shaft—nearly double of what it is in the other case ; and the power for this amounts to 950 HP. We have therefore to provide on top of shaft, For compression of water . . . Nil. „ pumping up the cream . . . 9 5 0 HP. Total

.

.

.

.

950HP.

as against 1,565 HP. to be provided for in the system.

high-pressure

I have shown, I think, that the system I propose to use offers some considerable advantages as regards the traffic in and out of the tunnel, and the consequent facilities afforded for the speedy lining of the tunnel; hence it cannot but result in a great saving in time and expense. For the few men employed at the boring machinery, I propose to carry sufficient air under compression with the water; which being discharged at the motors will keep that part of the tunnel cool and well ventilated. For the transport of the comparatively small amount of material1 for lining the tunnel (about one-fifth of the whole quantity to be moved), electric or air locomotives may be employed. If the latter, the air discharged would ventilate the tunnel; but in either case a very small pipe would convey sufficient air for that purpose. It may be observed in conclusion, that I consider the power required by the hydraulic system will not amount to one-third of that required when the chalk is cut in a dry state by compressed-air machinery, and has to be conveyed by air-locomotives in trucks, and* lifted to the surface by the ordinary means.

147

Fig . 1.

General

ctrrctiiyetment

for

xcavating

8

ft. heading .

Pig. 2. Elevation of Excavator for

"Fig .3.

8 ft.

heading.

Fig .4. End Elevation of Boring looking backwar.

8ft. heading .

Head,

Fig. 5.

Section

at XX, Fig. 2 looking forwards.

Beauchamp Tower (1845-1904) One of many engineers who entered their pupilage at Armstrong's Elswick Works in Newcastle-upon-Tyne, Tower's later professional experiences centred on the shipbuilding industry. He worked on the construction of iron steamers on Tyneside until 1868. Thereafter, he concentrated on developing his expertise in naval engineering, gaining the best possible start as assistant to William Froude (1810-1879), who was then engaged in designing mechanical apparatus for the Admiralty Experimental Works at Torquay. Tower alternated between work for Froude, including developing a marine engine dynamometer, and for William Armstrong and Co, for whom he conducted experiments on torpedoes in the years 1874—1875. The firm's interest in this area would be developed shortly afterward by the young Charles Parsons. By 1877 Tower had followed the ailing Froude to South Africa, where the latter died. From 1878, Tower was an independent engineer. His subsequent career fell into two areas, the importance of each being perceived quite differently by contemporary and modern engineering worlds. Tower expended considerable energy on behalf of the Admiralty in the development of a gyroscopic gun platform for use at sea during the 1880s, but the invention was never used. This was "the great work of his life" according to the obituary published in the Institution's Proceedings. Far more significant was his involvement with the IMechE's Committee on Friction Experiments. His role as experimenter was relayed to Members in four committee reports. Here Tower laid the foundations of modern lubrication theory.

151

Nov. 1883.

632

FIKST REPORT ON FRICTION EXPERIMENTS. BY MR. BEAUCHAMP TOWER, or LONDON. (Adopted by the Committee on Friction, and presented to the Council, Sept. 28,1883.)

I. DESCRIPTION OF MACHINE. In experimenting on the friction of lubricated bearings, and on the value of different lubricants, one of the difficulties which is first met with is the want of a method of applying the lubricant, which can be relied upon as sufficiently uniform in its action. All the common methods of lubrication are so irregular in their action that the friction of a bearing often varies considerably. This variation, though small enough to be of no practical importance, and to pass unnoticed, in the working of an ordinary machine, would be large enough utterly to destroy the value of a set of experiments, say, on the relative values of various lubricants; for it would be impossible to know whether an observed variation was due to a difference in the quality of the oil, or in its rate of application. The first problem therefore which presented itself, in the present experiments, was to devise a method of lubrication such as would be perfectly uniform in its action, and would form an easily reproducible standard with which to compare other methods. These conditions were best fulfilled by making the bearing run immersed in a bath of oil. By this method the bearing is always supplied with as much oil as it can possibly take; so that it represents the most perfect lubrication possible, and is a good standard with which to compare other methods. It is at all times perfectly uniform in its action. It is very easily defined and reproduced; and it also has the advantage that the temperature of the bearing can be easily regulated by gas jets under the bath. Experiment showed that the bath need not be full; the results obtained were the same when it was so nearly empty that the bottom of the journal only just touched the oil.

153

Nov. 1883.

FRICTION EXPEEIMENTS.

633

The journal experimented on (see Figs. 1 and 2, Plates 88 and 89) was of steel, 4 in. diameter and 6 in. long, with its axis horizontal. A gun-metal brass A, embracing somewhat less than half the circumference of the journal, rested on its upper side. The exact arc of contact of this brass was varied in the different experiments, as will be seen by reference to the appended Tables of results. Resting on this brass was a cast-iron cap B, from which was hung by two bolts a cast-iron cross-bar C carrying a knife-edge. The exact distance of this knife-edge below the centre of the journal was five inches. On this knife-edge was suspended the cradle D which carried the weight applied to the bearing. The cap, bolts, and cross-bar were put together in such a manner as to form a rigid frame, connecting the brass with the knife-edge. If there had been no friction between the brass and the journal, the weight would have caused the knifeedge to hang perpendicularly below the axis of the journal. Friction however caused the journal to tend to carry the brass, and the frame to which it was attached, round with it, until the line through the centre of journal and the knife-edge made such an angle with the perpendicular that the weight multiplied by the distance from the knife-edge to that perpendicular offered an opposing moment just equal to the moment of friction. Suppose r = radius of the journal (Fig. 3, Plate 88). s = distance of the knife-edge from the perpendicular. w = the weight. Then, from above, s x w = the moment of friction. Now the friction at the surface of the journal

Hence the coefficient

of friction =

So that the coefficient of friction is indicated by s in terms of r, no matter what the weight is. As an example, suppose s was equal to r; the coefficient of friction would obviously be 1; or if s was1/10of r, then the coefficient of friction would be 1/10. In order to avoid the difficulty of determining accurately when

154

634

FRICTION EXPERIMENTS.

Nov. 1883.

the knife-edge was perpendicularly under the centre of the journal (a knowledge which was necessary in order to obtain a measurement of s, and which was very difficult to obtain owing to the considerable friction between the brass and the journal when at rest), each experiment was tried with the journal revolving in both directions, and the sum of the values of s on both sides was measured; and then the coefficient of friction was indicated by the chord of the whole angle, included between the two lines of inclination caused by the friction, with the rotation in the two directions, the chord being expressed in terms of the diameter of the journal (see Fig. 4, Plate 88). Each result was thus a mean of two experiments, one with the axle running in one direction, and the other with it running in the other direction. In order to read the value of the coefficients thus obtained, a light horizontal lever L was attached (as shown in Fig. 1) to the frame connecting the brass to the knife-edge. It was 621/2inches long, or 121/2times the distance between the centre of the journal and the knife-edge ; so that at the end of the lever the chord indicating the coefficient of friction was magnified121/2times. As a chord of 4 inches at the knife-edge represents a coefficient of 1, a chord of 50 inches at the end of the lever also represents a coefficient of 1, while 5 inches represents a coefficient of1/10,1/2-inch of1/100,and 1/20-inch of 1/1000. The position of the end of the lever during each experiment was recorded by a tracing point attached to the end of the lever, and marking on metallic paper carried upon a revolving vertical cylinder P. The distance between the two lines obtained by running the axle both ways, when measured on the above scale, indicated the value of the coefficient. II. METHOD OP EXPERIMENTING. Early in the experiments it was found that, immediately after the motion of the shaft was reversed, the friction was greater than it was when the shaft had been running in the same direction some time. This increase of friction, due to reversal, varied considerably. It was greatest with a new brass, and diminished as the brass became worn, so as to fit the journal more perfectly. Its greatest observed

155

Nov.

1883.

FRICTION EXPERIMENTS.

635

amount was at starting and was about twice the normal friction, and it gradually diminished until the normal friction was reached after about ten minutes continuous running. This increase of friction was accompanied by a strong tendency to heat and seize, even under a moderate load. In the case of one brass, which had worked for a considerable time without accident, and had consequently become worn so as to fit the journal very accurately, this tendency to increase of friction after reversal almost entirely disappeared; and it could be reversed under a full load without appreciable increase of friction or a tendency to heat or seize. The phenomenon must be due to the surface fibres of the metal, which have been for some time stroked in one direction, meeting point to point and interlocking when the motion is reversed. The very perfectly fitting brass was probably entirely separated from the journal by a film of oil; and there being no metallic contact the phenomenon did not show itself. In consequence of the above facts, it was found necessary to proceed with the experiments in the following order. A complete series of experiments, with a gradually increasing load, was taken with the journal running in one direction; the load was then diminished by the same steps as it had been increased, and the experiments thus repeated, the shaft still running in the same direction, until the load had thus been reduced to 100 Ibs. per sq. in., which was the load clue to the unweighted cradle. The direction of motion was then reversed, and the shaft run for half-an-hour, so as to get it thoroughly used to going the other way ; after this the load could be increased and the experiments taken without any difficulty. The experiments, as before, were taken at each step of both increasing and decreasing the load ; so that each recorded result is really the mean of four experiments, which have in many instances been taken several hours apart. This method of obtaining a direct indication of the coefficient of friction, by the angular displacement of the frame connecting the brass and knife-edge, would undoubtedly have been the best had the coefficient of friction been nearly as constant as it has hitherto been supposed to be. But as shown by the Tables of results, the coefficient

156

636

FRICTION EXPERIMENTS.

Nov.

1883,

of friction was found, instead of being constant, to vary nearly inversely as the load, and also to be much smaller in quantity than was expected; the consequence was that with high loads the height of the diagram was very small. In the cases where with the greatest loads a coefficient of only 1/1000 was observed, the distance between the two lines was only1/20inch. The results shown in Tables I., II., III., IV., were obtained in this way. Owing to these experiments showing that the moment of friction was much more nearly constant than the coefficient, it was resolved to alter the method of observation, and to measure the moment directly instead of the coefficient. For this purpose the paper cylinder was removed, and a small lever M (see Fig. 1A, Plate 88) was connected to the main indicating lever in such a manner that the motion of the end of the main lever was magnified five times at the end of the small lever. The end of the small lever was pointed; and when the machine was working, this point was brought exactly opposite a fixed mark by putting weights into a scale-pan on the end of the main lever. The main lever was so overbalanced that under all circumstances some weight was required to be added in the scale-pan, in order to bring the end of the small lever to the mark, even when, in addition to the friction being greatest, the direction of motion of the journal tended most to depress it. The method of running in both directions, and loading and unloading, was followed as before. The weights in the scale-pan being noted, the moment of friction was given by half the difference between the weights in the scale-pan when running in one direction and in the other. The results given in Tables V., VI., VII., VIII., were obtained in this manner. Experiment showed that the friction varied considerably with temperature (see Table IX.) All the oil-bath experiments were therefore taken at a nearly uniform temperature of 90°; the variation above or below this temperature was never allowed to be more than11/2°.

157

Nov. 1883.

FRICTION EXPERIMENTS.

III. KESULTS OF EXPERIMENTS. In the experiments shown in Tables I., II., and III., care was taken not to load the bearing up to seizing, in order that the condition of the brass might not be disturbed. In Table IV. the bearing seized unintentionally. In Tables V., VI., VII., and VIII., the bearing was intentionally loaded up to seizing. The experiments shown in Tables V. and VI. were specially made for the purpose of ascertaining the greatest load which could be carried with rape and mineral oil in the oil bath. The greatest load carried with the rape oil was 573 Ibs. per. sq. in., and the greatest load carried with the mineral oil was 625 Ibs. In both of these cases the experiment was repeated after the brass had been taken out and scraped up, but with no better result. The general results of, the oil-bath experiments may be described as follows. The absolute friction, that is the actual tangential force per sq. in. of bearing, required to resist the tendency of the brass to go round with the journal, is nearly a constant under all loads, within ordinary working limits. Most certainly it does not increase in direct proportion to the load, as it should do according to the ordinary theory of solid friction. The ordinary theory of solid friction is that it varies in direct proportion to the load; that it is independent of the extent of surface; and that it tends to diminish with an increase of velocity beyond a certain limit. The theory of liquid friction, on the other hand, is that it is independent of the pressure per unit of surface, is directly dependent on the extent of surface, and increases as the square of the velocity. The results of these experiments seem to show that the friction of a perfectly lubricated journal follows the laws of liquid friction much more closely than those of solid friction. They show that under these circumstances the friction is nearly independent of the pressure per sq. in., and that it increases with the velocity, though at a rate not nearly so rapid as the square of the velocity. The experiments on friction at different temperatures, shown in Table IX., indicate a very great diminution in the friction as the

158

638

FEICTION EXPERIMENTS.

Nov.

1883.

temperature rises. Thus, in the case of lard oil, taking a speed of 450 revolutions per minute, the coefficient of friction at a temperature of 120° is only one-third of what it was at a temperature of 60°. A very interesting discovery was made when the oil-bath experiments were on the point of completion. The experiments being carried on were those on mineral oil; and the bearing having seized with 625 Ibs. per sq. in., the brass was taken out and examined, and the experiment repeated. While the brass was out, the opportunity was taken to drill a1/2-in.hole for an ordinary lubricator through the castiron cap and the brass. On the machine being put together again and started with the oil in the bath, oil was observed to rise in the holewhich had been drilled for the lubricator. The oil flowing over the top of the cap made a mess, and an attempt was made to plug up the hole, first with a cork and then with a wooden plug. When the machine was started the plug was slowly forced out by the oil in a way which showed that it was acted on by a considerable pressure. A pressure-gauge was screwed into the hole, and on the machine being started the pressure, as indicated by the gauge, gradually rose to above 200 Ibs. per sq. in. The gauge was only graduated up to 200 Ibs., and the pointer went beyond the highest graduation. The mean load on the horizontal section of the journal was only 100 Ibs. per sq. in. This experiment showed conclusively that the brass wasactually floating on a film of oil, subject to a pressure due to the load. The pressure in the middle of the brass was thus more than double the mean pressure. No doubt if there had been a number of pressure-gauges connected to various parts of the brass, they would have shown that the pressure was highest in the middle, and diminished to nothing towards the edges of the brass. The experiments with ordinary lubrication were begun with a needle lubricator, the hole from which penetrated to the centre of the brass. A groove in the middle of the brass, and parallel to the axis of the journal, extended nearly to the ends of the bearing for distributing the oil (see Figs. 5 and 6, Plate 90). It was found that with this arrangement the bearing would not run cool when loaded with only 100 Ibs. per. so. in. ; and that not a drop of oil

159

NOV. 1883.

FRICTION EXPERIMENTS.

630

would go down even when the needle-lubricator was removed and the hole filled completely with oil, thus giving a head of 7 inches of oil to force it into the brass. It appeared as though the hole and groove, being in the centre of pressure of the brass, allowed the supporting oil-film to escape. This view was confirmed by the following experiment. The oil-hole being filled up to the top, the weight was eased off the journal for an instant. This allowed the oil to sink down in the hole and lubricate the journal; but immediately the load was again allowed to press on the journal the oil rose in the hole to its former level, and the journal became dry, thus showing that this arrangement of hole and groove, instead of being a means of lubricating the journal, was a most effectual one for collecting and removing all oil from it. It should be mentioned that care was taken to chamfer the edges of the groove, so as to prevent any scraping action between them and the journal. As the centre of the brass was obviously the wrong place to introduce the oil, it was resolved to try to introduce it at the sides. Accordingly the centre hole and groove were filled up, and two grooves were made. These grooves were parallel to the axis of the journal, extending nearly to the ends of the brass, and were placed at equal distances on either side of the centre; they formed boundaries to an arc of contact, the chord of which was 31/4 in. (see Figs. 7 and 8, Plate 90). With this arrangement of groove the lubrication appeared to be satisfactory, the oil going down into the journal and the bearing running cool. The results of the experiments with this arrangement of brass are given in Table VII. The bearing nevertheless seized with an actual load of only 380 Ibs. per sq. in. The arrangement of grooves was then altered to that usual in locomotive axle-boxes (see Figs. 9-11, Plate 89). The oil was introduced through two holes, one near each end of the brass, and each connected to a curved groove; the two curved grooves nearly enclosing an oval-shaped space in the centre of the brass. At the same time the arc of contact was reduced till its chord was only 21/4in. This brass refused to take its oil or run cool. It would sometimes run for a short time with an actual load of 178 Ibs. per

160

640

FRICTION EXPEEIMENTS.

Nov.

1883.

sq. in., but rapidly heated on the slightest increase of the load. The brass having been a good deal cut about by altering and filling up grooves, it was considered desirable to have a new brass, and one was accordingly obtained. The grooves being made exactly the same as in the last experiment with the old one, this brass seized with an actual load of only about 200 Ibs. per sq. inch. The oilbox was completely cut away so as to allow a freer current of air round the bearing, and the lubricator pipes were soldered into the brass. The wicks were taken out of the lubricators and the lubricators filled full of oil, by which means oil was supplied to the brass under a full head of 9 in.; and yet the oil refused to go. down, and the underside of the journal felt perfectly dry to the hand, and speedily heated with a load of only 200 Ibs. per. sq. in. The fact that this arrangement of grooves, which is found to answer in the axles of railway vehicles, was found to be perfectly useless in this apparatus, can only be accounted for by the fact that a railway axlc has a continual end play while running, which prevents the brass from becoming the perfect oil-tight fit which it became in this apparatus. The attempts to make this arrangement of lubrication answer were not abandoned until after repeated trials. It now became clear that there was no use in trying to introduce the oil directly to the part of the brass against which the pressure acted, and that the only way to proceed was to oil the lower side of the journal, and trust to the oil being carried round by the journal to the seat of the pressure. The grooves and holes in the brass were accordingly filled up, and an oily pad, contained in a tin box full of rape oil, was placed under the journal, so that the journal rubbed against it in turning. The pad was only supplied with oil by capillary attraction from the oil in the box, and the supply of oil to the journal was thus very small; the oiliness in fact was only just perceptible to the touch, but it was evenly and uniformly distributed over the whole journal. The results are given in Table VIII. The bearing fairly carried 551 Ibs. per sq, in., and three observations were obtained with 582 Ibs., but the bearing was on the point of seizing and did seize after running a few minutes with this load. It will be observed that in this instance

161

Nov. 1883.

FRICTION EXPERIMENTS.

641

the bearing seized with very nearly the same load as it did in the oil-bath experiment with rape oil. These experiments with the oily pad show a nearer approach to the ordinarily received laws of solid friction than any of the others. The coefficient is approximately constant, and may be stated as about 1/100 on an average. There does not in this case appear to be any well-defined variation of friction with variations of speed, according to any regular law. The results of the experiments with rape oil, fed by a syphon lubricator to side grooves (Table VII.), follow nearly the same law as the results obtained from the oil-bath experiments, as far as the approximate constancy of the moment of friction is concerned; but the amount of the friction is about four times the amount in the oil-bath. It should be stated that though only these two Tables are given as the results of the experiments on what is called ordinary lubrication, that is, lubrication by means other than that of the oil-bath, they represent only a small part of the experiments or attempts at experiments which were made on this subject. But they are the only experiments the results of which were sufficiently regular to make them worthy of record. Indeed the results, generally speaking, were so uncertain and irregular that they may be summed up in a few words. The friction depends on the quantity and uniformity of distribution of the oil, and may be anything between the oil-bath results and seizing, according to the perfection or imperfection of the lubrication. The lubrication may be very small, giving a coefficient of 1/100; but it appeared as though it could not be diminished and the friction increased much beyond this point without imminent risk of heating and seizing. The oil-bath probably represents the most perfect lubrication possible, and the limit beyond which friction cannot be reduced by lubrication; and the experiments show that with speeds of from 100 to 200 ft. per minute, by properly proportioning the bearing-surface to the load, it is possible to reduce the coefficient of friction as low as 1/1000. A coefficient of 1/500 is easily attainable, and probably is frequently attained, in ordinary engine-bearings in which the direction of the force is rapidly alternating and the oil given an 3 A

162

FRICTION EXPERIMENTS.

Nov. 1883.

opportunity to get between the surfaces, while the duration of the force in one direction is not sufficient to allow time for the oil-film to be squeezed out. The extent to which the friction depends on the quantity of the lubrication is shown in a remarkable manner in Table X.; which proves that the lubrication can be so diminished that the friction is seven times greater than it was in the oil bath, and yet that the bearing will run without seizing. Observations on the behaviour of the apparatus gave reason to believe that with perfect lubrication the speed of minimum friction was from 100 to 150 feet per minute; and that this speed of minimum friction tended to be higher with an increase of load, and also with less perfect lubrication. By the speed of minimum friction is meant that speed in approaching which, from rest, the friction diminishes, and above which the friction increases.

163

Nov. 1883.

FRICTION

EXPERIMENTS.

643

TABLE I.—BATH OF OLIVE OIL. TEMPERATURE 90° F. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF THE ARC OF CONTACT = 3.92 IN. * Nominal LOAD 100 rev. Lbs. 105 ft. per sq. in. per min.

COEFFICIENTS OF FKICTION, for speeds as below. 400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

•0014

•0015

•0017

•0015

•0017

•0018

•002

•0015

•0017

•0019

•0021

•0024

•0016

•0017

•0019

•002

•0022

•0025

•0015

•0017

•0019

•0021

•0022

•0024

•0027

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

520

•0008

•001

•0012

•0013

468

•0011

•0013

•0014

415

•0012

•0014

363

•0013

310

300 rev. 314 ft. per min.

350 rev. 3G6 ft. per min.

Lbs.

258

•0011

•0017

•002

•0023

•0025

•0026

•0029

•0031

205

•0018

•0021

•0025

•0028

•003

•0033

•0036

•004

153

•0023

•003

•0035

.004

•0044

•0047

•005

•0057

100

•0036

•00-15

•0055

•0063

•0069

•0077

•0082

•0089

The above coefficients x the nominal load = nominal frictional resistance. per sq. inch of bearing. NOMINAL FACTIONAL RESISTANCE per sq. in. of bearing.

* Nominal LOAD. 100 rev. Lbs. 105 ft. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

520

•416

•520

•624

•675

•728

•779

•883

468

•514

•607

•654

•701

•794

•841

•935

415

••198

•580

•622

•705

•787

•870

•995

363

•472

•580

•616

•689

•725

•798

•907

310

•464

•526

•588

•650

•680

•742

•835

Lbs.

Lb.

258

•361

•438

•515

•592

•644

•669

•747

•798

205

•36S

•43

•512

•572

•613

•675

•736

•818

153

•351

•458

•535

•611

•672

•718

•764

•871

100

•36

•45

•55

•63

•69

•77

•82

•89

The nominal load per sq. in. is the total load divided by (4 x 6).

3 A2 164

644

Nov. 1883.

FRICTION EXPERIMENTS.

TABLE II.—BATH OF LAED OIL. TEMPERATURE 90° F. 4-IN. JOURNAL6-iN. LONG. CHORD OF ARC OF CONTACT OF BRASS = 3.92 IN. COEFFICIENTS OF FKICTION, for speeds as below. Nominal LOAD 100 rev. Lbs. 105 ft. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft, per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

520

•0009

•001

•0011

•0013

•0015

•0015

•0017

415

•0012

•0014

•0015

•0016

•0018

•0019

•0021

310

•0014

•0017

•002

•0022

•0025

•0026

•0029

Lbs.

205

•0017

•0020

•0023

•0028

•0031

•0034

•0039

•0042

153

•0022

•0027

•0032

•0037

•0041

•005

•0051

•0052

100

•0035

•0042

•005

•006

•0067

•0076

•0081

•009

The above coefficients x the load = nominal frictional resistance per sq. in. of bearing. NOMINAL FRICTIONAL RESISTANCE per sq. in. of bearing. Nominal 100 rev. LOAD 105 ft. Lbs. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

520

•468

•52

•572

•675

•779

•779

•883

415

•498

•58

•621

•663

•746

•788

•870

310

•434

•526

•619

•680

•774

•804

•898

Lbs.

Lb.

205

•348

•409

•47

•572

•634

•696

•798

•859

153

•336

•412

•489

•565

•626

•764

•779

•795

100

•35

•42

•5

•6

•67

•76

•81

•9

165

Nov. 1883.

PBICTION EXPERIMENTS.

645

TABLE III.-—BATH OF MINERAL GREASE. TEMPERATURE 90° F. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF ARC OF CONTACT OF BRASS = 3 • 92 IN. COEFFICIENTS OF FRICTION, for speeds as below. Nominal 100 rev. LOAD 105 ft. Lbs. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

Lbs.

625

•001

•0012

•0014

•0014

•0016

•0018

•002

520

•0014

•0016

•0018

•0019

•002

•0021

•0022

415

•0016

•0019

•0021

•0023

•0025

•0026

•0027

310

•002

•0022

•0026

•0029

•0032

•0035

•0038

•004

205

•0026

•0034

•0040

•0047

•0053

•0058

•0062

•0066

153

•0028

•0038

•0048

•0057

•0065

•0071

•0077

•0083

100

•0054

•0076

•0094

•0109

•0123

•0133

•0142

•0151

The above coefficients x the nominal load = nominal frictional resistance per sq. in. of bearing. NOMINAL FKICTIONAL RESISTANCE per sq. in. of bearing. Nominal 100 rev. LOAD 105 ft. Lbs. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

625

•625

•75

•875

•875

•999

1.125

1.25

520

•727

•831

•935

•987

1.04

1.091

1.143

415

•663

•787

•87

•953

1.036

1.074

1.12

Lbs.

166

Lb.

310

•62

•682

•805

•899

•992

1.085

1.18

1.24

205

•531

•696

•818

•962

1.085

1.188

1.27

1.35

153

•428

•581

•734

•871

1.085

1.177

1.27

100

•54

•76

•94

1.33

1.42

1.51

1.09

•994

1.123

646

Nov. 1883.

FRICTION EXPERIMENTS.

TABLE IV.—BATH OF SPERM OIL. TEMPERATURE 90° F. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF ARC OF CONTACT OF BRASS = 3 • 92 IN. COEFFICIENTS OF FRICTION, for speeds as below. Nominal LOAD 100 rev. Lbs. 105 ft. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

415

•0015

•0017

•0018

•0019

310

•0011

•0012

•0014

400 rev. 419ft. per min.

450 rev. 471 ft. per min.

•002

•0021

•0021

•0016

•0017

•0018

•0019

350 rev. 366 ft. per min.

Lbs.

520

Seized

205

•0013

•0016

•0018

•0021

•0023

•0024

•0025

•0027

153

•0016

•0019

•0023

•0028

•0030

•0033

•0035

•0037

100

•0025

•003

•0038

•0044

•0051

•0057

•0061

•0064

The above coefficients x the nominal load = nominal friction resistance per sq. in. of bearing. NOMINAL FBICTIOXAL RESISTANCE per sq. in. of bearing. Nominal 100 rev. LOAD 105 ft. Lbs. per sq. in. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

400 rev. 419 ft. per min.

450 rev. 471 ft. per min.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

Lb.

415

•621

•705

•746

•788

•829

•87

•87

310

•341

•372

•434

•495

•526

•557

•588

Lbs.

Lb.

520

Seized

205

•266

•327

•368

•43

•471

•491

•512

•552

153

•244

•291

•352

•428

•459

•505

•535

•566

100

•25

•3

•38

•44

•51

•57

•61

•64

167

Nov. 1883.

FRICTION EXPERIMENTS.

647

TABLE V.—BATH OF RAPE OIL. TEMPERATUBE 90° F. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF ARC OF CONTACT OF BRASS = 3.92 IN. Nominal LOAD Lbs. per sq. in.

COEFFICIENTS OF FRICTION, for speeds as below. 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 450 rev. 262 ft. 105 ft. 157 ft. 209 ft. 314 ft, 419 ft. 366 ft. 471 ft. per min. per min. per min. per min. per min. per min. per min. per min.

Lbs. 573

•00102

520

•000955 •00105 •00115 •00125 •00133 •00142 •00148

415

•00093

•00107 •00119 •0013

363

•00084

•0096

258

•00107 •00139

•00162 •00178 •00195 •00213 •00227 •00243

153

•00162 •0020

•00239 •00267 •003

100

•00277 •00357

•00423 •00503 •00576 •00619 •00663 •00714

•00108 •00118 •00126 •00132 •00139

•0011

•00140 •00149 •00158

•00122 •00134 •00147 •00155

•00334 •00367 •00396

The above coefficients x the nominal load — nominal frictional resistance per sq. in. of bearing. Nominal LOAD Lbs. per sq. in.

NOMINAL FKICTIONAL RESISTANCE per sq. in. of bearing. 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 450 rev. 471 ft. 419 ft. 366 ft. 314 ft. 262 ft. 209 ft. 157 ft. 105 ft. per mm. per min. per mm. per mm. per mm. per mm. per mm. per mm.

Lb. •583

Lb. •62

Lb. •678

Lb. •721

Lb. •794

Lb.

•758

520

•496

•546

•597

•648

•691

•735

•771

415

•386

•445

•495

•539

•582

•619

•655

363

'306

•35

•401

•444

•488

•532

•561

Lbs. 573

Lb.

Lb.

258

•277

•357

•416

•459

•503

•547

•583

•626

153

•248

•306

•364

•408

•459

•510

•561

•605

100

•277

•357

•423

•503

•576

•619

•663

•714

N.B.—The bearing seized on reversing with 573 Ibs. per square inch. The experiment was repeated, but the bearing refused to carry more weight. These quantities were obtained by a direct load on the lever, so that in these the coefficient is calculated from the force on the lever, instead of the force on the lever being calculated from the coefficient as was the case in the former experiments.

168

648

Nov. 1883.

FRICTION EXPERIMENTS.

TABLE VI.—BATH OF MINERAL OIL. TEMPERATURE 90° F. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF ARC OF CONTACT OF BRAss = 3.92 IN. Nominal LOAD Lbs. per sq. in.

COEFFICIENTS OF FRICTION, for speeds as below.

400 rev. 419 ft. per min.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

250 rev. 262 ft. per min.

300 rev. 314 ft. per min.

350 rev. 366 ft. per min.

Lbs. 625

•0013

•00139

•00147

•00157

•00165

520

•00123

•00139

•00.15

•00161

•0017

•00178

415

•00123

•00143

•0016

•00176

•0019

•002

310

•00142

•0016

•00184

•00207

•00225

•00241

100 rev. 105 ft. per min.

205

•00178

•00205

•00235

•00269

•00298

•00328

•0035

100

•00334

•00415

•00494

•00557

•0062

•00676

•0073

The above coefficients x the nominal load = nominal frictional resistance per sq. in. of bearing. Nominal LOAD Lbs. per sq. in.

NOMINAL FKICTIONAL RESISTANCE per sq. in. of bearing.

300 rev. 314 ft. per mm.

350 rev. 366 ft.

400 rev. 419 ft. per mm.

100 rev. 105 ft. per mm.

150 rev. 157 ft. per min.

200 rev. 209 ft. per min.

per mm.

Lb.

Lb. •81

Lb. •865

Lb. •92

Lb. •98

520

•64

•72

•782

•84

•886

•924

415

•51

•594

•664

•73

•785

•83

310

•44

•494

•57

•64

•695

•745

Lbs. (525

250 rev. 262 ft.

per mm.

Lb. 1.03

Lb.

205

•364

•419

•48

•55

•61

•67

•716

100

•334

•415

•494

•557

•62

•676

•73

N.B.—The bearing carried the 625 Ibs. per sq. in. running both ways, but seized on the weight being increased. These quantities were obtained by a direct load on the lever, as in Table V. This was a thinner sample of mineral oil than that used in the previous experiments; it was fluid at 50°, while the oil previously used could only be described as grease at 50°. This will account for these experiments showing less friction than the former, except with the highest load, at which, the thin oil being overloaded and on the point of seizing, the friction is greater than with the thick oil.

169

Nov. 1883.

649

FRICTION EXPERIMENTS.

TABLE VII.—RAPE OIL, FED BY SYPHON LUBRICATOR. 4-IN. JOURNAL, 6 IN. LONG. Nominal LOAD Lbs. per sq. in.

Actual LOAD Lbs. per sq. in.

Lbs. 258

Lbs. 317

205

252

100

123

CHORD OF ARC 31/4 IN.

COEFFICIENTS OF FRICTION, for speeds as below.

100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 209 ft. 262 ft. 157 ft. 314 ft. 105 it. 366 ft. 419 ft. per min. per min. per min. per min. per min. per min. per min.

•0056

•0057

•0063

•0068

•0132

•0098

•007

•0077

•0082

•0087

•0144

•0125

•0146

•0152

•0163

•0171

•0178

The above coefficients x the nominal load — nominal factional resistance per eq. in. of bearing. NOMINAL FRICTIONAL RESISTANCE per sq. in. of bearing.

Nominal LOAD Lbs. per sq. in.

Actual LOAD Lbs. per sq. in.

Lbs. 258

Lbs. 317

Lb.

Lb. 1.43

Lb. 1.46

Lb. 1.61

Lb. 1.76

Lb.

205

252

2.71

2.01

1.43

1.57

1.68

1.79

100

123

1.44

1.25

1.46

1.52

1.63

1.71

100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 262 ft. 419 ft. 314 ft. 366 ft. 209 ft. 157 ft. 105 ft. per min. per min. per n.in. per min. per min. per min. per min.

Lb.

1.78

The The The The

chord of the arc of contact of the brass =31/4in. nominal load per sq. in. is the total load divided by 4 x 6. actual load per sq. in. is the total load divided by31/4x 6. bearing seized on attempting to run with an actual load of 380 1bs. per sq. in. With nominal load of 258 1bs. per sq. in. the temperature of the bearing was 90°.

170

"

"

"

"

"

"

205 205 100

„

11

,,

"

"

,,

,,

"

80o

85° •

650

Nov. 1883.

FRICTION EXPERIMENTS.

TABLE VIII.—RAPE OIL, PAD UNDER JOURNAL. 4-IN. JOURNAL, 6 IN. LONG. CHORD OF ARC OF CONTACT OF BRASS =21/2IN. Nominal LOAD Lbs. per sq. in.

Actual LOAD Lbs. per sq. in.

328 310 293 275 258 205 153 100

582 551 520 498 458 364 272 178

COEFFICIENTS OP FJRICTION, for speeds as below. Temperature. Fahr. 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 105 ft. 157 ft. 209 ft. 262 ft. 314 ft. 366 ft. 419 ft. per min. per min. per min. per min. per min. per min. per min. 90°

82° 76° 77° 78° 82° 74° 75°

•0099 •0105 •0091 •0112 •0105 •0102 •009 •0105 •0099

•0107 •0099 •0105 •0091 •0095 •0087 •0096 •0109

•0102 •0092 •0097 •0095 •0088 •0085 •0102 •0122

•0098 •0099 •0097 •0103 •0084 •0078 •0105 •0133

•0082 •0085 •0119 •0144

•0083 •01 .0125 •0154

The above coefficients x the nominal load = nominal frictional resistance per sq. in. of bearing.

Nominal LOAD Lbs. per sq. in.

Actual LOAD Lbs. per sq. in.

328

582 551 520 498 458 364 272 178

310

293 275 258 205 153 100

NOMINAL FPJCTIONAL RESISTANCE per sq. in. of bearing. Temperature. Fahr. 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 400 rev. 105 ft. 157 ft, 209 ft. 262 ft. 314 ft. 366 ft. 419 ft. per min. per min. per min. per min. per min. per min. per min. Lb.

Lb.

90°

82° 76° 77° 78° 82° 74° 75°

1.56 1.0

3.06 3.06 2.49 2.89 2.145 1.37 0.992

Lb. 3.5 3.06

Lb. 3.35

2.84 3.06 2.84 2.62 2.49 2.44 2.28 1.78 1.735 1.473 1.56 1.093 1.225

Lb. 3.21

3.06 2-84 2.84 2.17 1.605 1.605 1.33

Lb.

Lb.

2.1 1.75 1.81 1.44

2.13 2 04 1.89 1.54

The chord of the arc of contact of the brass — 21/4in. The nominal load per sq. in. is the total load divided by 4 x 6. The actual load per sq. in. is the total load divided by 2| x 6. The results with the actual load of 582 Ibs. per sq. in. were obtained with difficulty, and the bearing seized with that load after running for a short time. The pad consisted of a piece of felt pressing against the journal, and resting on worsted immersed in a tin box full of oil.

171

Nov. 1883.

651

FRICTION EXPERIMENTS.

TABLE IX.—BATH OF LARD OIL. VARIATION OF FRICTION WITH TEMPERATURE. NOMINAL LOAD 100 LBS. PER SQ. IN. Temperature Fahr.

COEFFICIEKTS OF FRICTION, for speeds as below. 100 rev. 150 rev. 200 rev. 250 rev. 300 rev. 350 rev. 40ft rev. 450 rev. 471 ft. 314 ft. 105 ft. 262 ft. 366 ft. 419 ft. 157 ft. 209 ft. per mm. per mm. per mm. per mm. per mm. per mm. per mm. per mm.

120°

•0024

•0029

•0035

•004

•0044

•0047

•0051

•0054

110°

•0026

•0032

•0039

•0044

•005

•0055

•0059

•0064

100°

•0029

•0037

•0045

•0051

•0058

•0065

•0071

•0077

90°

•0034

•0043

•0052

•006

•0069

•0077

•0085

•0093

80°

•004

•0052

•0063

•0073

•0083

•0093

•0102

•0112

70°

•0048

•0065

•008

•0092

•0103

•0115

•0124

•0133

60°

•0059

•0084

•0103

•0119

•013

•014

•0148

•0156

TABLE X.—COMPARISON OF THE FRICTION WITH THE DIFFERENT METHODS OF LUBRICATION, UNDER AS NEARLY AS POSSIBLE THE SAME CIRCUMSTANCES. LUBRICANT EAPE OIL, SPEED 150 REVOLUTIONS PER MIN. Actual LOAD Lbs. per sq. in.

172

Coefficient of Friction.

Comparative Friction.

Oil Bath

263

•00139

1

Syphon lubricator

252

•00980

7.06

Pad under journal

272

•00900

6.48

652

Nov. 1883.

FRICTION EXPERIMENTS.

TABLE XI.—COMPARISON OF THE FEICTION WITH THE VARIOUS LUBRICANTS TRIED, UNDER AS NEARLY AS POSSIBLE THE SAME CIRCUMSTANCES. TEMPERATURE 90°, LUBRICATION BY OIL BATH. Mean Resistance.

Per Cent.

Sperm Oil

Lb. 0.484

100

Rape Oil

0.512

106

Mineral Oil .

0.623

129

0.652

135

Olive Oil

0.654

135

Mineral Grease

1.048

217

Lubricant.

Lard O i l

. . . .

N.B.—The above figures (calculated from Tables I.-VI.) are the means of the actual frictional resistances at the surface of the journal per sq. in. of bearing, at a speed of 300 revs, per min., with all nominal loads from 100 Ibs. per sq. in. up to 310 Ibs. per sq. in. They also represent the relative thickness or body of the various oils, and also in their order, though perhaps not exactly in their numerical proportions, their relative weight-carrying power. Thus sperm oil, which has the highest lubricating power, has the least weight-carrying power; and though the best oil for light loads, would be inferior to the thicker oils if heavy pressures or high temperatures were to be encountered.

173

Pig. 1. of

Sectional Tesybing

Fig. 3.

.ElevationMachine.

Fig. 4.

Fig 1 A.

Second

arrandement

of

Index.

Figr. 2.

End

View of Testing Machine.

Final arrangement for Lnbricativn.

Fig. 9.

Fig. 10.

Scale I to 24.

Fig: II.

Fig. 14.E n l a r g e m e n tatA Fig. 12.

Fig . 5

Fig 7

Fig. 12

Fig. 8.

Fig. 13.

Hiram Stephens Maxim (1840-1916) This American-born and self-proclaimed "chronic inventor" acquired his engineering training by serving first as a carriage maker, then as a draughtsman to his uncle, Levi Stephens, in Fitchburg, Masschusetts. Following a relatively aimless career and various small inventions, Maxim became chief engineer to the first US electric lighting company. In this capacity, he attended the Paris Exhibition of 1881 and developed a taste for European life. He settled in London shortly afterwards, and opened a workshop in Hatton Garden. Maxim formed the Maxim Gun Company in 1884, the same year in which he joined the Institution of Mechanical Engineers. The company was chaired by Albert Vickers (1838-1919). After amalgamation with Nordenfeldt, an earlier machine-gun concern, in 1888, both later became Vickers, Sons and Maxim. Maxim's machine-gun was not a new idea - Nordenfeldt, Gatling and others had previously produced rapid-firing weapons - but the single barrel and automatic mechanism of Maxim's gun were major advances. The weapon was adopted by the British Army in 1889 and later by the Royal Navy, and ensured financial success for its originator. The gun was the subject of Maxim's only paper for the Institution's Proceedings. Other achievements included a contribution to the early history of aeronautics. From 1889, Maxim experimented with models and finally a full-sized flying machine. In 1894, this aerial monster of 105 foot wingspan briefly, and accidentally, left its guide rail at Bexley, Kent, to achieve a "hop" of powered flight. The weight of the machine - it was powered by steam engines, prevented its further success.

177

SHOWING THE GUN TO MY GRANDSON

APRIL 1885.

167

DESCRIPTION OF THE MAXIM AUTOMATIC MACHINE-GUN. BY MB. HIRAM S. MAXIM, OF LONDON.

Previous Machine-Guns.—The first practical Machine-Gun is believed by the writer to be that made by Dr. Gatling, an American inventor, by whom it was presented to the United States government about 1868. It is shown in general elevation in the accompanying sketch, Fig. 1, Plate 19. In its earliest form it was not made for firing metallic cartridges, but was constructed with a series of steel sections, which, after having been previously loaded by hand, were one after another brought up to the breech of the barrel and pressed firmly against it at the instant of firing. None of the Gatling guns however appear to have been actually used in the field before the close of the American war in 1865. The next machine-gun of any note was that of Hotchkiss, another American inventor, who took it to France, where he established a large factory, and has supplied his guns to nearly all the principal governments in the world. This gun was followed by the Nordenfelt and the Gardner, shown in general elevation in the sketches, Figs. 2 to 4, Plate 19. All four of these machine-guns depend upon hand power for performing the various operations of loading, firing, and extracting the empty shells. Three of them are worked by a crank, while the Nordenfelt gun is worked by means of a lever, like an ordinary pump. As considerable force is required for working either the crank or the lever, the gun has to be mounted on a very firm stand or base, in order that it may not be rendered unsteady by the motion given to the handle. This necessity precludes the possibility of turning these guns with any degree of freedom: excepting the Hotchkiss gun, which is essentially a slow-firing gun, firing only about forty shots per minute, and is the only one that can be moved freely while firing.

179

168

AUTOMATIC MACHINE-GUN.

APRIL 1885.

These guns are each provided with a magazine of ammunition. The Hotchkiss magazine holds about a dozen cartridges; and the quick-firing Gatling, the Nordenfelt, and the Gardner, have each a magazine holding about a hundred. The magazine is placed on the top of the gun, and with any great rapidity of firing has of course to be replenished very often : for which purpose two men at least are required, who are compelled to expose themselves above the gun, both the magazine and the men presenting a target to the enemy's fire. The workmanship of all four of these guns is exquisite. Their weak point does not lie here, but arises from another cause which would be very difficult to remedy in them. It is said by some military men that no machine-gun has ever been brought into action which has not become " jammed " at the critical moment. Even if that be not strictly true, still the liability to accident from this cause is very great. A certain percentage of all cartridges fail to explode promptly at the instant of being struck: to use the technical expression, they " hang fire." Suppose that, while the handle of the gun is being worked at its highest speed, one of these sluggish cartridges happens to enter the barrel. It is struck ; and instantly, before it explodes, the breech is opened, and the cartridge begins to be withdrawn again out of the barrel. At this instant the explosion takes place, breaks the shell in two, drives the front half forwards into the barrel, and blows the rear half out at the breech, and sometimes blows up the magazine. At any rate, it always drives the forward end of the cartridge firmly into the chamber of the barrel; and if the magazine does not explode, the next rotation of the crank drives a loaded cartridge into this chamber; the gun then becomes blocked or jammed, and is of no further use. Maxim Automatic Gun.—It is many years since the writer conceived the idea of making a gun in which the recoil should be utilised for loading and firing; but it was not until 1883 that he had any time to devote to this problem. Before commencing experiments he considered carefully the different methods which might be employed for working an automatic gun by means of power derived

180

APRIL 1885.

AUTOMATIC MACHINE-GUN.

169

from the burning powder. In those which he afterwards experimented upon, the power required was derived in the six following ways:— (1) power derived from the gases escaping from the muzzle of the gun, either by utilising their pressure directly, or by employing them as an ejector to produce a vacuum in a chamber near the muzzle of the gun; (2) power derived from the recoil of the entire gun; (3) power derived from the recoil of the barrel, the breech-block, and the lock; (4) power derived from a backward motion of the cartridge in the chamber at the instant of exploding; (5) power derived from only a portion of the cartridge moving backwards; (6) power derived from the elongation of the cartridge at the instant of exploding. Experimental Apparatus.—The apparatus is now exhibited that was used for conducting experiments on a gun in which the power was derived according to the third of the above methods—namely from the recoil of the barrel, the breech-block, and the lock. As the writer was the first to make a gun of this kind, he had no data whatever to go upon, and had therefore to contrive some kind of device for ascertaining both the quantity and the character of the power to be dealt with. This apparatus consists of two parallel steel bars, clamped into supports, and having the barrel and the breech-block mounted between them. The whole is so constructed that all of the parts are adjustable. The distance through which the barrel recoils, before the breech-block becomes detached from it, is adjustable ; the further distance that the barrel travels backwards, after the block becomes detached from it, is also adjustable; the travel given to the striker is adjustable; the angle at which the crank stands at the instant of explosion is adjustable ; the amount of weight in the rotating parts and their distance from the centre of rotation are also adjustable. This experimental apparatus was made as far as possible of soft steel and brass, in order that the action of the gunpowder might be observed upon the various parts. If any part showed signs of yielding under the strain, it was obvious that this part required strengthening in the gun ultimately to be made. The apparatus has already fired about a thousand cartridges, and at the present time is still in

181

170

AUTOMATIC MACHINE-GUN.

APRIL 1885.

condition to be fired. With one hand on the muzzle the barrel can be pushed back with sufficient force to perform the whole cycle of operations for firing, the push of the hand taking the place of the recoil of the barrel. In this way it has been found that a pressure of about 60 1bs. travelling through a distance of seven-eighths of an inch is the power required for working the gun : which is very much less than the actual power derived from the recoil, as determined approximately by the writer in experiments with a Winchester rifle of the " express " pattern. First Gun.—Having in this way determined the character and quantity of the power to be dealt with, and having ascertained the time required for the gas to escape from the barrel, as well as the strength required for the several parts, and the distance through which they had to travel, the writer proceeded to make his first gun, the construction of which is shown in the front and side elevations, Figs. 5 and 6, Plates 19 and 20, and in detail to a larger scale in Figs. 7 to 33, Plates 21 to 25. The barrel A, Figs. 7 to 10, Plates 21 and 22, is encased throughout the greater portion of its length within a water-jacket, Fig. 10, which projects from the front end of the casing that encloses the machinery. Its backward motion, produced by the recoil of a shot just fired, is at first opposed by a pair of flat springs, one at each side, Fig. 8, which have to be forced apart by toggle struts TT recessed into the sides of the barrel, as shown in the plans, Figs. 12 to 14, Plate 23. As soon as these struts have passed the centre, the springs close together again, and thereby aid the further backward travel of the barrel. During the first half of the seveneighths inch backward travel of the barrel, it carries the breechblock B with it at the same speed, the two being secured fast together by the locking catch C, Fig. 7, which is held firmly down under the crossbar D. But when the barrel has moved backwards through 7-16ths inch, the catch is free from under the crossbar, Fig. 9, and is lifted by a projection on its front end coming in contact with the face of the- crossbar, thereby releasing the breech-block from the barrel. At the same instant a straight lever L centered

182

APRIL 1885.

AUTOMATIC MACHINE-GUN.

171

upon the barrel, Fig. 8, encounters with one of its arms a stop S fixed in the casing, while its other arm bears against the toe of a sliding rod or striking bar R, the rear end of which bears against the breech-block B. Both the stop S and the toe of the bar R are made with long bearing faces slightly curved, as shown one-third full size in Figs. 12 to 14, Plate 23; and as the slightly curved arms of the transfer lever L roll over them, the respective leverages change very rapidly in relation to each other, and become inverted, with the result of imparting a rapidly accelerated motion to the breech-block. Consequently while the barrel A travels through the remaining 7-16ths inch of its backward motion, the breech-block B is driven backwards with sufficient force to carry the crank K over the back centre, Fig. 9; the radius of the crank is 3 inches, and the connectingrod G from the breech-block is 6 inches long. The action of the transfer lever L in accelerating the breech-block has also the converse effect of simultaneously retarding the backward travel of the barrel; and all the backward motions cease at the instant of the crank passing the back centre, Fig. 9. With freshly loaded cartridges the momentum given to the crank and its attachments is found sufficient to drive the breech-block forwards again into its firing position and to fire the next shot. But with old and weaker cartridges the gun is found to work with greater certainty if a strong helical spring P is used, Fig. 8, to assist in drawing the breech-block forwards again into its firing position after the crank has passed the back centre. Consequently all guns made after the very first experimental gun are now provided with this spring. Immediately upon the breech-block quitting the barrel, the tail of the extractor E, Fig. 9, Plate 22, and Figs. 32 and 33, Plate 25, which is a forked lever centered upon the barrel, comes against a stop fixed in its path; and the forked end of the lever, which takes hold of the cartridge rim or flange at each side, withdraws the empty cartridge shell about1/4inch out of the barrel, Fig. 9. Its extraction is then completed by a hook I, Figs. 9 and 29, attached to the same crosshead as the breech-block. The hook runs underneath a pair of long fixed springs F, Figs. 7, 9, and 18, by which it is pressed

183

172

AUTOMATIC MACHINE-GUN.

APRIL 1885.

down upon the cartridge so as to keep a secure Hold while extracting it; but at each end the springs are curved slightly upwards, in order to reduce the pressure upon the hook in its foremost position, where it has to lift for catching hold of the cartridge flange, as well as in its hindmost position, where it has to be lifted again by the fixed incline J, Fig. 29, for releasing its hold of the cartridge. In this way the empty cartridge-case is drawn back, Fig. 30, into one of the grooves or pockets in the rim of the magazine or feeding cylinder M, Fig. 9, which is mounted upon an axis immediately beneath and parallel with the line of travel of the breech-block. The magazine M, Figs. 18 to 20, Plate 24, is rotated intermittently by an arrangement of spiral ratchet-wheel and paul, shown developed flat in plan in Fig. 28, Plate 25, which is somewhat similar to the arrangement commonly employed in rock-drills for rotating the drill automatically between each blow : the motion of the breech-block in the last part of its backward travel, when the empty cartridge is entirely drawn out of the barrel, rotates the cylinder M through half the pitch or distance to the next groove or pocket, which has already been charged with a fresh cartridge; and the first part of the forward travel rotates it through the remaining half, bringing the fresh cartridge into the exact line of the barrel before the front end of the cartridge has reached the rear end of the barrel. The grooves or pockets in the magazine are charged one at a time in each back-stroke, from a belt of cartridges that passes over a flanged wheel W, Figs. 7, 9, and 17, Plates 21 to 24, which is situated in front of the magazine M, and is geared to it. The flanged wheel W has recesses in each flange for the ends of the cartridges to lie in. A hook or extractor H carried on the crosshead of the breech-block catches a cartridge in the backward travel of the breech-block, and draws it out of the belt into one of the pockets on the underside of the magazine M, where it remains while carried upwards step by step to the barrel by the intermittent rotation of the magazine. The empty belt passes out through an opening in the left-hand side of the gun casing, Figs. 17 and 20, Plate 24; and through an upper opening on the same side, Fig. 20, the empty cartridges drop out one by one from the pockets of the magazine as it rotates. During the

184

APRIL 1885.

AUTOMATIC MACHINE-GUN.

173

backward travel of the breech-block a pivoted cover-plate N, Figs. 8, 18, 19, and 31, Plates 21 to 25, is thrown across laterally over the magazine by the pressure of a coiled spring, in order to prevent any risk of the cartridge being jerked upwards out of the magazine. As the crank approaches the back centre, towards the end of the backward travel of the breech-block, the tail of the cocking lever O, Fig. 9, Plate 22, which is pivoted upon the crosshead, comes against a fixed stud; in the remainder of the backward travel the cocking lever then compresses the main spring, which is a helical spring coiled round the striking pin; and a suitable catch or sear Q, also hinged upon the crosshead, finally catches the nose of the cocking lever, and holds the striker cocked in readiness for firing the next shot. The second half of the cycle of operations comprises those which are effected by the forward travel of the breech-block with its connections; and consists in pushing the fresh cartridge home into the barrel, locking the breech-piece, and releasing the sear Q for firing the shot. As soon as the crank has passed the back centre, it begins to push the breech-block forwards, with the fresh cartridge in front of, it; and through the transfer lever L centered upon the barrel, Fig. 8, Plate 21, the quick travel of the breech-block B imparts a slow forward travel to the barrel A, sufficient to carry it forwards until the toggle struts TT pass the centre, and the flat side-springs are then in a position to urge it forwards to the end of its travel with the breech-block locked fast against it. In this final travel of 7-16ths inch under the action of the side springs, the sear Q coming into contact with a cam U, Fig. 7, releases the striker, which fires the cartridge. The cam is connected with an ordinary cataract or hydraulic buffer V, Fig. 7, of which the by-pass is throttled by an adjustable plug, Figs. 21 to 25, Plate 25 ; a hand-lever on the plug regulates the rate of firing so as to deliver any number of shots from two or three per minute up to as many as six hundred under favourable conditions. The handle on the cataract plug serves as a trigger for firing the gun by hand; for if the by-pass be opened while the gun is loaded, the explosion follows instantly; and

185

174

AUTOMATIC MACHINE-GUN.

APRIL 1885.

if it be entirely closed, the gun though loaded cannot be fired at all. The crank-shaft K is fitted with a handle Z outside the casing and opposite to the crank, Figs. 6 and 8, Plates 20 and 21, by which the gun is worked by hand at starting, until the first shot has been fired; the recoil then comes into action for continuing the firing automatically. The gun is 4 feet 9 inches long over all, from the muzzle to the rear of the casing that contains the firing mechanism, Fig. 6, Plate 20. It stands about31/2feet high upon its tripod. The belt supplying the cartridges is made of two lengths of canvas, Figs. 26 and 27, Plate 25, riveted together at regular intervals with brass eyelets and strips, so as to form a succession of loops, into each of which a cartridge is inserted by hand. When any belt is running out, a fresh one is hooked on to its tail end, without causing any delay to the continuous firing of the gun. By means of a simple appliance attached to the muzzle of the gun, Fig. 34, Plate 26 (see page 192), the smoke can be deflected in any direction desired, sideways or upwards, so as to give the gunner at all times a clear view of his aim. A gun of rifle calibre can also be made practically noiseless. The simple water-jacket encasing the barrel of the gun is found to answer very well for preventing excessive heating of the barrel, as the amount of heat required to evaporate water is so very great. Working of Gun.—The gun is mounted here upon a tripod, Figs. 5 and 6, Plates 19 and 20, which is a very convenient stand for exhibiting its working. The tripod of course is not a necessary part of the gun, the mounting of which depends altogether upon whether it is wanted in the field or on shipboard. The crank handle Z projecting at the rear from the right-hand side of the gun is necessary in order to work the crank for bringing the first cartridge into the barrel of the gun; also for removing from the barrel any cartridge which may have failed to explode. The cartridges for supplying the gun are placed in a box B beneath,

186

APRIL 1885.

AUTOMATIC MACHINE-GUN.

175

which may be made large enough to contain almost any number. In a light field-carriage about 2000 would be a fair supply; the box here shown holds 333, or one-sixth of that supply. In turning the crank handle forwards and backwards, the feeding wheel and magazine are moved forwards one tooth at each turn: so that when the end of a belt of cartridges is introduced into the feed chamber, one cartridge is drawn in at each turn of the crank handle. Seven turns will draw, seven cartridges out of the belt, and there will then be six cartridges in the magazine of the gun, and one in the barrel. On pulling the trigger, the cartridge in the barrel will explode, and its empty shell will be expelled from the breech of the barrel, which will instantly receive a fresh cartridge from the magazine; and the magazine in its turn will be supplied with another cartridge from the belt of cartridges. When a cartridge enters the barrel, and the breech-block presses it firmly home, the block closes the breech of the barrel securely, the two being firmly locked together during the explosion of the cartridge, Fig. 7, Plate 21; and the breech cannot be opened again until the barrel, together with all its attachments which participate with it in the recoil, has moved backwards through 7-16ths of an inch, being half of its total travel of seven-eighths of an inch. By the time they have moved back through this distance, 7-16ths inch, the shot is already some distance out of the barrel, and the pressure of the gases is so far reduced that it is quite safe to open the breech. As the barrel recoils further, it becomes detached from the other moving parts; and while it is stopping, the breech-block with its attachments is sent rapidly backwards to a still greater distance. This further backward movement removes the empty shell from the barrel, and cocks the hammer; and then the return or forward travel of the breech-block pushes a fresh cartridge into the barrel, closes the breech, and pulls the trigger. Seven cartridges having been drawn from the belt into the magazine of the gun, there are now six of them in the magazine, and one in the barrel. The regulating lever R of the trigger, Fig. 6, Plate 20, has been set forwards into such a position as will cause the gun to fire very slow. (Seven shots were here fired slow.) In

187

176

AUTOMATIC MACHINE-GUN.

APRIL 1885.

this slow firing the whole of the various operations take place very rapidly, with the single exception of finally pulling the trigger for firing the shot. The speed of firing or rapidity of pulling the trigger is regulated by a piston working in the controlling cylinder or cataract, Fig. 24, Plate 25, which has an adjustable by-pass from one end to the other; the area of this passage way is regulated by a cock, and the cock is attached to the trigger lever. When the seven cartridges were fired very slowly, it was because the trigger lever R, Fig. 6, was pulled back only far enough to open the cock very slightly. If the lever be pulled further back, the speed of firing will be greater; while if it be pulled completely back against its stop, the whole seven cartridges will be fired in less than one second of time. When the barrel with its attachments has recoiled through half its travel, and has thus reached the point at which the breech-block becomes detached from it, the remaining half of its travel has then to be occupied in bringing the barrel itself to rest, while at the same time withdrawing the breech-block rapidly from it through a much longer range of backward travel. This is accomplished by the single transfer-lever L, Plate 23, so contrived as to afford a great range of variation in its effective leverage. The breech-block B is thereby driven rapidly backwards at the expense of stopping the barrel A completely. Upon the return or forward movement of the block, the impact which would be expected from its closing against the end of the barrel is prevented by this same lever arrangement, whereby the barrel is caused to be already moving forwards in the same direction as the block before they come in contact, thus preventing any destructive hammering between them. In Fig. 36, Plate 26, is shown, one-third full size, a diagram of the relative movements of the barrel and of the breech-block, set out upon a base-line KK which represents the length of the circular path described by the crank. The darker shaded area shows the short travel of the barrel; and the lighter area shows where the breech-block parts company with the barrel at one of the points P, and how it then continues its much longer travel while the barrel is stationary, and how it afterwards overtakes and rejoins the barrel

188

APRIL 1885.

AUTOMATIC MACHINE-GUN.

177

at the other point P in the return or forward travel. Below the actual base-line KK is added also a dotted line DD, for indicating what would have been the virtual base-line, had the crank been arranged to make a complete revolution: instead of which it is a fundamental feature of this gun that the crank stops short on each side of its front centre, and rotates forwards and backwards alternately through somewhat less than a whole circle, as indicated in Figs. 6 and 7, Plates 20 and 21. On introducing a belt of cartridges into the gun, and turning the crank handle, the cartridges are drawn in one by one, until the magazine is full, that is, until it is filled round half its circumference up to the barrel, and empty round the other half; and the empty part of the belt hangs out from the opposite side of the gun, Pigs. 17 and 20, Plate 24. On pulling the trigger by hand, the first one of these cartridges is fired; and the gun will then supply itself from the belt and continue firing automatically as long as there are any cartridges in the belt. The firing can be stopped after only a single cartridge has been fired, or after two, three, or any other number up to a whole volley have been fired; and the gun can be made to fire either slow or with great rapidity. It is the first shot alone that requires to be fired by hand, after which the firing goes on automatically until stopped. Hang Fire.—Suppose that any of the cartridges supplied to the gun should hang fire. At the instant of the cartridge being struck the breech is closed, and there is no power to open it automatically except the power locked up in the cartridge itself. This power does not develop itself until the cartridge itself explodes; consequently the breech does not open until after it has exploded, and the cartridge cannot be withdrawn in the act of exploding, as it can in other guns. Thus the serious trouble occurring with all other machine-guns in cases of hang fire cannot here occur, being rendered impossible by the very principle upon which the gun is constructed. The case of a cartridge entering the barrel and failing to go off at all may be exemplified by placing in the magazine of the gun two good cartridges, then a bad one, and then two good ones. As

189

178

AUTOMATIC MACHINE-GUN.

APRIL 1885.

soon as the first cartridge is in the barrel and has been fired by pulling the trigger by hand, the gun will fire the second cartridge automatically, and will attempt to fire the third or bad cartridge, but will fail. The bad cartridge has then to be removed from the barrel by working the crank handle by hand; the fourth cartridge has, like the first, to be fired by hand; after which the gun will fire the fifth automatically. The whole operation of passing the bad cartridge will occupy about half a second. Hence, if a cartridge hangs fire, the gun waits for it; if it fails to go off at all, it must be removed by hand, which is done in about half a second. Adjustment.—As the gun requires no external power for working it, being wholly self-contained, it may of course be turned freely in any direction while firing. For target practice, and for accurate shooting at long range, it is convenient to train it with screws, as in the case of other machine-guns; and suitable adjusting screws are accordingly provided, as shown in Figs. 5 and 6, Plates 19 and 20. Suppose however that it be desired to give a quick adjustment as to elevation. For this purpose the clamping screw C is slackened on the telescopic elevating strut E, Fig. 6, Plate 20, and in a second the required elevation is approximately given; after which the clamping screw C is tightened up, and a final true adjustment is given in the ordinary manner by the fine regulating screw S. Again if it be desired to spread the fire horizontally between two given points, as across a bridge or a ford or a pass;—or to take accurate aim in the daytime upon a part of the enemy's position where he is expected to be at work at night;—in such cases the horizontal adjustment towards the right-hand side is made with one adjustable stop, and towards the left-hand side with another. The two stops can be set so as to give any spread that is wished, and to give it in any position desired. When so adjusted the gun can remain in this position till wanted, and can be fired in the night or at any other time with the certainty of covering everything between these two points without getting out of adjustment. For receiving cavalry or a sudden charge, the gun is clamped in elevation, but is left

190

APRIL 1885.

AUTOMATIC MACHINE-GUN.

179

free to be moved completely round horizontally. For firing upon a quickly-moving torpedo boat, it is wholly undamped so as to be moved freely in all directions. The gun here shown is the size known as rifle calibre, which would doubtless be of great service in the field. Naval Gun.—For naval purposes a very much larger gun is required, having sufficient power to destroy a torpedo boat completely at a single shot. For this purpose the writer is now making a gun large enough to take a cartridge" with a shell of 17/8 inch diameter and 6 inches length, the shell or projectile having a percussion fuse and sufficient powder to give great penetration and a very long range, with a muzzle velocity of 2000 feet per second. This gun will be capable of being fired at the rate of about 150 shots per minute; but as the time that a torpedo boat will be in sight before she is near enough to destroy the ship is only from about ten to twenty seconds, it is not expected that the gun will ever have to be worked during a whole minute at its full speed. What is required is to fire a very large number of shots in a very short space of time. A six-pounder capable of piercing about 4 inches of steel plate could be fired about 50 times per minute.

Discussion.

Mr. MAXIM showed the experimental apparatus described in the paper (p. 169), and the first gun made (p. 170); and also the last gun made, which had just been finished and mounted on its tripod stand in readiness to be shown at the Inventions Exhibition. He illustrated the description given in the paper by firing the first gun under the several conditions of slow and quick fire, and hang fire; and showed the handling and adjustment of the gun for meeting all the requirements arising in its use. Y

191

Fig. 1.

Gatlinq

Gun.

Fig. 2. Nordenfelt Field

Gun.

Fig.3. Nordenfelt Marine

Gun Fig. 5.

Fig. 4. Gardner Gun .

192

Maxiun

Automatic Gun.

AUTOMATIC MACHINE-GUN. Plate 20. Fig. 6. Side Elevation of

Maxin

Gun.

193

Fig. 7.

Longitudinal Section in position of firing.

Fig. 8. Sectional Plan.

Fig. 9. Longitudinal Section showing Breech-Block fully back.

Fig. 10. Section

along

Water - Jacket

encasing Barrel.

Fig.12.

Fig.15. Transverse Section at Transfer Lever.

196

Fig.13.

Fig.16

Fig.14.

Fig. 17. Section at XX (Fig.7)

Fig. 18. Section

Fig. 19.

Fig. 20.

Fig.21.

Plate 25. Fig. 29. Extracting Hook, on Breech -Block.

Fig. 22. Transverse Section. Spiral Ratchet - Wheel developed flat.

Cataract.

Belt of Scale

1

/3rd.

Cartridges. Fig. 26. Plan.

Fig.28. Fig. 30. Inverted Plan showing Stop for extracted Carbridge -case.

Fig.23.Plan

Full size Section. Fig. 24.

Fig. 27 Section.

Fig. 25. End View of Piston.

Extracting Lever on Barrel.

Fig. 32.

Fig. 31.

Fig. 33.

Fig 36. Relative Travel of Breech -Block

and Barrel. Scale

Fig. 35. Fig. 34. Section of Smoke- deflecting Cap on muzzle of gun.

Escape of Smoke from gun muzzle.

1

/3rd

This page intentionally left blank

William Thomson (1824-1907), 1st Baron Kelvin of Largs Recognised as one of the 19th century's great men of science, William Thomson was born into a distinguished scientific family. His father was a professor of mathematics and his elder brother a professor of engineering. However inevitable his own rise to academic prowess would appear to be, few would have predicted that his singular talent would see him matriculated at Glasgow University in 1834 at the tender age of ten. By the age of 22 he had gained the Chair of Natural Philosophy, having already produced a considerable body of work. In 1847, Thomson began his famous collaboration with James Prescott Joule (1818-1889) on the relationship between mechanical power, heat, and electricity. This work culminated, in part, with Thomson outlining the first and second laws of thermodynamics (equivalence and transformation) to the Royal Society of Edinburgh (1851-1854). He also introduced the absolute scale of temperature (the Kelvin scale). In addition, Thomson conducted fundamental research in electromagnetic field theory. His paper On transient electric currents (1853) laid the foundations for the study of electrical oscillations. Thomson was not merely a theoretician however. The practicalities of wireless telegraphy were encapsulated within this paper and Thomson became involved in laying submarine cables from 1858. The success of subsequent transatlantic schemes was largely attributable to Thomson's energy. He invented instruments to facilitate this work, including the mirror galvanometer. The inventor's love of the sea became intertwined with his professional labours. Thomson sailed his own yacht, the Lalla Rookh, was proud of his navigational skills and made substantial improvements to the maritime compass. His later researches, in vortex theory, tides, and waves were the scientific manifestations of these pleasures. Greatly honoured, as President of the Royal Society and holder of the Order of Merit, Thomson retained a touching dedication to first principles: he declared "I can never satisfy myself until I can make a mechanical model of a thing...as long as I cannot...I cannot understand".

201

AUG. 1887.

409

ON SHIP WAVES. BY SIR WILLIAM THOMSON, Kt., F.R.SS. L. AND E., LL.D., PKESIDENT R.S.E. Lecture delivered at the Conversazione in the Science and Art Museum, Edinburgh, on Wednesday evening, 3rd August 1887.

"Waves" is a very comprehensive word. It comprehends waves of water, waves of light, waves of sound, and waves of solid matter such as are experienced in earthquakes. It also comprehends much more than these. " Waves " may be defined generally as a progression through matter of a state of motion. The distinction between the progress of matter from one place to another, and the progress of a wave from one place to another through matter, is well illustrated by the very largest examples of waves that we have —largest in one dimension, smallest in another—waves of light, waves which extend from the remotest star, at least a million times as far from us as the sun is. Think of ninty-three million million miles, and think of waves of light coming from stars known to be at as great a distance as that! So much for the distance of propagation or progression of waves of light. But there are two other magnitudes concerned in waves : there is the wave-length, and there is the amount of displacement of a moving particle in the wave. Waves of light consist of vibrations to and fro, perpendicular to the line of progression of the wave. The length of the wave—I shall explain the meaning of " wave-length " presently: it speaks for itself in fact if we look at waves of water, as shown in Fig. 2, Plate 80—the length from crest to crest in waves of light is from one thirty-thousandth to one fifty-thousandth or one sixty-thousandth of an inch ; and these waves of light travel through all known space. Waves of sound differ from waves of light in the vibration of the moving particles being along the line 2 N

203

410

SHIP WAVES.

AUG. 1887.

of propagation of the wave, instead of perpendicular to it. Waves of water agree more nearly with waves of light than do waves of sound; but waves of water have this great distinction from waves of light and waves of sound, that they are manifested at the surface or termination of the medium or substance whose motion constitutes the wave. It is with waves of water that we are concerned to-night; and of all the beautiful forms of water waves that of Ship Waves is perhaps the most beautiful, if we can compare the beauty of such beautiful things. The subject of ship waves is certainly one of the most interesting in mathematical science. It possesses a special and intense interest, partly from the difficulty of the problem, and partly from the peculiar complexity of the circumstances concerned in the configuration of the waves. Canal Waves.—I shall not at first speak of that beautiful configuration or wave-pattern, which I am going to describe a little later, seen in the wake of a ship travelling through the open water at sea; but I shall, as included in my special subject of ship waves, refer in the first place to waves in a canal, and to Scott Russell's splendid researches on that subject, made about the year 1834—fifty-three years ago—and communicated by him to the Royal Society of Edinburgh. The diagrams copied from his paper in the Transactions of that Society will serve for a preliminary explanation or illustration of the meaning of the term " wave." I gave a very general and abstract definition; let us now have it in the concrete: a wave of water produced by a boat dragged along a canal. In Plate 80 is reproduced one of Scott Russell's pictures illustrating some of his celebrated experiments. In Fig. 2 is shown a boat in the position that he called behind the wave ; and in the rear of the boat is seen a procession of waves. It is this procession of waves that we have to deal with in the first place. We must learn to understand the procession of waves in the rear of the canal boat, before we can follow, or take up the elements of, the more complicated pattern which is seen in the wake of a ship travelling through open water at sea. Scott Russell made a fine discovery in the course of those experiments. He found that it

204

AUG. 1887.

SHIP WAVES.

411

is only when the speed of the boat is less than a certain limit that it leaves that procession of waves in its rear. Now the question that I am going to ask is, how is that procession kept in motion ? Does it take power to drag the boat along, and to produce or to maintain that procession of waves? We all know it does take power to drag a boat through a canal; but we do not always think on what part of the phenomena, manifested by the progress of the boat through the canal, the power to drag the boat depends. I shall ask you for a time to think of water not as it is, but as we can conceive a substance to be—that is, absolutely fluid. In reality water is not perfectly fluid, because it resists change of shape; and non-resistance to change of shape is the definition of a perfect fluid. Is water then a fluid at all? It is a fluid because it permits change of shape; it is a fluid in the same sense that thick oil or treacle is a fluid. Is it only in the same sense ? I say yes. Water is no more fluid in the abstract than is treacle or thick oil. Water, oil, and treacle, all resist change of shape. When we attempt to make the change very rapidly, there is a great resistance; but if we make the change very slowly, there is a small resistance. The resistance of these fluids to change of shape is proportionate to the speed of the change: the quicker you change the shape, the greater is the force that is required to make the change. Only give it time, and treacle or oil will settle to its level in a glass or basin, just as water does. No deviation from perfect fluidity, if the question of time does not enter, has ever been discovered in any of these fluids. In the case of all ordinary liquids, anything that looks like liquid and is transparent or clear — or, even if it is not transparent, anything that is commonly called a fluid or liquid— is perfectly liquid in the sense of exerting no permanent resistance to change of shape. The difference between water and a viscous substance, like treacle or oil, is defined merely by taking into account time. Now for some motions of water (as capillary waves), resistance to change of shape, or as we call it viscosity, has a paramount effect; for other cases viscosity has no sensible effect. I may tell you this—I cannot now prove it, for my function this 2N2

205

412

SHIP WAVES.

AUG. 1887.

evening is only to explain and bring before you generally some results of mathematical calculation and experimental observation on these subjects—I may tell you that great waves at sea will travel for hours or even for days, showing scarcely any loss of sensible motion—or of energy, if you will allow me so to call it— through viscosity. On the other hand, look at the ripples in a little pond, or in a little pool of fresh rain water lying in the street, which are excited by a puff of wind; the puff of wind is no sooner gone than the ripples begin to subside, and before you can count five or six the water is again perfectly still. The forces concerned in short waves such as ripples, and the forces concerned in long waves such as great ocean waves, are so related to time and to speed that, whereas in the case of short waves the viscosity which exists in water comes to be very potent, in the case of long waves it has but little effect. Allow me then for a short time to treat water as if it were absolutely free from viscosity—as if it were a perfect fluid; and I shall afterwards endeavour to point out where viscosity conies into play, and causes the results of observation to differ more or less—very greatly in some cases, and very slightly in others—from what we should calculate on the supposition of water being a perfect fluid. If water were a perfect fluid, the velocity of progression of a wave in a canal would be smaller the shorter the wave. That of a " long wave "—whose length from crest to crest is many times the depth of the canal — is equal to the velocity which a body acquires in falling from a height equal to half the depth of the canal. For brevity we might call this height the "speed height"—the height from which a body must fall to acquire a certain speed. Examples : a body falls from a height of 16 feet, and it acquires a velocity of eight times the square, root of the height, or 32 feet per second; a body falls from a height of 4 feet, the velocity is therefore only 16 feet per second; and so on. Thus in a canal 8 feet deep the natural velocity of the "long wave" is 16 feet per second, or about 11 miles per hour. If water were a perfect fluid, this would be the state of the case: a boat dragged along a canal at any velocity less than the natural

206

AUG. 1887.

SHIP WAVES.

413

speed of the long wave in the canal would leave a train of waves behind it of so much shorter length that their velocity of propagation would be equal to the velocity of the boat; and it is mathematically proved that the boat would take such a position as is shown in Fig. 2, Plate 80, namely just on the rear slope of the wave. It was not by mathematicians that this was found out; but it was Scott Russell's accurate observation and well devised experiments that first gave us these beautiful conclusions. To go back: a wave is the progression through matter of a .state of motion. The motion cannot take place without the displacement of particles. Vary the definition by saying that a wave is the progression of displacement. Look at a field of corn on a windy day. You see that there is something travelling over it. That something is not the ears of corn carried from one side of the field to the other, but is the change of colour due to your seeing the sides or lower ends of the ears of corn instead of the tops. A laying down of the stalk is the thing that travels in the wave passing over the corn field. The thing that travels in the wave in Fig. 2, Plate 80, is an elevation of the water at the crest and a depression in the hollow. You might make a wave thus. Place over the surface of the water in a canal a wave-form, made from a piece of paste-board or of plastic material such as gutta-percha that you can mould to any given shape; and take care that the water fills up the wave-form everywhere, leaving no bubbles of air in the upper bends. Now you have a constant displacement of the water from its level. Now take your gutta-percha form, and cause it to move along — drag it along the surface of the canal—and you will thereby produce a wave. That is one of the best and most convenient of mathematical ways of viewing a wave. Imagine a wave generated in that way; calculate what kind of motion can be so generated, and you have not merely the surface motion produced by the force that you applied, but you have the water-motion in the interior. You have the whole essence of the thing discovered, if you can mathematically calculate from a given motion at the surface what is the motion that necessarily follows throughout the interior; and that can be done,

207

414

SHIP WAVES.

AUG.

1887.

and is a part of the elements of the mathematical results which I have to bring before you. Now to find mathematically the velocity of progression of a free wave, proceed thus. Take your gutta-percha form and hold it stationary on the surface of the water; the water-pressure is less at the crest and greater at the hollow; by the law of hydrostatics, the deeper down you go, the greater is the pressure. Move your form along very rapidly, and a certain result, a centrifugal force, due to the inertia of the flowing water, will now cause the pressure to be greatest at the crest and least at the lowest point of the hollow, Fig. 2, Plate 80. Move it along at exactly the proper speed, and you will cause the pressure to be equal all over the surface of the gutta-percha form. Now have done with it. We only had it in imagination. Having imagined it and got what we wanted out of it, discard it when moving at exactly this proper speed, and then you have a free wave. That is a slight sketch of the mode by which we investigate mathematically the velocity of the free wave. It was by observation that Scott Russell found it out; and then there was a mathematical verification, not of the perfect theoretic kind, but of a kind which showed a wonderful grasp of mind and power of reasoning upon the phenomena that he had observed. But still the question occurs to everybody who thinks of these things in an engineering way, how does that procession require work to be done to keep it up ? or does it require work to be done ? May it not be that the work required to drag the boat along the canal has nothing to do with the waves after all ? that the formation of the procession of waves once effected leaves nothing more to be desired in the way of work ? that the procession once formed will go on of itself, requiring no work to sustain it? Here is the explanation. The procession has an end. The canal may be infinitely long, the time the boat may be going may be as long as you please; but let us think of a beginning—the boat started, the procession begun to form. The next time you make a passage in a steamer, especially in smooth water, look behind the steamer, and you will see a wave or two as the steamer gets into motion. As it goes faster and faster, you will see a wave-pattern spread out; and if you were on shore,

208

AUG. 1887.

SHIP WAVES.

415

or in a boat in the wake of the steamer, you would see that the rear end of the procession of waves follows the steamer at an increasing distance behind. It is an exceedingly complicated phenomenon, and it would take a great deal of study to make out the law of it merely from observation. In a canal the thing is more simple. Scott Russell however did not include this in his work. This was left to Stokes, to Osborne Reynolds, and to Lord Rayleigh. The velocity of progress of a wave is one thing; the velocity of the front of a procession of waves, and of the rear of a procession of waves, is another thing. Stokes made a grand new opening, showing us a vista previously unthought of in dynamical science. As was his manner, he did it merely in an examination question set for the candidates for the Smith prize in the University of Cambridge. I do not remember the year, and I do not know whether any particular candidate answered the question; but this I know, that about two years after the question was put Osborne Reynolds answered it with very good effect indeed. In a contribution to the Plymouth Meeting of the British Association in 1877 (see " Nature" 23 Aug. 1877, pages 343-4), in which he worked out one great branch at all events of the theory thus pointed out by Stokes, Reynolds gave this doctrine of energy that I am going to try to explain; and a few years later Lord Rayleigh took it up and generalised it in the most admirable manner, laying the foundation not only of one part, but of the whole, of the theory of the velocity of groups of waves. The theory of the velocity of groups of waves, on which is founded the explanation of the wave-making resistance to ships whether in a canal or at sea, I think I have explained in such a way that I hope every one will understand the doctrine in respect to waves in a canal; it is more complex in respect to waves at sea. I shall try to give you something on that part of the subject; but as to the dynamical theory, you will see it clearly in regard to waves in a canal. If this drawing, Fig. 2, Plate 80, were continued backwards far enough, it would show an end to the procession of waves in the rear of the boat; and the distance of that end would depend on the time the boat had been travelling. You will

209

416

SHIP WAVES.

AUG. 1887.

remember that we have hitherto been supposing water to be free from viscosity; but in reality water has enough of viscosity to cause the cessation of the wave procession at a distance corresponding to 50 or 60 or 100 or 1000 wave-lengths in the rear of the ship. In a canal especially viscosity is very effective, because the water has to flow more or less across the bottom and up and down by the banks; so that we have not there nearly the same freedom that we have at sea from the effects of viscosity in respect to waves. The rear of the procession travels forward at half the speed of the ship, if the water be very deep. What do I mean by very deep ? I mean a depth equal to at least one wave-length; but it will be nearly the same if the depth be three-quarters of a wave-length. For my present purpose, in which I am not giving results with minute accuracy, we will call very deep any depth more than three-quarters of the wave-length. For instance, if the depth of the water in the canal is anything more than three-quarters of the length from crest to crest of the wave, the rate of progression of the rear of the procession will be half the speed of the boat. Here then is the state of the case. The boat is followed by an ever-lengthening procession of waves; and the work required to drag the boat along in the canal—supposing that the water is free from viscosity—is just equal to the work required to generate the procession of waves lengthening backwards behind the boat at half the speed of the boat. The rear of the procession travels forwards at half the speed of the boat; the procession lengthens backwards relatively to the boat at half the speed of the boat. There is the whole thing; and if you only know how to calculate the energy of a procession of waves, assuming the water free from viscosity, you can calculate the work which must be done to keep a canal boat in motion. But now note this wonderful result: if the motion of the canal boat be more rapid than the most rapid possible wave in the canal (that is, the long wave), it cannot leave behind it a procession of waves—it cannot make waves, properly so called, at all; it can only make a hump or a hillock travelling with the boat, as shown in Fig. 3, Plate 80. What would you say of the work required to

210

AUG. 1887.

SHIP WAVES.

417

move the boat in that case ? You may answer that question at once: it would require no work; start it, and it will go on for ever. Everyone understands that a curling stone projected along the ice would go on for ever, were it not for the friction of the ice; and therefore it must not seem so wonderful that a boat started moving through water would also go on for ever, if the water were perfectly fluid : it would not, if it is forming an ever-lengthening procession of waves behind it; it would go on for ever, if it is not forming a procession of waves behind it. The answer then simply is, give the boat a velocity greater than the velocity of propagation of the most rapid wave (the long wave) that the canal can have; and in these circumstances, ideal so far as nullity of viscosity is concerned, it will travel along and continue moving without any work being done upon it. I have said that the velocity of the long wave in a canal is equal to the velocity which a body acquires in falling from a height equal to half the depth of the canal. The term " long wave " I may now further explain as meaning a wave whose length is many times the depth of the water in the cana l—50 times the depth will fulfil this condition—the length being always reckoned from crest to crest. Now if the wave-length from crest to crest be 50 or more times the depth of the canal, then the velocity of the wave is that acquired by a body falling through a height equal to half the depth of the canal ; if the wave-length be less than that, the velocity can be expressed only by a complex mathematical formula. The results have been calculated; but I need not put them before you, because we are not going to occupy ourselves with them. The conclusion then at which we have arrived is this : supposing at first the velocity of the boat to be such as to make the waves behind it of wave-length short in comparison with the depth of water in the canal: let the boat go a little faster, and give it time until steady waves are formed behind it; these waves will be of longer wavelength : the greater the speed of the boat, the longer will be the wave-length, until we reach a certain limit; and as the wave-length begins to be equal to the depth, or twice the depth, or three times the depth, we approach a wonderful and critical condition of affairs—we approach the case of constant wave velocity. There will still be a

211

418

SHIP WAVES.

AUG. 1887.

procession of waves behind the boat, but it will be a shorter procession and of higher waves ; and this procession will not now lengthen astern at half the speed of the boat, but will lengthen perhaps at a third, or a fourth, or perhaps at a tenth of the speed of the boat. We are approaching the critical condition: the rear of the procession of waves is going forward nearly as fast as the boat. This looks as if we were coming to a diminished resistance; but it is not really so. Though the procession is lengthening less rapidly relatively to the boat than when the speed was smaller, the waves are very much higher; and we approach almost in a tumultuous manner to a certain critical velocity. I will read you presently Scott Russell's words on the subject. Once that crisis has been reached, away the boat goes merrily, leaving no wave behind it, and experiencing no resistance whatever if the water be free from viscosity, but in reality experiencing a very large resistance, because now the viscosity of the water begins to tell largely on the phenomena. I think you will be interested in hearing Scott Russell's own statement of his discovery. I say his discovery, but in reality the discovery was made by a horse, as you will learn. I found almost surprisingly in a mathematical investigation, " On Stationary Waves in Flowing Water," contributed to the Philosophical Magazine (Oct. Nov. Dec. 1886 and Jan. 1887), a theoretical confirmation, 491/2years after date, of Scott Russell's brilliant " Experimental Researches into the Laws of Certain Hydrodynamical Phenomena that accompany the Motion of Floating Bodies, and have not previously been reduced into conformity with the known Laws of the Resistance of Fluids."* These experimental researches led to the Scottish system of fly-boats carrying passengers on the GlasgoAV and Ardrossan Canal, and between Edinburgh and Glasgow on the Forth and Clyde Canal, at speeds of from eight to thirteen miles an hour, each boat drawn by a horse or pair of horses galloping along the bank. The method originated from the accident of a spirited horse, whose duty it was to drag the boat along at a slow walking speed, taking fright and * By John Scott Russell, Esq., M.A., F.R.S.E. Read before the Royal Society of Edinburgh, 3 April 1837, and published in the Transactions in 1810.

212

AUG. 1887.

SHIP WAVES.

419

running off, drawing the boat after him ; and it was discovered that, when the speed exceeded the velocity acquired by a body falling through a height equal to half the depth of the canal (and the horse certainly found this), the resistance was less than at lower speeds. Scott Russell's description of how Mr. Houston took advantage for his Company of the horse's discovery is so interesting that I quote it in extenso:— " Canal navigation furnishes at once the most interesting illustrations of the interference of the wave, and most important opportunities for the application of its principles to an improved system of practice. " It is to the diminished anterior section of displacement, produced by raising a vessel with a sudden impulse to the summit of the progressive wave, that a very great improvement recently introduced into canal transports owes its existence. As far as I am able to learn, the isolated fact was discovered accidentally 011 the Glasgow and Ardrossan Canal of small dimensions. A spirited horse in the boat of William Houston, Esq., one of the proprietors of the works, took fright and ran off, dragging the boat with it, and it was then observed, to Mr. Houston's astonishment, that the foaming stern surge which used to devastate the banks had ceased, and the vessel was carried on through water comparatively smooth with a resistance very greatly diminished. Mr. Houston had the tact to perceive the mercantile value of this fact to the canal company with which he was connected, and devoted himself to introducing on that canal vessels moving with this high velocity. The result of this improvement was so valuable, in a mercantile point of view, as to bring, from the conveyance of passengers at a high velocity, a large increase of revenue to the canal proprietors. The passengers and luggage are conveyed* in light boats, about * This statement was made to the Royal Society of Edinburgh, in 1837, and it appeared in the Transactions in 1840. Almost before the publication in the Transactions the present tense might, alas, have been changed to the past— " passengers were conveyed/' Is it possible not to regret the old fly-boats between Glasgow and Ardrossan and between Glasgow and Edinburgh, and their beautiful hydrodynamics, when, hurried along on the railway, we catch a

213

420

SHIP WAVES.

AUG. 1887.

sixty feet long, and six feet wide, made of thin sheet iron, and drawn by a pair of horses. The boat starts at a slow velocity behind the wave, and at a given signal it is by a sudden jerk of the horses drawn up on the top of the wave, where it moves with diminished resistance, at the rate of 7, 8, or 9 miles an hour."' Scott Russell was not satisfied with a mere observation of this kind. He made a magnificent experimental investigation into the circumstances. An experimental station at the Bridge of Hermiston on the Forth and Clyde Canal was arranged for the work, as represented in Fig. 1, Plate 80. The experimental station was so situated that there was a straight run of 1500 feet along the bank, and three pairs of horses are seen galloping along. They seem from the drawing to be galloping on air, but are of course on the towing path; and this remark may be taken as an illustration that, if the horses only galloped fast enough, they could gallop over the water without sinking into it, as they might gallop over a soft clay field. That is a sober fact with regard to the theory of waves; it is only a question of time how far the heavy body will enter into the water, if it is dragged very rapidly over it. This however is a digression. The very ingenious apparatus of Scott Russell's is delineated in Fig. 1. There is a pyramid 75 feet high, supporting a system of pulleys which carry a heavy weight suspended by means of a rope. The horses are dragging one end of this rope, while the other end is fastened to a boat which travels in the opposite direction. It is the old principle applied by Huyghens, and still largely used, in clockwork. Scott Russell employed it to give a constant dragging force to the boat from the necessarily inconstant action of the horses. I need not go into details, but I wish you to see that Scott Russell, in devising these experiments, adopted methods for accurate measurement in order to work out the theory of those results, the general natural history of which he had previously observed. I will now read certain results from Scott Russell's paper that I think are interesting. The depth of the canal at the experimental .glimpse of the Forth and Clyde Canal still used for slow goods traffic; or of some .swampy hollows, all that remains of the Ardrossan Canal on which the horse and Mr. Houston and Scott Russell made their discovery ?

214

AUG. 1887.

SHIP WAVES.

421

station was about 4 or 5 feet on an average ; it was really 51/2feet in the middle, but a proper average depth must have been about 41/2 feet, because Scott Russell found by experiment that the natural speed of the long wave was about 8 British statute miles an hour or 12 feet per second. Here then are some of the results, The "Raith," a boat weighing 10,239 Ibs. (nearly 5 tons), took the. following forces to drag it along at different speeds :—at 4 • 72 milesan hour 112 Ibs.; at 5. 92 miles an hour 261 Ibs.; and at 6 • 19 milesan hour 275 Ibs. There is no observation at the critical velocity of about 8 miles an hour. The next is at 9 • 04 miles an hour, and the force is 250 Ibs., as compared with 275 Ibs. at 6.19 miles an hour. Then at a higher speed, 10 . 48 miles an hour, the force required to drag it increases to 2681/2Ibs. This illustrates that water is not a perfect fluid. It also illustrates the theoretical result in a beautiful and interesting way. If water were a perfect fluid, the forces at the lower speeds would be somewhat less than he has given, perhaps not very much less: at all speeds above 8 miles the force would be nothing; the boat once started, the motion would go on for ever. On the same canal another boat, weighing 12,579 Ibs. (nearly 6 tons), gave these still more remarkable results :—at 6.19 miles an hour 250 Ibs.; at 7.57 miles an hour 500 Ibs.; at 8.52 miles an hour 400 Ibs.; and at 9 • 04 miles an hour only 280 Ibs. That is a striking confirmation of the result of the previous observations. Scott Russell says also: "I have seen a vessel in 5 feet water, and drawing only 2 feet, take the ground in the hollow of a wave having a velocity of about 8 miles an hour, whereas at 9 miles an hour the keel was not within 4 feet of the bottom." Again he says: " Two or three years ago, it happened that a large canal in England was closed against general trade by want of water, drought having reduced the depth from 12 to 5 feet. It was now found that the motion of the light boats was rendered more easy than before; the cause is obvious. The velocity of the wave was so much reduced by the diminished depth, that, instead of remaining behind the wave, the vessels rode on its summit." He also makes this interesting statement: "I am also informed by Mr. Smith of Philadelphia, that he distinctly recollects the circumstance of having

215

422

SHIP WAVES.

AUG. 1887.

travelled on the Pennsylvania canal in 1833, when one of the levels was not fully supplied with water, the works having been recently executed, and not being yet perfectly finished. This canal was intended for 5 feet of water, but near Silversford the depth did not exceed 2 feet; and Mr. Smith distinctly recollects having observed to his astonishment, that, on entering this portion, the vessel ceased to ground at the stern, and was drawn along with much greater apparent ease than on the deeper portions of the canal." Even if one regretted the introduction of railways, do not imagine that it can be set forth on mechanical grounds that traction in a canal can compete for any considerable speeds with traction on a railway. Taking again some of the figures already given, a boat weighing 10,239 Ibs. required 112 Ibs., or about 1-100th of its weight, to drag it at 43/4miles an hour. So that to drag a boat at that moderate speed took the same force as would be required to drag it on wheels up an incline of 1 in 100, supposing there to be no friction in the wheels on a railway. But at the higher speed of 9 miles an hour, taking advantage of the comparatively smaller force due to having passed the velocity corresponding with the long wave, we have 250 Ibs., which divided by 10,239 is about 1 in 40; so that the force required to drag the boat along at the rate of 9 miles an hour was what would be required to drag it on wheels up an incline of 1 in 40. Sad to say, I am afraid the wheels have it in an economical point of view. Ship Waves at Sea.—I must now call your attention to the most beautiful, the most difficult, and in some respects the most interesting part of my subject, that is, the pattern of waves formed in the rear of a ship at sea, not confined by the two banks of a canal. The whole subject of naval dynamics, including valuable observations and suggestions regarding ship waves, was worked out with wonderful power by William Froude; and the investigations of the father were continued by his son, Edmund Froude, in the Government Experimental Works at Haslar, Gun Creek, Gosport. William Froude commenced his system of nautical experiments in a tank made by

216

AUG. 1887.

SHIP WAVES.

423

himself at Torquay, in Devonshire; first wholly at his own expense for several years, and afterwards with the assistance of the Government he continued those experiments till his death. The Admiralty have taken up the work, and have made for it an experimental establishment in connection with the dockyard of Portsmouth; and now, after the death of William Froude, his son Edmund continues to carry out there his father's ideas, working with a large measure of his father's genius, and, with his father's perseverance and mechanical skill, obtaining results, the practical value of which it is impossible to over-estimate. It is certainly of very great importance indeed to this country, which depends so much on shipbuilding, and the prosperity of which is so much influenced by the success of its shipbuilders, to find the shapes of ships best suited for different kinds of work—ships of war, swift ships for carrying mails and passengers, and goods carriers. I may mention also that one of our great shipbuilding firms on the Clyde, the Dennys, feeling the importance of experiments of this kind, have themselves made a tank for experimental purposes on the same plan as Mr. Fronde's tank at Torquay; and Mr. Purvis, who, when a young man, was one of Mr. Froude's assistants, is taking charge of that work. The Dennys are going through, with their own ships, the series of experiments which Mr. Froude found so useful, and which the Admiralty now find so useful, in regard to the design of ships; and as the outcome of all this work a ship can now be confidently designed to go at a certain speed, to carry a certain weight, and to require a certain amount of horse-power from the engine. The full mathematical theory of ship waves has been exceedingly attractive in one sense, and in another sense it has been somewhat repulsive, because of its great difficulty, for mathematicians who have been engaged in hydrodynamical problems. Following out that principle of Stokes, which was further developed and generalised by Lord Eayleigh, we can see how to work out this theory in a thorough manner. In fact I can now put before you a model, Figs. 10 and 11, Plate 82, constructed from calculations which I have actually made, by following out the lines of theory that I have indicated. I find that the whole pattern of waves is comprised between two straight

217

424

SHIP WAVES.

AUG.

1887.

lines drawn from the bow of the ship and inclined to the wake on its two sides at equal angles of 19° 28'. It is seen in Figs. 9 and 10 that two such lines, drawn from the how or front shoulders of the ship, include the whole wave-pattern. There is some disturbance in the water abreast of the ship, before coming to these two lines. Theoretically there is a disturbance to an infinite distance ahead and in every direction ; but the amount of that disturbance practically is exceedingly small—imperceptible indeed—until you come to these two definite lines. You see the oblique wave-pattern— waves in echelon pattern. The law of that echelon is illustrated by the curves shown in Fig. 9, Plate 82. The algebraical equations of these curves arc

and

where x and y, according to ordinary usage, are measured along, and perpendicular to, the direction of motion from E towards A, and w is an arbitrary variable ; by assuming a series of arbitrary values for w, a corresponding series of values for x are found from the first equation, and thence the corresponding values of y from the second. I trace a complete curve thus — ABC and ADC; there is a perfect cusp in each curve at B and D respectively, although it cannot be shown perfectly in the drawing. Another formula, which need not be reproduced here, gives a wave-height for every point of those curves. Take alternate curves for hollows, and for crests; and now in clay or plaster of Paris mould a form corresponding with the elevation cine to the curve AB, plus the elevation due to the curve BC, adding the two together; thus you get for every point of your curves a certain wave-height. With the assistance of Mr. Maclean and Mr. Niblett the beautiful clay model which is before you (illustrated in Figs. 10 and 11, Plate 82) has been made, and it shows the results of the theory constructed from actual calculation. I will tell you how to construct the angle of 19° 28' made by each of those two straight lines AB and AD with the direction of motion CA. Draw a circle; produce the diameter from one end to a length equal to the diameter; and from the outer extremity of this projecting line draw two tangents to the circle. Each of those tangent lines makes an angle of 19° 28' with the

218

AUG. 1887.

SHIP WAVES.

425

produced diameter, that is, with the wake of the ship or with the line of progression of the ship. A little more as to the law of this diagram, Fig. 9, Plate 82. The echelon waves consist chiefly of the very steep waves at a cusp. The theoretical formula gives infinite height at the cusp; but that is only a theoretical supposition, though it gives an interesting illustration of mathematical " infinity." Blur it, or smooth it down, precisely as an artist does when he designedly blurs a portion of his picture to produce an artistic effect; blur it artistically, correctly, and mathematically, and you get the pattern. It will be impossible to realise that perfectly; but I have endeavoured to do it in the model illustrated in Figs. 10 and 11, necessarily with an enormous exaggeration however, as you will remark. While every other dimension is unchanged, you must suppose each wave to be reduced to about a fifth part of its height shown in this model; thus you will get the steep " steamboat waves," so much enjoyed by the little boys who, regardless of danger, row out their boats to them every day at the Clyde watering places. Theoretically these waves are infinitely steep ; practically they are so steep that the boat generally takes in a little water, and is sometimes capsized. There is a distance of perhaps a couple of feet from crest to crest, and the wave is so steep and " lumpy" on the outer border of the echelon that there is frequently broken water there fifty or a hundred yards from the ship. One point of importance in the geometry of this pattern is that each echelon cusp, represented in Fig. 9 at B or D, bisects the angle between the flank line AB or AD and the thwart-ship line BD: the angle in question being 70° 32' (90° — 19° 28'). An observation of this angle was actually made for me by Mr. Purvis. He observed, from the towing of a sphere instead of a boat (so as to get a more definite point), the angle between the flank line AB and the direction of motion CA, and found it to be 191/4°.The theoretical angle is 19° 28', and we have therefore in this observation a very admirable and interesting verification of the theory. The doctrine embodied in the wave-model illustrated in Figs. 10 and 11, Plate 82, may be described in a very general way thus. Think 2 o

219

426

SHIP WAVES.

AUG. 1887.

of a ship travelling over water. How is it tliat it makes the wave ? Where was the ship when it gave rise to the wave BCD in Fig. 9 ? Answer: the portion BCD of the wave-pattern is due to what the ship did to the water when the ship was at E, the point E being at the same distance behind C that the point A is in front of C; when it was at E it was urging the water aside, and the effect of the ship pushing the water aside was to leave a depression. Now suppose the ship to be suddenly annihilated or annulled, what would be the result ? The waves would travel out from it, as in the case of a stone thrown into the water. Again suppose the ship to move ten yards forward and then stop, what would be the result ? A set of waves travelling forward while the disturbance that the ship made by travelling ten yards remains. Now instead of stopping, let the ship go on its course : the wave disturbance is going its course freed from the ship, and travels forward. When the disturbance originated which has now reached any point C, the ship was as far behind that point C as it now is before it. Calculate out the result from the law that the group-velocity is half the wave-velocity—the velocity of a group of waves at sea is half the velocity of the individual waves. Follow the crest of a wave, and you see the wave travelling through the group, if it forms one of a group or procession of waves. Look, quite independently of the ship, at a vast procession of waves, or imagine say fifty waves; look at one of those waves, follow its crest; in imagination fly as a bird over it, keeping above the crest as a bird in soaring does sometimes, and, beginning over the rear of the procession, a hundred yards on either side of the ship's wake, you will find the waves get larger and larger as you go forwards. Then go backwards through the procession, and you will see the waves get smaller and smaller and finally disappear. You have now gone back to the rear of the procession ; a small wave increases and travels uniformly forward, and, while the crest of each wave always goes on with the velocity corresponding to the length of the wave, the rear of the procession travels forward at half the speed of the wave: so that every wave is travelling forward through the procession from its rear at a speed which is the same relatively to the rear of the procession as the speed of the rear of the

220

AUG. 1887.

SHIP WAVES.

427

procession relatively to the water. Thus each separate wave is travelling at the ship's speed, which is twice as fast relatively to the water as the rear of the procession of waves is travelling. The wave is the progression of a form; the velocity of a wave is clearly intelligible; the velocity of a procession of waves is still another thing. The penetrating genius of Stokes originated the principle, admirably worked out by Osborne Reynolds and Lord Rayleigh, who have given us this in the shape in which we now have it. Now I must call your attention to some exceedingly interesting diagrams that I am enabled to show you through the kindness of Mr. W. H. White, director of Naval Construction for the Admiralty, and Mr. Edmund Froude, to whose work I have already referred. Fig. 12, Plate 82, shows a perspective view of echelon waves taken from Mr. William Froude's paper, " Experiments upon the Effect Produced on the Wave-making Resistance of Ships by Length of Parallel Middle Body" (Institution of Naval Architects, vol. xviii 1877, page 77). The three diagrams from Mr. White, Figs. 6, 7, and 8, Plate 81, show profiles of the thwart-ship waves of various ships at different speeds. Look first at Fig. 6, showing the wave profile for H.M.S. '' Curlew" at a speed of nearly 15 knots an hour. Note how the water, after the first elevation, dips down below the still-water line; rises up to a ridge at a distance back from the first nearly but not exactly equal to the wave-length corresponding with the speed ; and then falls clown again, experiencing various disturbances. From the appearance of the waves raised by ships going at high speeds, we may learn to tell how quickly they are going. The other day, at the departure of the fleet from Spithead after the great naval review, a ship was said to be going at 18 knots, while it was obvious from the waves it made that it was not going more than 12. In Fig. 7 we have wave profiles for another ship at two different speeds. The upper line corresponds to a speed of 18.4 knots; the lower line to a speed of 17 knots. In the first case the water shoots up to its first maximum height close to the bow, sinks to a minimum 2 o 2

221

428

SHIP WAVES.

AUG. 1887.

towards midships, and flows away past the stern slightly above still-water level. In the second case the character of the wave is somewhat similar, but smaller in height; and there is a marked difference at the stern, due to other disturbing causes. In Fig. 8 we have three different speeds for H.M.S. " Orlando" similarly represented. There is still another very interesting series of diagrams, Figs. 13 to 19, Plates 83 and 84, taken from Mr. Edmund Froude's paper "On the Leading Phenomena of the Wave-making Kesistance of Ships," read before the Institution of Naval Architects, 8th April 1881. In Figs. 13 to 17 are shown the waves produced by a torpedo launch at speeds of 9, 12, 15, 18, and 21 knots per hour. We need not here go into the law of wave-length, but I may tell you that it is as the square of the velocity: the wave-length is four times as great for 18 as for 9 knots. Look now at the pattern of the waves in Figs. 9 and 10, Plate 82. Look at the echelon waves and the thwart-ship waves. Mr. Froude had not worked out the theory that has given the curvature of the transverse ridge exactly; but he drew the waves from general observation, and it is wonderful to see how nearly they agree with the theoretical curves, Fig. 9, and the model, Fig. 10. Velocity and Length of Waves. Velocity of Wave. Length of Wave. Feet. Knots per hour. 6 7 8 9 10 11 12 13 14 15 16

222

19.513 26.559 34.690 43.904 54.203 65.585 78.052 91.602 106.238 121.956 138.760

Velocity of Wave. Knots per hour.

Length of Wave. Feet.

17 18 19 20 22 24 26 28 30 35 40

156.646 175.618 195.672 216.812 262.343 312.209 366.412 424.952 487.827 663.987 867-248

AUG. 1887.

SHIP WAVES.

429

That is a most interesting series of diagrams in Plates 83 and 84, and as a lesson it conveys more than any words of mine. Here is a table (page 428) giving the length of a free wave; and remember, when once the waves are made and are left by the ship, they are then and thereafter free waves. At 6 knots per hour the wave-length is 191/2feet; at 12 knots it is four times as great. At 10 knots it is 54 feet; at 20 it is four times as much. The greatest speeds in Froude's diagrams give about 240 feet length of wave. Now that is a very critical point with respect to the length of the wave and the speed of the ship. I may tell you that Froude the elder and his son Edmund have made most admirable researches in this subject, and have poured a flood of light on some of the most difficult questions of naval architecture. Parallel Middle Body.—I should like to say something about the practical question of parallel middle body. When I first remember shipbuilding on the Clyde, and its progress towards its present condition, a very frequent incident was that when a ship was floated it was found to draw too much water forward, in other words to be down by the head. When this happened, the ship was taken out of the water again, and a parallel piece, 10 or 20 or 30 feet long, was put into the middle: a parallel middle body, curved transversely, but with straight lines in the direction of its length. Many a ship was also lengthened with a view to add to its speed. William Froude took up the question of parallel middle body, and the effect of the entrance and run. The entrance is that part of the ship forward, where it enters the water and swells out to the full breadth of the ship ; the run is the after part, extending from where the ship begins to narrow to the stern. A ship may consist of entrance, parallel middle body, and run. Froude investigated the question, Is the parallel middle body inserted in a ship an advantage or a disadvantage, in some cases or in all cases ? He found it a very complex question. According to the relation of the wave-length to the length of the ship, it produces a good or a bad effect. A ship with a considerable length of parallel middle body shows very curious phenomena regarding the resistance at different speeds. As

223

430

SHIP WAVES.

AUG. 1887.

the speed is raised, the resistance increases; but on a further increase of speed, it seems as if it was beginning to diminish; the resistance never quite diminishes however with increase of speed, but only increases much less rapidly. The curve indicating the relation of the speed to the velocity has a succession of humps or rises, each showing a rapid increase of resistance; between these it becomes almost flat, showing scarcely any increased resistance. Froude has explained that thoroughly by the application of this doctrine of ship waves which I have endeavoured to put before you. When the effect of the entrance or bow, and the effect of the run or stern, are such as to annul or partially to annul each other's influence in the production of waves, then we have a favourable speed for that particular size and shape of ship. On the other hand, when the crest of a wave astern due to the action of the bow agrees with the crest of a wave astern due to that of the stern, then we have an unfavourable speed for that particular size of ship. Thus Froude worked out a splendid theory, according to which, for the speed at which a ship is to go, a certain length of parallel middle body may, if otherwise desired, be an advantage. But on the whole the conclusion was that—if the ship is equally convenient, if it is not too expensive, if it can pass through the lock gates &c., and if all the other practical conditions can be fulfilled, without a parallel body—it is better that the ship should be all entrance and run, according to Newton's form of least resistance: fine lines forward, swelling out to greatest breadth amidships, and tapering finely towards the stern. In other words, the more ship-shape a ship is, the better. I wish to conclude by offering one suggestion. I must premise that, when I was asked by the Council to give this lecture, 1 made it a condition that no practical results were to be expected from it. I explained that I could not say one word to enlighten you on practical subjects, and that I could not add one jot or tittle to what had been done by Scott Kussell, by Eankine, and by the Froudes, father and son, and by practical men like the Dennys, W. H. White, and others ; who have taken up the science and worked it out in practice. But there is one suggestion founded on the doctrine of

224

AUG. 1887.

SHIP WAVES.

431

wave-making, which 1 venture to offer before I stop. I have not explained how much of the resistance encountered by a ship in motion is due to wave-making, and how much to what is called skin resistance. I can briefly give you a few figures on this point, which have been communicated to me by Mr. Edmund Froude. For a ship A, 800 feet long and 311/2feet beam and 2634 tons displacement, a ship of the ocean mail steamer type, going at 13 knots an hour, the skin resistance is 5 • 8 tons, and the wave resistance 3 • 2 tons, making a total of 9 tons. At 14 knots the skin resistance is but little increased, namely 6.6 tons; while the wave resistance is nearly double, namely 6 • 15 tons. Mark how great, relatively to the skin resistance, is the wave resistance at the moderate speed of 14 knots for a ship of this size and of 2634 tons weight or displacement. In the case of another ship B, 300 feet long and 46 • 3 feet beam and 3626 tons displacement—a broader and larger ship with no parallel middle body, but with fine lines swelling out gradually—the wave resistance is much more favourable. At 13 knots the skin resistance is rather more than in the case of the other ship, being 6 • 95 tons as against 5 • 8 tons; while the wave resistance is only 2 • 45 tons as against 3 • 2 tons. At 14 knots there is a very remarkable result in this broader ship with its fine lines, all entrance and run and no parallel middle body:—at 14 knots the skin resistance is 8 tons as against 6.6 tons in ship A, while the wave resistance is only 3.15 tons as compared with 6 • 15 tons. Another case which I can give you is that of a torpedo boat 125 feet long, weighing 51 tons. At a speed of 20 knots an hour the skin resistance is 1 • 2 ton, and the wave resistance 1 • 1 ton; total resistance 2 • 3 tons. To calculate the horsepower you multiply the speed in knots per hour by 62/3,and -then multiply the resistance in tons by the product so obtained; and the result for the torpedo boat going at 20 knots an hour is 307 horsepower to overcome a resistance of 2 • 3 tons or l-22nd of her weight (51 tons). Again the ship B of 300 feet length, going at 20 knots an hour with an expenditure of 4550 horse-power, experiences a resistance of 34 tons, or about 1-110th of her weight (3626 tons). Thus the energy actually expended in propelling these vessels at 20 knots an hour at sea would be sufficient, if they were supported

225

432

SHIP WAVES.

AUG. 1887.

on frictionless wheels, to drag them at the same speed up railway inclines, of 1 in 22 for the' torpedo boat, and 1 in 110 for the ship B. My suggestion is this, and I offer it with exceedingly little confidence, indeed with much diffidence; but I think it is possibly worth considering. Inasmuch as wave resistance depends almost entirely on action at the surface of the water, and inasmuch as a fish swimming very close to but below the surface makes very little wave disturbance, it seems to me that by giving a great deal of body below the water line we may relatively dimmish the wave disturbance very much. To get high speeds of 18 and 20 knots an hour, it is probable that, by swelling out the ship below like the old French ships, instead of having vertical sides—making the breadth of beam say five feet more below the water than at the water line—there may be obtained a large addition to the displacement or carrying-power of the ship, with very little addition to the wave disturbance, and therefore with very little addition to the wave resistance, which is most important at high speeds. I think it may be worth while to consider this in regard to the designs of ships. In conclusion, I should like to urge you to look at these phenomena for yourselves. Look at the beautiful wave-pattern of capillary waves, which you will find produced by a fishing line hanging vertically from a rod, or from an oar, or from anything carried by a vessel moving slowly through smooth water at speeds of from about1/2knot to 2 knots an hour. Again look at the equally beautiful wave-pattern produced by ships and boats, as illustrated in Plate 82. But you can scarcely see the phenomena more beautifully manifested than by a duck and ducklings. A full-sized duck has a splendidly shaped body for developing a wave-pattern, and going at good speed it produces on the surface of a pond very nearly the exact pattern of ocean waves. A little duckling going as fast as it can, perhaps about a knot an hour, shows very admirably the capillary waves,* differing manifestly from the ocean waves formed in * For information regarding capillary waves, see Scott Russell's Report on Waves (British Association, York, 1844, pages 311-390); also parts III,

226

AUG. 1887.

SHIP WAVES.

433

the front and at the rear of a larger body moving more rapidly through the open water. I call attention to this, because, having given you perhaps a rather dry statement of scientific facts, if I can say a word that will lead you each to use your eyes in looking at ships, boats, ducks, and ducklings, moving on water at different speeds, and to observe these beautiful phenomena of waves, I think, even were you to remember nothing of this lecture, you would have something to keep in your minds for the rest of your lives.

227

Fig .1. ExperimeniaL Apparatus 1835. Scal 1/700th.

Fig. 2. Boal behind the wave.

Fig. 3. Boat upon the wave.

Fig. 4. At 15 knots.

Fig. 5. At 20 1/4 knots.

Fig. 6. Wave Profile,

H.M.S."Girlew"

Fig. 7. Wave Profiles.

Fig. 8. Wave, Profiles, H.M.S. "Orlando ."

Kg. 13. At 9 Knots.

Fig.14. At 12 Knots.

Kg. 15. (See also Fig 4 .)

At 15 Krwts.

Fig. 16. At 18 knots.

Fig.17. At 21 knots (See also Fig.5)

231

Fig. 18.By83feetLaunch.

Fig. 19 .

232

By 333 feet Ship.

Charles Algernon Parsons (1854-1931) Charles Parsons was born into a family of gifted scientists. His father, the 3rd Earl of Rosse, was President of the Royal Society and one of Ireland's foremost astronomers. Parsons grew up under the shadow of the largest reflecting telescope in the world, located at his home at Birr Castle, Parsonstown. He attended Dublin and Cambridge Universities before joining W G Armstrong & Company of Elswick. His early experimental research on torpedoes was followed by an interest in power generation. By 1884 Parsons had become a partner in the Electrical Department of Clarke Chapman & Company of Gateshead, where he concentrated on inventions. He made rapid progress in developing turbines for ship and shore-based electric lighting. In placing multiple blades on a single shaft (the rotor) and driving high velocity steam through a fixed casing and vanes (the stator) Parsons achieved rotor speeds in excess of 15,000 rpm. He designed complementary dynamos from 1885 and the combined "sets" became a working proposition. By 1888, over 200 sets were in marine service, powering ships' lights. The success of Parsons' ideas, culminated in the first use of a steam turbine in a public power station (Forth Banks, Newcastle-upon-Tyne) led him to form C A Parsons & Company in 1889. The business expanded rapidly to a 25 acre complex, supplying turbo-generators world-wide. In his original 1884 patents he had noted the potential for using turbines in marine propulsion. This led to the eventual construction of the steam yacht Turbinia,. Its cavalier debut, outrunning the British fleet at the 1897 Naval Review, is probably the most famous single incident in the engineer's career. Steam turbines were adopted in naval ships (Viper and Cobra, 1899), passenger vessels (King Edward, 1901 and Mauretania, 1906) and, with the development of reducing gears, added to cargo vessels (Vespasia, 1909). Parsons thus dominated land and sea-based power systems. Interestingly, although Parsons experimented with steam-driven flying models, he appears not to have considered the turbine in an aeronautical context. From 1921, Parsons began in earnest to transform and modernise the British optics industry. Control of what became Grubb, Parsons & Company allowed him to produce large telescopes, even bigger than that at Birr Castle.

233

480

OCT. 1888.

DESCRIPTION OF THE COMPOUND STEAM TUEBINE AND TUKBO-ELECTKIC GENERATOE. BY THE HONOURABLE CHARLES A. PARSONS, OF GATESHEAD.

The Compound Steam Turbine has now been developed into a motor which utilizes steam with a high degree of economy. It possesses considerable simplicity, and its speed of revolution is high; and as dynamos working at a high speed combine cheapness and efficiency, the application of the steam turbine for driving them is at first sight a good one. The combination of the steam turbine and dynamo has involved a considerable departure from existing practice, and has necessitated much experimental work, and investigations on entirely new ground. The first Turbo-Electric Generator, completed about four years ago, ran at 18,000 revolutions per minute, and gave six electrical horse-power; it has been in almost constant use since that time, and has done a large amount of work. The second, made shortly afterwards, runs at 10,000 revolutions per minute; it was placed on the Tyne Steam Shipping Co.'s steamer "Earl Percy," and has worked her 60 lamps ever since to their entire satisfaction ; the cost of fuel and maintenance is very small, and the light remarkably steady. Generators were then made for supplying up to 250 lamps, and a large number of installations were carried out, which have given excellent results; the consumption of steam was about equal to that of a good high-pressure engine with single slide when working with the same steam pressure and driving a good dynamo; but so marked has been the economy realised in regard to lamp renewals, oil, attendance, and other items, that the generators have almost without exception given great satisfaction. It became essential however, if these generators were to be successfully adopted for large installations, that higher degrees of economy should be realised, more nearly approaching those of the best compound engines. Theory based on the authenticated performances of

235

OCT. 1888.

COMPOUND STEAM TURBINE.

481

water turbines and the laws of the flow of steam and gases showed that the turbo-electric generator possessed the elements of the highest economy, not merely comparable with the best known performances, but even superior to them. How far practice has come up to theory may be judged by the results given at the end of this paper, which it will be seen approach nearly the best results of ordinary engines working with the same steam pressures. Compound Steam Turbine.—The compound steam turbine T, Figs. 1 and 2, Plates 85 and 86, consists of two series of parallel-flow or Jonval turbines, set one after the other on the same spindle S, so that each turbine takes steam from the one before and passes it on to the one following. In this way the steam entering all round the spindle from the central inlet I, Fig. 1, passes right and left through the whole of each series of turbines to the exhaust E at each end. The steam expands as it loses pressure at each turbine; and by successive steps the turbines are increased in size or area of passage-way, so as to accommodate the increase of volume, and to maintain a suitable distribution of pressure and velocity throughout the whole series of turbines. The areas of the successive turbines are so arranged that the velocity of the flow of steam shall bear throughout the series about the same ratio to the speed of the blades; and as far as possible this ratio of velocity is so fixed as to give each turbine of the series its maximum efficiency. The two equal series of turbines on each side of the central steam inlet I balance each other as regards any end pressure on the spindle of the motor, and thus remove any tendency to undue wear on the collars of the bearings B. The turbines are constructed of alternate revolving and stationary rings of blades. The revolving blades r, Fig. 9, Plate 90, are cut with right or left-hand obliquity on the outside of a series of brass rings, which are threaded upon the horizontal steel driving spindle S, and secured upon it by feathers; the end rings form nuts, which are screwed upon the spindle and hold the rest of the rings upon it. The stationary or guide blades g are cut with opposite obliquity on the inside of another series of larger brass rings, which are cut in halves, and are held in the top and bottom halves of the

236

482

COMPOUND STEAM TURBINE.

OCT. 1888.

cylindrical casing by feathers. The set of blades on each revolving ring runs between a pair of sets of the stationary or guide blades. The passages between the blades in the alternating rings form a longitudinal series of zigzag channels when the machine is standing still, as seen at Z in Plates 86 and 90. Bearings.—In Fig. 6, Plate 87, is shown full size a longitudinal section of one of the bearings. As it is impossible to secure absolute accuracy of balance, the bearings are of special construction so as to allow of a certain very small amount of lateral freedom. For this purpose the bearing is surrounded by two sets of steel washers l-16th inch thick and of different diameters, the larger fitting close in the casing C and about 1-32nd inch clear of the bearing, and the smaller fitting close on the bearing and about l-32nd inch clear of the casing C. These are arranged alternately, and are pressed together by the spiral spring N. Consequently any lateral movement of the bearing causes them to slide mutually against one another, and by their friction to check or damp any vibrations that may be set up in the spindle. The tendency of the spindle is then to rotate about its axis of mass, or principal axis as it is called; and the bearings are thereby relieved from excessive pressure, and the machine from undue vibration. The automatic oiling of the bearings by the screw J, Plate 86, almost entirely prevents friction and wear. The circulation is continuous, the oil being used over and over again; and as it deteriorates very slowly, and there is little waste, the consumption may be said to be unusually small. The oil is raised up to the screw J by the suction of the fan F acting upon its free surface in the stand-pipe P. By the screw J it is fed into the adjoining bearing, and is also forced along the pipe H to the two other bearings of the spindle. After passing through the bearings the oil flows back along the pipe K to the reservoir W, Plate 85, to be again drawn up thence through the pipe U by the fan and fed into the bearings by the screw. The throttle-valve V is worked by the movement of a leather diaphragm L, which the suction of the fan F tends to close against the tension of the spring A.

237

OCT. 1888.

COMPOUND STEAM TURBINE.

483

Turbo-Electric Generator.—In Plates 85 and 86 is represented a turbo generator of 25 horse-power actual. All the turbines are here of the same diameter, and the expansive action of the steam is utilized by varying the depth and pitch of the -blades. In Plates 88 to 90 is shown a 50 horse-power turbo generator, which may be said to be of the triple-expansion type, from the fact that it is made with three different diameters of turbines for the purpose of dealing more advantageously with the increasing volume of steam as it expands. The three barrels H K M of different diameters, Plate 89, contain the three successive sizes of turbines. In each barrel the blades are continuously varied in pitch, so that an almost perfect distribution of steam is attained; and each barrel by itself may be compared in some respects with a cylinder in a triple compound engine. In the larger sizes the blades are accurately curved, as shown at Z, Plate 90, as in the best water turbines. To prevent end pressure, the spaces at the ends of the corresponding barrels are connected by equalizing passages Q, Plate 89. Including fluid friction, the theoretical efficiency of each turbine in the set is about 89 per cent.; and the mean efficiency of the whole set is theoretically about 87 per cent, of the power which should be given out in the adiabatic expansion of the steam. At each turbine the flow of steam is continuous, and proceeds unchecked to the next. The steam expands slightly in passing each set of blades, but without shock and with gradually diminishing pressure, the whole energy of expansion being utilized to carry the steam through the subsequent turbines. With the continuous lubrication and small pressure on the bearings, there is no material wear; and as the steam has no cutting action on the turbines, the initial clearances remain the same. Therefore the consumption of steam in the turbo motor does not increase under the conditions of every day running, and after long periods of work has been found to remain almost the same as on the trial run. The power absorbed in friction in the bearings has been estimated: when they are cold it is considerable, amounting to over one-third of a horse-power per bearing; but when the oil

238

484

COMPOUND STEAM TURBINE.

OCT. 1888.

becomes heated to its normal temperature, it falls to less than twenty per cent, of this amount. Dynamo.—The motor is coupled to the dynamo D, Plates 85 and 88, by a coupling socket with squared hole J, Figs. 8 and 9, which fits accurately upon the squared ends of the two spindles ; this admits of the armature being easily withdrawn. The magnets are entirely of cast-iron, and usually are made with simple shuntwinding only. The armature, Fig. 5, Plate 87, is of the drum type. The body is built up of thin iron discs, only 1-100th of an inch thick and insulated from each other by tracing paper; it is turned up, and grooves are milled out to receive the conducting wires. For pressures of from 60 to 80 volts, there are usually thirty grooves. The course of the wire is as follows :—starting at a it is led spirally through a quarter of a turn & round the cylindrical portion c ; then passing longitudinally along a groove in the core it is again led spirally through a quarter turn d round the other cylindrical portion c, then through the end washer, and then back similarly through a quarter spiral turn e, and along the diametrically opposite groove in the core, and lastly through a little more than a quarter spiral turn /, back to g, where it is coupled to the next convolution. The commutator is formed of rings of sections. Each section is formed of short lengths, and each length is dovetailed and interlocked between conical steel rings; the whole is insulated with asbestos, and when screwed up by the end nut forms with the steel bush a compact whole. There are fifteen sections in the commutator, and each coupling is connected to a section. The whole armature is bound entirely from end to end with pianoforte or brass wire. Efficiency of Dynamo.—For a normal output of 400 amperes and 80 volts, the resistance of the armature from brush to brush is only 0 • 0025 ohm. The resistance of the field magnets is 23 ohms, or an electrical efficiency of just 98 per cent. There are losses due to eddy currents in the core and wire of the armature, and to magnetic retardation resulting from change of polarity of the core. These

239

OCT. 1888.

COMPOUND STEAM TURBINE.

485

losses have been ascertained by separately exciting the magnets from another dynamo, and observing the change of steam pressure required to maintain the speed constant; the corresponding power was then calculated. The commercial efficiency of this dynamo has been found to be about 95 per cent. Electrical Control Governor.—On the magnet yoke is the electrical control governor G, Plates 85 and 88, shown one quarter full size in Figs. 3 and 4, Plate 87, the movement of which is caused by the attraction of the magnet yoke upon a small iron bar or needle n, finely balanced and pivoted on a vertical spindle; a spiral spring s resists this attraction. A double finger or arm r is keyed on the same vertical spindle ; the end of each finger r is flat, and when opposite the inlet i to the air-pipe Y closes it. The spiral spring s is so adjusted by the movable head li that the greater the attraction the more is the inlet i closed by one of the fingers r. When the inlet i is open, the inrush of air along the pipe Y partially neutralizes the suction of the fan F, and allows the diaphragm L to extend, and so to open the throttle-valve V. The combination of the fan F, the diaphragm L, and the spring A, forms a good centrifugal governor ; but alone it is not accurate enough in its action for electrical purposes, and requires supplementing by the delicate control of the finely balanced and pivoted magnetic needle n. So accurate is the governor with this addition, that, when the load is gradually varied from nothing up to the maximum, the variation in volts at the terminal is less than one per cent. Steam Consumption.—As the result of careful tests made when exhausting into the atmosphere and giving off 32,000 watts, the consumption of steam per electrical horse-power per hour has been found to be 42 Ibs. with a steam pressure of 61 Ibs. at the inlet; and 35-1 Ibs. with a steam pressure of 92 Ibs. at the inlet. Tests made at Portsmouth Dockyard, and at Messrs. Weyher and Kichemond's in Paris, have agreed closely with the tests made on the same turbo generators before they left the works at Gateshead. These tests have therefore confirmed the accuracy of the figures above given.

240

486

COMPOUND STEAM TUEBINE.

OCT. 1888.

Durability.—After three years' working of ten hours daily, the wear on the bearings has been found to be very small; in some cases almost inappreciable. The blades or vanes of the turbines show no cutting action from the steam. The commutators in the larger sizes have stood this amount of work well, and when carefully looked after have suffered very little wear. Advantages.—The characteristic advantages of the turbo-electric generators may be summed up as follows :—steadiness of the electric current produced, arising from the high speed and the momentum stored in the moving parts; freedom from accident, on account of simplicity and direct action; small first cost, and small cost of maintenance of machine and lamps; small consumption of oil; little attention required; small size and weight for the power developed, which is about nine watts per Ib. of weight in the whole machine, including both engine and dynamo. Consequently they are specially suitable for torpedo boats and fast cruisers, where weight and space are of the utmost importance. The number of these generators already (August 1888) supplied for ship and land installations represents an aggregate of more than 2,000 electrical horse-power.

Discussion, 1 August 1888. The Hon. CHARLES A. PARSONS exhibited a couple of specimens of the engine of different sizes, one with the cover removed to show the construction and arrangement of the series of turbines. The PRESIDENT said all would admit that this was a new departure in high-speed steam engines, and the Members were very much obliged to the author for having brought this paper before the Institution. Judging from the drawings and the engine exhibited,

241

OCT. 1888.

COMPOUND STEAM TURBINE.

487

it appeared to him that the author had successfully overcome a great many difficulties which had hitherto beset high-speed engines. Considering that his engine could run at 18,000 revolutions a minute, without developing a large amount of friction, and without great consumption of oil and steam, it was clear that he had pretty well solved a problem which a good many engineers had been trying to solve.

Subsequent to this paper being presented, there followed lengthy discussion and correspondence which was reproduced in the Proceedings. This indicated the high level of interest that the paper engendered at the time. These 30 pages have been omitted from this facsimile reproduction but are held in the archives of the Institution.

242

Fig. 1.

Side

Elevation.

Fig. 2. Longitudinal Section

of

Turbine.

Fig. 5. Winding

of

Armature. Scale 1/4

th

Fig 3. Plan.

Fig. 6.

Longitudinal Full

Fig. 4.

Side

Elevation.

Section size

of

Bearing.

•

Fig.7.

Side

Elevation.

Fig: 8.

Longitudinal

Section

of

Turbine.

Fig. 9. Enlargement

from

Fig.

8

Scale

half

size.

Richard Francis Trevithick (1845-1913) One of the third generation of Cornish Trevithicks to be associated with railway engineering, Richard Francis' grandfather was the Richard Trevithick (1771-1833), who built the first locomotive to run on rails. Richard Francis Trevithick was educated at Cheltenham College, and served his apprenticeship at Harvey & Company's Hayle Foundry. From there he progressed to the London and North Western Railway workshops at Crewe, before pursuing his profession overseas. By the time of his election to the Institution of Mechanical Engineers in 1876, Trevithick was based in South America. He supervised the locomotive and carriage department of the Central Argentine Railway, based at Rosario. Moving to Asia, Trevithick was appointed Chief Mechanical Engineer to the Ceylon Government Railways, before joining a wave of British engineers assisting in the rapid industrialisation of Japan. Many of the Institution's members worked in Japan during the Meiji era (1868-1912). Generally these were in the key industries of textiles, shipbuilding (Archibald Francis McNab), railway engineering (Benjamin Frederick Wright) and the Imperial Mint (Robert Maclagan). The Imperial University at Tokyo owed its engineering department to the efforts of Robert Henry Smith and Charles Dickinson West. Their students became the first Japanese-born members of the IMechE eventually staffing the locomotive works at Kobe, where Trevithick worked. Trevithick not only had the distinction of building the first Japanese-made locomotive, but was also the last foreign railway official in Japan. The two achievements were related of course. In correspondence with Edgar Worthington, the Institution's Secretary in 1904, Trevithick noted the spirit and determination for independent engineering design in Japan "the evidence is plain as to the direction in which their ambition runs".

249

298

APrIL 1895.

LOCOMOTIVE BUILDING IN JAPAN. BY MR. RICHARD F. TREYITHICK, LOCOMOTIVE AND CARRIAGE SUPERINTENDENT, IMPERIAL GOVERNMENT RAILWAYS OF JAPAN.

The first Locomotive built in Japan, No. 221, is also the first Compound Locomotive employed in this country. Its general appearance and construction are shown in Plates 48 to 56. It wasturned out of the Kobe shop on 26th May 1893, and has worked continuously ever since without giving trouble in any way, and with a mileage of a little over 33,700 during the fifteen months to 31st August 1894. In respect of efficiency, it takes its turn with any other main-line engine, and has always done the work with a smaller consumption of coal, as shown by the comparative statement in Table 1 appended of the working of this engine and of an imported non-compound locomotive No. 88 during eight months' running in competition with each other. It has not been found to require more steam-shed repairs than the non-compound engines, and has never occasioned delay to the traffic from any cause; and the fact that it has not yet been necessary to send it to the shops for general repairs gives reason to believe that its mileages between general overhauls will not be inferior to those of any other class of main-line engines. In first cost it is by far the cheapest main-line engine on the Imperial Government Railways of Japan, having cost £1,350 ; and the particulars of the first cost of this and other engines are given in Table 2. The construction of this engine on the writer's initiative and from his designs was sanctioned as an experiment, subject only to the condition that a tank engine was to be built. That the experiment is considered by the railway administration to have been brought to a successful issue is sufficiently proved by the

251

APRIL 1895.

JAPANESE LOCOMOTIVE BUILDING.

299

sanction given for a repetition on a larger scale, namely for the construction of eight non-compound tender locomotives, which are now being built. Leading Dimensions.—The following are the leading dimensions of this outside-cylinder compound tank locomotive, No. 221, the gauge of the Imperial Government Railways of Japan being 3 feet 6 inches ; see Plates 48 to 56. High-pressure cylinder 15 inches diameter and 20 inches stroke; steam ports 12 inches X13/8inch, exhaust port 12 inches X 23/4inches ; Trick phosphor-bronze slidevalve, lap 11/8inch, inside clearance 1/4 inch, maximum travel 4 5/16- inches full, lead constant 1-16th inch = l-8th inch with ordinary valve. Low-pressure cylinder 221/2inches diameter and 20 inches stroke; steam ports 16 inches X15/8inch, exhaust port 16 inches X 31/4inches ; Trick phosphor-bronze slide-valve, lap 15-16ths inch, inside clearance nil, maximum travel 41/4inches, lead constant 1-16th inch = l-8th inch with ordinary valve. Intermediate receiver double the capacity of the high-pressure cylinder. Joy's radial valvemotion for both valves. Single slide-bars ; wearing surface of crosshead slides 45 square inches. Boiler shell made of best Yorkshire iron plates 1/2 inch thick; inside fire-box of best Yorkshire iron plates 7-16ths inch thick, except copper tube-plate 3/4 inch thick where pierced for tubes and 1/2 inch thick below; barrel 45 inches inside diameter throughout; distance between tube-plates 9 feet 91/4inches. Crown sheet of fire-box stayed with 72 Yorkshire iron stays of7/8inch diameter, having ends enlarged for thread to be cut with 1-inch die. Brass tubes, 157 in number, thickness No. 12 B.W.G. = 0.110 inch, external diameter 13/4inch at fire-box end and l13/16inch at smoke-box end, pitch 2f inches. Heating surface in tubes 703 • 36 square feet, in fire-box above foundation ring 66.21; total 769.57 square feet. Grate area 12.4 square feet. Working pressure 150 1bs. per square inch. Eight wheels, four coupled 53 inches diameter with tires 3 inches thick; coupled-wheel axles of steel, with journals 8 inches long by 61/2inches diameter, and gunmetal axle-boxes without loose steps. Leading and trailing radial wheels 38 inches diameter with tires 3 inches thick; axles of best

252

300

JAPANESE LOCOMOTIVE BUILDING.

APRIL 1895.

Yorkshire iron with journals 81/2inches long by 6 inches diameter, fitted with Webb's radial axle-boxes. Adjusting screws of all bearing springs in tension. Extreme wheel-base 191/2feet, rigid wheel-base 71/2feet. Weight of engine in working order 40 tons; available for traction 211/2tons. Capacity of water tanks 987 gallons. Weight of coal usually put on engine 25 cwts. The non-compound engine No. 88 imported from England, which was run in competition with No. 221 compound, has cylinders 14 inches diameter and 20 inches stroke; steam ports 12 inches X 11/4inch, exhaust port 12 inches X 23/4inches; ordinary slide-valves, lap 3/4 inch, inside clearance nil, maximum travel 31/16inches, lead constant1/8inch. Joy's radial valve-motion. Two slide-bars to each cylinder; wearing surface of cross-head slides 52 square inches. Boiler shell of best Yorkshire iron plates 7-16ths inch thick; inside fire-box of copper plates 1/2 inch thick, with copper tube-plate of same thicknesses as in the compound No. 221; barrel telescopic, smallest inside diameter 43 inches, largest 443/4inches; distance between tube-plates 9 feet 81/4inches. Crown sheet of fire-box stayed by six roof bars, attached to shell and to crown sheet in usual manner. Brass tubes, 147 in number, of exactly same thickness, diameter, and pitch as in compound. Heating surface in tubes 653 square feet, in fire-box above foundation ring 641/2;total 7171/2square feet. Grate area 12 square feet. Working pressure 140 Ibs. per square inch. Eight wheels, four coupled 52 inches diameter with tires 21/2inches thick; coupled-wheel axles of steel, with journals 7 inches long by 6 inches diameter, and steel axle-boxes with loose steps; adjusting screws of driving springs in compression. Leading and trailing radial wheels 37 inches diameter with tires 21/2inches thick ; axles of steel, with journals 81/2inches long by 5 inches diameter, fitted with Webb's radial axle-boxes; adjusting screws of bearing springs in tension. Extreme wheel-base 191/2feet, rigid wheel-base 71/2feet. Weight of engine in working order 35 tons ; available for traction 201/2tons. Capacity of water tanks 1,000 gallons. Weight of coal usually put on engine 25 cwts. The particulars of the valve setting of the two engines are given in Tables 3 and 4 appended. Those of No. 221 compound

253

APRIL 1895.

JAPANESE LOCOMOTIVE BUILDING.

301

were taken from the engine when it was turned out of the shop; in forward gear the usual angles of the quadrant when running trains are 14° and 13° and 12°; in backward gear the engine has scarcely ever had to run a train. The particulars of No. 88 noncompound were taken from the maker's tracing. The road over which these engines—No. 221 compound and No. 88 non-compound—have been worked is fairly level with three ruling gradients of 1 in 100, none of which exceeds three-quarters of a mile in length. The sharpest curves, four in number, on the portion of the line over which No. 221 has worked, are of 20 chains radius, and each of them is about a quarter of a mile long. There is one curve of 36 chains radius; and the others are of 40 chains radius and upwards. One of the 20-chain curves is on a gradient of 1 in 100 ; the others are on practically level road. Particulars of Building at Kobe Works.—Two frame-plates of mild Siemens-Martin steel were obtained from England, planed on both sides, and shaped on edges to tracings sent from Japan; all holes and gaps for horn-blocks were drilled or cut out in Kobe. Four steel plates forming the radial axle-box guides were sent out from England, flanged and bent to required radius in conformity with tracing supplied from Japan. The remainder of the material required for constructing the frame was supplied from the store, out of the ordinary stock of iron plates, bars, and angles. Four straight axles, two of best Yorkshire iron and two of best Siemens-Martin steel, and eight steel tires, were sent out from England as supplied from forge or steel works; all turning and fitting were done in Kobe. The wheels, eight in number, were forged in Kobe from scrap iron, and formed the crucial job in smith work. All tires now used on these railways are of steel; they are made in England, sent out as received from the steel works, and bored out and finished as required in Japan. The valve gear, spring gear, brake gear, draw-bar hooks and attachments were made in Kobe, almost entirely from scrap iron. Springs were made from steel out of stores. The buffers were made partly of Yorkshire iron plate, and partly of scrap iron.

254

302

JAPANESE LOCOMOTIVE BUILDING.

APRIL 1895.

The piston-rods were made of round steel supplied from store, but the cross-heads were made of scrap iron. The low-pressure piston of steel was sent out from England as received from the steel works; all turning and fitting were done in Kobe. The high-pressure piston being of cast-iron was made in Kobe. The slide bars were made of steel bars out of stock. The connecting rods and coupling rods were forged from 6-inch square bars of best Yorkshire iron; the crank pins from 6-inch diameter round bars of same quality. The cylinders, axle-boxes, and all other fittings of cast-iron or brass were cast and finished complete in the Kobe shops. The boiler was made of plates, bars, and angles of best Yorkshire iron supplied from store, all flanging &c. being done in the shops. The foundation ring and fire-hole ring were made from scrap iron. All copper tube-plates used in Kobe for boiler making or repairs are ordered from England of best quality; the flanging is entirely done in Kobe, after which the tube and rivet holes are drilled. The dome with its seating was obtained from England, though not specially ordered for this engine. All boiler mountings were made in Kobe. Tanks, coal bunkers &c. were made of best best Staffordshire plates and angles from stock. Besides the dome and its seating, the only finished pieces from abroad put into this engine are one Bourdon pressure-gauge, one vacuum-brake ejector, one vacuum-brake duplex pressure-gauge, and one vacuum-oil sight-feed lubricator. All boiler tubes used in Japan are of course imported; likewise the bulk of the piping, both copper and iron, required in a locomotive; and indiarubber springs &c. It will thus be seen that in this job the builders have done nearly every part of the work it was possible for them to undertake unless possessed of ironworks. The workmanship, both in detail and in the engine as a whole, leaves nothing to be desired, and compares favourably with that of the best engines imported. The significance of this result is emphasized when it is understood that it is entirely the product of Japanese labour led by Japanese foremen, no foreign foremen being employed in the Kobe workshops.

255

APRIL 1895.

JAPANESE LOCOMOTIVE BUILDING.

303

The position of the compound system for locomotives in Japan is at present merely that of trial. The view entertained after the first fifteen months' working of No. 221 compound, that there was no reason to believe it would prove costly in maintenance, has been confirmed by subsequent experience. On 2nd October 1894, after sixteen months' running with a mileage of 35,573, the engine was sent to the repairing shop for wheel turning, and a few other small repairs were then executed. On 6th October it was again put to work, and by 31st May 1895 had brought its mileage up to 57,473. It has not been found that any of the non-compound engines do better. The Government Railways led the way by building this compound locomotive. Two private railways shortly afterwards followed suit by each purchasing one or two four-cylinder compound locomotives on the Vauclain system from the Baldwin Locomotive Works, Philadelphia, Pennsylvania; and finally in August 1894 the Japan Railway put in traffic a couple of two-cylinder compounds built by Messrs. Neilson and Co. of Glasgow. The Baldwin engines, for which possibly too much was claimed, have not quite realized all that was expected of them, nor have they in any way disparaged the locally built engine. Some particulars of the working of these fourcylinder compounds are given in Table 5 in comparison with No. 221 two-cylinder compound and No. 179 non-compound; the latter is a duplicate of No. 88 non-compound, and was the first engine worked against No. 221 compound. It will thus be seen that a locomotive has been built in Japan, which in first cost and efficiency will bear comparison with imported locomotives; and also that Japan, possessing cheap labour and coal, may at no distant date find it unnecessary to go to foreign makers for her locomotives.

256

304

JAPANESE LOCOMOTIVE BUILDING.

APRIL 1895.

TABLE 1. Comparative Working of Compound and Non-compound Locomotives during eight months ending 31 August 1894. Radial tank engine with four wheels coupled { Engine miles, total miles Number of vehicles per train, average Weight of train including load, average, tons Ton-miles, total ton-miles Coal Consumption. Working trains, total Getting up steam, total Total coal consumption Working trains, per engine mile a Total consumption per engine mile b Working trains, per ton-mile c Total consumption per ton-mile d Ton-miles per ton of total coal ton-miles Increase of ton-miles by Compound Saving by Compound, line a "

"

"

"

b

"

"

"

"

c

"

"

"

"

d

No. 221 Compound. 15,672 16.43 122.07 1,913,081

No. 88 Non-compound, 15,603.6 15.73 118.65 1,851,367

Lbs. 301,008 47,301

Lbs. 346,397 40,758

348,809

387,155

19.21 22.20 22.22 24.81 0.1573 0.1871 0.1821 0.2091 12,301 10,713 14 • 82 per cent. 13 .47 per cent. 10 .44 per cent. 15.93 per cent. 12.91 per cent.

The train service from and to Kobe consists of passenger, mixed, and goods trains. Engines Nos. 221 and 88 were worked against each other with all three 3dnds of trains. Of the 15,672 miles run by the compound No. 221 during the eight months, 9,436 were with passenger trains, 2,447 with mixed, and 3,789 miles with goods trains. Exclusive of engine, the weight of a passenger train ranged from a minimum of 41 tons to a maximum of 151, and averaged 90 tons; a mixed train from 74 to 212, averaging 138 tons; and a goods train from 40 to 318, averaging 135 tons. These mileages and weights are from the record of No. 221 compound; but for any practical purpose they apply equally to No. 88 noncompound. Passenger and mixed trains run to the same time-table ; goods trains are timed rather slower. On the section of line worked over by these two engines the stations are approximately five miles apart; and the time allowed requires passenger trains to maintain between stations a uniform speed of about 22 miles per hour, and goods trains about 16. Whence it follows that, in order to keep time, a heavy train has to attain a higher speed than a light one, so as to make up for time lost in starting and stopping. In order to diminish the trouble of obtaining the above comparative statement, the engines were never told off for shunting. No allowance was made for the little shunting sometimes required at roadside stations, putting off or taking on wagons; and therefore enginemiles and train-miles are the same in amount. The reason why the coal used in getting up steam has been kept separate from that used in actually working the trains is that No. 221 compound invariably required more coal to get up steam than did No. 88 non-compound, owing to No. 221 having a slightly larger boiler ; and in this way 3 per cent, of its working economy was sacrificed. Steam is more easily kept up in No. 221 compound than in No. 88 non-compound, especially when the load is heavy. The coal used on the Japanese railways is all mined in this country. It is very suitable for locomotives, and complaints are rare of engines steaming badly. The present cost of coal at the Kobe end of the line is about 5-39 yen per ton ; at the present exchange value of the silver yen, which is generally rather less than two shillings, this cost is equivalent to rather less than 10s. 9d. a ton.

257

APRIL 1895.

JAPANESE

LOCOMOTIVE BUILDING.

305

TABLE 2. First Cost of Compound and Non-compound Locomotives. The first cost of No. 221 Compound built at Kobe was 8,992 yen, made up as follows, a silver yen being then worth three shillings:— Material . . . . . . Labour . . . . . p Coal, oil, coke, and other small stores, cost working shop engines overtime, &c. . Drawing-office materials . • • Total

,

. . Of

$

5,337 = 2,930 =

663 = /

62 =

£

801 440 99 9

. $8.992 = £1,349

Recently three Non-compound radial tank engines, similar to No. 88, cost £1,550 each free on board in an English port. Freight, insurance, and other charges brought the cost of each engine up to £1,713 on board in Kobe. Adding 5 per cent, duty on £1,550, amounting to £77 10s., and cost of landing and placing in Kobe locomotive yard, amounting to £1 10s., the cost is raised to £1,792. The engines had then to be taken out of their packing cases, erected, and pointed; so that by the time they were put into traffic £1,800 would be about the cost of each. There is thus a difference of £450 in favour of the engine built in Japan. The radial tank engines are the smallest main-line engines now used on these railways, and therefore the least expensive in first cost. The construction of No. 221 Compound did not entail the least alteration or addition to the Kobe •workshop plant, nor the employment of a single extra workman of any kind. Whenever it seemed at all likely to encroach on tho requirements of the routine work, inconvenience was avoided by overtime ; and the cost of the extra running of machinery was charged against the new engine, as in the above statement.

TABLE 3.— Valve Setting in No. 88 Non-compound Locomotive., taken from maker's tracing. Lead constant at both ends l-8th inch. Maximum Opening of Steam Port

FORWARD GEAR. BACKWARD GEAR. Percentage of stroke. Percentage of stroke. CUT-OFF. RELEASE. CUT-OFF. RELEASE. Bac-k. Front. Diff. Back. Front. Diff. Back. Front. Diff. pack. Front. Diff.

258

306

JAPANESE LOCOMOTIVE BUILDING.

APRIL 1895.

TABLE 4.—Valve Setting in No. 221 Compound Locomotive) taJcen from enqine when turned out of shop.

* Engine has scarcely ever had to run a train in backward gear. 14° and 13° and 12° are the usual angles of quadrant when running trains. f = full. b = bare.

259

TABLE 5.— Comparative Working of Compound and Non-compound Locomotives on Imperial Government Railways and Chikuho Kogyo Tetsudo Kaislia Hailway. Name of Railway Number of Engine Compound or Non-compound Where built Engine miles, total Number of vehicles per train, average Weight of train including load, average Ton-miles, total Coal Consumption, total „ „ per engine-mile „ „ per ton-mile Ton-miles per ton of total coal Increase of ton-miles over Ghikuho No. 7

miles tons ton-miles Ibs. Ibs. Ib. ton-miles per cent.

Imperial 221 Compound Japan 6403.5 12.89 94.36 604,203 121,170 18.92 0.2005 11,172 30.42

Imperial 179 Non-comp. England 6390.5 13.24 94.71 605,272 138,445 21.66 0 2287 9,794 14.34

Chikuho 9 Compound America 3259.3 26.81 180.70 588,956 135,200 41.48 0.2296 9,756 13.89

Chikuho 3 Non-comp. America 2874.0 23.80 165.20 474,785 124,020 43.15 0.2612 8,576 0.12

Chikuho 7 Nou-comp. America 3228.6 25.67 172.40 556,611 145,530 45.08 0.2615 8,566 —

No. 179 is a duplicate of No. 88 non-compound, and was the first engine worked against No. 221 compound. No. 9 is a four-cylinder compound, having two high-pressure cylinders 11 inches diameter, and two low-pressure 19 inches diameter, all 22 inches stroke; six coupled wheels 4 feet diameter; heating surface 1,379 square feet; grate area 21 square feet. Nos. 8 and 7 non-compounds have cylinders 17 inches diameter and 22 inches stroke; and agree with No. 9 in the other particulars given. All three American engines—Nos. 9, 8, 7—are from the Baldwin Locomotive Works, and all have 291/2tons on drivers, and their working pressure is said to be 140 Ibs. per square inch. A duplicate engine of No. 9 four-cylinder compound from the Baldwin Locomotive Works was borrowed from the Sanyo Railway, and tried on the gradients of 1 in 40 on a section of the Government Railways between Kioto and Baba; in full gear and with the pressure gauge showing 175 Ibs. per square inch it took with difficulty over this section a train of 158 tons exclusive of engine, at a speed of only about five miles an hour.

JAPANESE

LOCOMOTIVE

BUILDING.

Plate 48.

Fig. 1. First Locomotive built in Japan.

261

JAPANESE

LOCOMOTIVE BUILDING.

Fig. 2. First Locomotive built in Japan.

Plate 49.

Fig. 3. Elevation, of

High-pressure

side.

Fig. 4.

Longitudinal

Section.

Fig:5.

Plan.

JAPANESE

Fig. 6.

266

LOCOMOTIVE BUILDING. Plate53.

Front

Elevation.

JAPANESE Outsides - cylinder Fig.

7.

LOCOMOTIVE BUILDING. Plate,64.

Compound

Transverse

Tank

Section

-Locomotive, No 221.

through

Smoke-box.

267

JAPANESE outside- cylinder Fig. 8. Half

268

LOCOMOTIVE

Compound

Transerve

Section

BUILDING. Plate 55.

Tank Locomotive No 221. through

Fire

-box

JAPANESE OutsideFig

cylinder 9.

LOCOMOTIVE

Compound

Transverse

Tank

Section

BUILDING. Plate 5ft.

Locomotive

through

No

221.

Foot plate.

269

Works Visits The Institution was formed in the regions, with an emphasis on the need for local networking between engineers. Special gatherings of members took place to mark occasions such as the 1851 Great Exhibition. These were formalised into an Annual Provincial meeting under Joseph Whitworth's Presidency. The first of these meetings took place in Glasgow in 1856. where he expounded on his vision of "this wonderful power of producing wealth". Thereafter, the Institution met throughout the UK, Europe and in America. An important feature of these summer meetings was the Works Visit. Local committees compiled elaborate activity programmes, including factory inspections, tours of engineering landmarks and temptingly free transport. This benefited participating members, and sites, raw materials, processes and inventions were written up for the Proceedings. The result was a series of useful snapshots of regional manufacturing. The practise continued until well after the First World War. Reproduced here are three specimen visits, including a description of Daimler's automobile works, in only its second year of operation!

270

JULY 1889.

561

THE EIFFEL TOWER. The Eiffel Tower, Plate 61, consists essentially of an iron pyramid composed of four great curved columns, independent of each other, and connected together only by belts of girders at the different stories, until the columns unite at the top of the tower, where they are connected by ordinary bracing. The leading principle followed in the design was that adopted by M. Eiffel in all his lofty structures, namely to give the corners of the tower such a curve that it should be capable of resisting the transverse effects of wind pressure without necessitating the connection of the members forming the corners by diagonal bracing. The actual work of the foundations was commenced in January 1887 ; but a great number of borings had previously been made on the Champ de Mars, which revealed the existence of a bed of hard compact clay 52 feet thick, resting on a chalk substratum, and capable of carrying with safety a load of from three to four tons per square foot. The bed of clay dips slightly from the Slcole Militaire towards the Seine, and underlies a deposit of compact sand and gravel, which affords good material for foundations. At the two foundations furthest from the river the bed of gravel is about 18 feet thick; but at the other two it is much reduced in thickness, and is only met with at a depth of 16 feet below the mean water level of the Seine, being overlaid moreover with soft and permeable deposits, and in these cases it became necessary to employ caissons sunk by the aid of compressed air. The piers are numbered 1,2,3, 4, and are respectively north, east, south, west; the east and south piers are the two furthest from the river, which here flows from north-east to south-west. At these two piers the gravel was met with at a depth of 23 feet. At the north and west piers the sinking was carried through the thinner gravel, and the foundations were made on the underlying bed of fine sand. Each of the four foundations consists of four component piers, which in general are erected on a mass of concrete 32 feet 9 inches long, 19 feet 8 inches wide, and 6 feet 6 inches thick. For one component pier in each foundation however the concrete is 46 feet long by 24 feet wide, being prolonged to the centre of the main pier, so as to form a platform for the elevator

271

562

EIFFEL TOWER.

JULY 1889.

machinery. Each component pier is built with one face vertical towards the centre of the tower, the outer corresponding face being inclined at the same angle as the column of the tower ; the two other faces are vertical and parallel, and the top has been made at right angles to the outer face and therefore normal to the springing of the column. Two bolts, about 4 inches diameter and 4 feet 10 inches apart, are built to a depth of 20 feet into the piers, and are secured to mooring plates 8 inches deep. The concrete used was Boulogne cement; the bedstones on the tops of the piers are from the quarries of Chateau-Landon, and have a crushing strength of about 17,500 Ibs. per square inch, whilst the maximum load to which they are exposed does not exceed 427 Ibs. per square inch. The load on the ground beneath the masonry of the piers is from 3 • 0 to 3.4 tons per square foot. The centres of the four component piers of each foundation are 49 feet 21/2inches apart, and the four foundations form a square of about 412 feet side. The work on them lasted about six months, during which 40,000 cubic yards of earth were excavated, and 16,000 cubic yards of masonry completed. The erection of the lower portion of the columns was effected without difficulty, and the only appliances employed were derricks and winches. The former, though 72 feet in height, were of the simplest possible construction and were made of timber. The four standards, placed one at each angle of the four columns, measure about 31 inches on a side, and were delivered on the ground in lengths weighing from two to three tons; these were handled by means of the derricks, and were bolted one upon the other as the work advanced; the standards were connected by the permanent crossbracing, which held them in position and consolidated the structure. The bolts by which the various pieces were first connected together were afterwards replaced by rivets, as soon as it was ascertained that the different parts of the work were in their proper places. When a height of 50 feet was reached however, this plan had to be abandoned; and for the remainder of the work to the summit cranes were employed which were fastened to the work and carried up as it proceeded. These cranes consisted of a long arm, turning on a pivot, and mounted on a frame in the form of a triangular pyramid

272

JULY 1889.

EIFFEL TOWER.

563

upside down ; the pivot supported a long vertical post, to which the crane arm was hung at about half its length; and the post carried a platform, from which the crane was worked; at the bottom of the frame was another small platform. As the work of building up the columns advanced, there was erected within each of them an inclined path following the same angle as the column, and consisting of two girders, the upper flanges of which were intended to serve as a roadway for the elevators; the upper flanges of these girders were pierced with a series of holes at equal distances apart to allow of the crane being fixed to them at any desired height. Similar holes were made in the lower framework of the crane, which could thus be bolted to the girders and held securely in place. As soon as all the pieces within range of the crane had been raised and riveted, the crane itself was moved upwards: a strong iron cross-beam, through the centre of which passed a screwed bolt, was secured at its ends to the two elevator girders about 8 feet above the crane ; the bolt which passed through the hole in this beam was attached to the crane, and its nut was put in place above the beam; the fastenings of the crane to the riveted work were then removed, leaving it suspended by the screw alone, so that it could be raised to its new position by simply rotating the nut on the screw; after which the crane was again secured, and the cross-beam removed. Four of these cranes were used up to a height of 380 feet, but beyond this two only were employed on a somewhat modified plan. Each crane weighed 12 tons, and had a normal working load of about 2 tons. When 98 feet height was reached, it became necessary to prevent the inclined columns from falling over by their own weight. For this purpose a strong scaffolding 100 feet high was erected on timber piles, driven into the ground to prevent settlement; by these the columns were supported on their inner sides through the intervention of sr.nd boxes, such as are commonly used for the centres of arch bridges. These were found useful for allowing the different members to be easily adjusted when necessary. Altogether twelve stagings were erected, 20,000 cubic feet of timber being employed in their construction ; this however was all the scaffolding required, for as soon as the first story was completed, which is 189 feet above

273

564

EIFFEL TOWER.

JULY 1889.

the ground, the four columns mutually supported one another. To facilitate the erection of the second story, a circular railroad was laid down on the first floor, as well as a ten horse-power portable engine working a crane, from which the chain passed through a square opening in the centre of the platform; and the ironwork when delivered on the ground was hoisted by this crane into wagons on the circular railroad, and distributed by them to the different columns, into which it was raised by the cranes already described. As the work advanced, the dimensions of the iron became lighter and the progress more rapid. A height of 380 feet was reached on 14 July 1888, where the second story is situated. From this point two cranes only could be used, and they were braced firmly together, so as to form in a manner a single structure. The time required to raise them into a fresh working position was forty-eight hours; once fixed, no further change of position was necessary until a complete panel from 30 feet to 40 feet in height had been erected. In addition to these two rising cranes and the one on the first story, another was erected on the second story, and still another on the midway platform which was constructed when a height of 643 feet had been reached. During all stages of the work movable platforms were employed which could be placed in any desired position, so as to bring the riveters within reach of their work. These platforms were protected by handrails and screens, and the precautions taken were such that one man only is said to have fallen from them during the whole course of the work. The tower terminates at a height of 906 feet above the ground with a platform about 53 feet square; the width of the column at this level is 33 feet, the gallery being carried by brackets which project sufficiently to afford a considerable area of platform. Above the platform rises the campanile, in the lower part of which is a spacious and well equipped laboratory, intended for the prosecution of scientific researches. Four arched lattice girders rise diagonally from each corner of the tower, and unite at a height of about 54: feet above the third or top platform. By means of a spiral staircase, yet another small gallery is reached, about 19 feet diameter, surrounding

274

JULY 1889.

EIFFEL TOWER.

565

the lantern which crowns the edifice and brings the total height of the structure to 984 feet. Provision is made for protecting the structure from lightning by cast-iron pipes 19 inches diameter, which pass through the water-bearing strata below the level of the Seine to a depth of 60 feet, their upper ends being connected with the ironwork of the tower. The total weight of wrought and cast iron used in the structure is 7,300 tons, not including the weight of the caissons employed in the foundations nor of the elevator machinery. Iron, and not steel, was used in the construction throughout. The first story, which has an area of 38,000 square feet, is chiefly occupied by restaurants. The second floor, with a surface of 15,000 square feet, is surrounded by a covered gallery 8 feet 6 inches wide, having a total length of 490 feet. The central portion of this floor is occupied by the elevator service, considerable space being necessary to provide for the ascending and descending traffic. On the top platform of the tower there is a large hall covered in on all sides with glass, from which, when the weather is favourable, a magnificent panorama is visible.

275

EIFFEL TOWER LIFTS. Fig.l. Elevation, of Eiffel Feet 984

906-

Midway Platform

276

Engineers 188 9 )

Tower.

Plate 61

DAIMLER MOTOE MILLS, COVENTRY. The Daimler Motor Co., formed in March 1896, was the first company started in this country with the object of entering upon the manufacture of horseless carriages or autocars; and the factory today is the largest and best organized for this purpose in the country. The manufacture is not restricted to autocars, but Daimler motors are made both for the propulsion of launches, and also for use as stationary engines for driving small machinery. The present factory was acquired in May 1896, with the surrounding eleven acres of ground. The premises consist of two blocks of buildings at right angles to each other. Next to the manager's and clerks' offices comes the heavy machine-tool shop, measuring 172 feet long by 35 feet wide, and containing the heavier tools required for turning fly-wheels and heavy crank-shafts, planing gear-boxes and motorframes, and similar heavy work. Above this is the brass-workers' shop, in which are manufactured all the brass and gun-metal fittings. At right angles to these shops are the principal machine-tool shop, 85 feet long by 56 feet wide, and the tool-makers' department; the former is filled with lathes, milling machines, and other tools of a lighter class than those in the first shop. Another workshop of the same size is devoted to the erection of completed motors, where the various parts are assembled together, and fitted into place. Adjoining on one side are the store rooms for finished goods, and on the other side a spacious and well arranged smiths' shop. Another workshop,

277

420

DAIMLER MOTOR MILLS, COVENTRY.

JULY 1897.

like the smiths' shop, has recently been erected close by, in which the motors are tested and finally adjusted, until they are shown to be in proper working order and giving off the required horse-power. Operations having already been hampered by want of room, a new workshop for the special purpose of the motor-carriage building department has just been completed on the other side of the railway line which runs through the works. This shop measures 120 feet long by 130 feet wide, and is being equipped with the necessary machinery for the manufacture of the different portions of the frame and gearing, and with a number of fitting benches and an erecting pit, for properly building up the complete autocars. The drawing office, pattern-makers' shops, stores for finished patterns, and tinsmiths' department, are upon an upper floor, whilst the painting and finishing of the completed articles is carried on in a separate building close by, abutting upon the private railway line on one side, and upon the canal on the other. Between 200 and 300 men are now employed, and for some time autocars have been delivered at the rate of about three per week; this number is expected to be largely increased when the new workshops are in full working order. In the hands of its continental manufacturers the Daimler motor has been used upon the winning carriages in all the chief continental road competitions, notably that held last year, when three carriages fitted with Daimler motors were entered, and finished first, second, and third, accomplishing the journey of over 1,000 miles, under most adverse circumstances as to wind and weather, at an average speed throughout of over 16 miles per hour. The size of motor upon which the works are principally engaged is the 4 horse-power ; and this is so arranged that without alteration of the machinery it can be fitted either to light parcels-delivery vans, or to carriages of a number of different shapes for four to six persons. Other sizes also in hand are 2, 3, 6, and 10 horse-power. Every carriage on completion is put through a road test of at least 100 miles before being delivered to the purchaser.

278

JUNE 1900.

471

THE COLONIAL CONSIGNMENT AND DISTRIBUTING CO.'S FROZEN AUSTRALASIAN MEAT STOEE, LAMBETH. These premises, called " Nelson's Wharf," cover 31/2acres, and are situated on the south side of the Eiver Thames between Blackfriars Bridge and Waterloo Bridge, in Commercial Eoad, Lambeth. They were constructed in 1891 for the purpose of forming a central depot, whence the frozen meat supplies from the Colonies might be received and distributed throughout the United Kingdom, inclusive of London itself. They were erected for Messrs. Nelson Brothers (now the above-named company) by Messrs. John Mowlem and Co., contractors, to the designs of Messrs. Bramwell and Harris, of Westminster. The buildings comprise general offices, living rooms for caretaker and for a foreman, club and mess rooms, &c., stable for nineteen horses on the first floor level, with harness, &c., rooms and van accommodation under; a large covered courtyard for railway and other vans loading from a distributing platform ; a tidal dock for two barges, open to the River Thames, with covered courtyard alongside; cold store for 250,000 sheep ; engine and boiler house ; carpenters', fitters', and electricians' shops; eighteen defrosting rooms, &c., with a jetty 90 feet long by 22 feet wide, having an iron overhanging roof 56 feet wide, sheltering two barges, the site being practically opposite the Temple Pier. A special feature of the cold store part of the building is that it is practically a brickbox on a concrete foundation without any side openings, the two lower floors being below the ground line. The timber structure supporting the six floors is self-contained and independent of the brick walls, the insulation being formed of hard-wood sawdust 12 to 15 inches thick encased in the usual double thickness of matchboarding with tarred paper sandwiched between. Cargoes of frozen beef and sheep are received in barges loaded from ships in the docks and in the river, and have to be raised to

279

472

COLONIAL FROZEN MEAT STORE.

JUNE 1900.

the sixth or "Receiving and Distributing Floor" by hydraulic cranes, and elevators termed " Sheep Syphons." The latter are endless chains having arms upon them across which the sheep are laid, and are driven by a 3-cylinder Brotherhood hydraulic engine, and deliver the sheep on to a receiving table on the distributing floor. The apparatus is adjustable to the position of the barge at any state of the tide and whilst at work. From the table the sheep are sorted on to trucks and conveyed by hydraulic jiggers with wire-rope suspended cages on to the various floor-levels of the cold rooms; from which they are again raised by similar means when required for distribution, the trucks being lowered to the courtyard-level van-loading platform by hydraulic direct-acting lifts. There are nineteen hydraulic elevating machines, made by Messrs. Way good and Co., eight of which, together with a steam-driven sheep-elevator on the jetty, are capable of dealing with barge cargoes. It is therefore a comparatively easy matter to receive 10,000 or 12,000 sheep per day as an ordinary day's work; whilst as many as from 17,000 to 20,000 sheep in and out of stock have been handled in one day. Two systems of refrigeration are installed; one, the De La Yergne direct-expansion of ammonia process, for which there are two machines, each of 80 horse-power with 12-inch by 24-inch double-acting compressors, and compound-condensing steamcylinders, with about 12 miles of 2-inch expansion pipes in the cold rooms, made by Messrs. L. Sterne tand Co., of Glasgow; the other system is a cold-air compression and expansion machine of about 300 horse-power, made by Messrs. Haslam and Co., of Derby, with airdistributing ducts formed in the concrete foundation of the building, but which is seldom used. The " defrosting " plant is the invention of Sir Montague Nelson, Chairman of the, Company, and consists of a series of rooms fitted with ammonia expansion pipes, and steamheating pipes, &c., the purpose of which is to thaw out, or " defrost" beef and mutton to the extent of 4,500 sheep and 2,000 quarters of beef per week, giving a daily output for market deliveries. This process is both natural and efficient.

280

JUNE 1900.

COLONIAL FROZEN MEAT STORE.

473

Electricity for lighting the cold rooms, offices, and premises generally, is generated by two of Messrs. J. and H. Gwynne and Co.'s vertical compound engines, each coupled direct to a Paterson and Cooper low-speed dynamo. Hydraulic-power water is supplied for the lifting apparatus at a pressure of 750 Ibs. on the square inch, by a Smith and Vale compound duplex steam-pump, having 6-inch water-rams, working on a 15 feet stroke, 12 inches diameter accumulator. The London Hydraulic Power Co.'s service water at the same pressure is laid on through meters to the ring-main water service-pipes as a stand-by, when the pump is shut down for repairs. The three Lancashire steel boilers 7 feet 6 inches diameter by 30 feet long, of which two are in use at one time, made by Messrs. Galloway, and fitted with Vicars' mechanical stokers, supply steam at 140 Ibs. pressure. Service pumps for condenser-water and boilerfeed water, a filter, heater, and general steam-condenser and airpump, complete an installation running day and night throughout the year. This establishment is now dealing with frozen colonial sheep and lambs at the rate of about 1,500,000 a year, also many thousands of single joints, and about 100,000 quarters of beef from New Zealand and Australia.

281

This page intentionally left blank

Henry Selby Hele-Shaw FRS (1854-1941) Apprenticed to Messrs Rouch and Leaker, Hele-Shaw was an early beneficiary of the engineering scholarship scheme established by Sir Joseph Whitworth (1803-1887). He won scholarships and prizes for the years 1876-1879 and these enabled him to study at the University of Bristol. He then became its first Professor of Engineering in 1881. Four years later he became the first incumbent of the Harrison Chair at Liverpool University College, where he established an active department based around the Walker Engineering Laboratories. There followed a varied and useful series of researches in areas as diverse as marine engines, liquid flow and magnetic fields. By the turn of the century, Hele-Shaw was devoting time to newer fields. He organised commercial vehicle trials on behalf of the Liverpool Automobile Club and became a founder of the Royal Automobile Club. His 1900 paper for the Proceedings was the earliest full paper on road vehicles to be published by the IMechE. Later, a more informal talk to the Institution on developments in flying machines completed a unique "double-first" in the promotion of transport engineering. Following a brief sojourn at the Transvaal Technical Institute, thereby founding his third university engineering department, Hele-Shaw settled in Westminster as a consulting engineer. His new location allowed him to spend considerable time in serving on the Institution of Mechanical Engineers' Council and committees. He was elected President in 1922. He developed many inventions, perhaps the most important of them being the variable pitch airscrew. In 1924 this was little used, but came to be regarded as essential by WWII. Hele-Shaw's Presidency was marked by his contributions to engineering education. He founded the Whitworth Society for former Whitworth scholars such as himself. More importantly, for the coming generation of engineers, he proposed the National Certificate in Mechanical Engineering, to be jointly awarded by the IMechE and the Board of Education. This set a pattern to be repeated for qualification in many other disciplines. On his retirement, and in recognition of his own origins, he instituted the Hele-Shaw National Certificate Prize, to which the Institution added the Hele-Shaw Medal.

283

APRIL 1900.

185

ROAD LOCOMOTION.

BY PROFESSOR H. S. HELE-SHAW, LL.D., F.E.S., Member, OF LIVERPOOL.

There are strong reasons for thinking that the subject of mechanical propulsion upon common roads has now reached a point when it deserves the very careful consideration of mechanical engineers. The idea of bringing the matter generally before the Institution for discussion is due to our President, whose far-reaching judgment will be admitted by all. The title of this Paper must be admitted to be very comprehensive, but it seems that what is needed at this time is a discussion of the general principles of the engineering features of the question, rather than a detailed description of any particular system. For many years the uses and importance of the traction engine have become more and more recognised, and its possibilities in connection with the present war have quite recently been brought very strongly before the public. This engine, the work of which covers only a portion of the field for mechanical propulsion on roads, has been very fully dealt with before this Institution and elsewhere, and it will be in the first place instructive to consider what has led to a general revival of a movement for lighter road-locomotives which about seventy years ago, in the days of Hancock and Gurney, reached a point that for a time appeared to be leading to permanent results of the most important kind, but which ended in complete failure. In one sense this revival is undoubtedly due to the passing of the Locomotives on Highways Act in 1896, previous to which, for more than twenty years, a law had existed, popularly known as the "Man with the Red Flag"

285

186

ROAD LOCOMOTION.

APRIL 1900.

Act, which made it impossible for any self-propelled vehicle to proceed at a rate of more than four miles an hour. The immediate cause of the passing of this Act was the attention aroused in this country by the successful introduction of the motor vehicle for purposes of pleasure in France, where the red flag was not at any rate used specially for obstruction on highways. This freedom from legal restriction enabled an enterprising paper, " Le Petit Journal," to organise in 1894 a trial of motor vehicles between Paris and Rouen, which was so successful as to lead to one on a larger scale between Paris and Bordeaux over a distance of more than seven hundred miles. These trials proved conclusively the great possibilities of motor vehicles, and attracted much attention in this country, where after the first exhibition and trial in England, which was originated by Sir David Salomons in 1895 in the grounds of the Local Agricultural Society at Tunbridge Wells, successful measures were taken to obtain a more enlightened legislative treatment. We must however look deeper for the real causes of the present movement, which can be traced to the gradual feeling amongst all classes of the community that modes of transport both for purposes of pleasure and business on the roads had not kept pace, or indeed had made little progress at all, compared with the great changes which had been effected in speed, comfort and convenience, in the direction of locomotion by rail. Mr. Samuel W. Johnson, speaking of the progress of railways in his Presidential Address* before this Institution, showed that in thirty years the annual train-mileage had increased from 200 millions to 850 millions, and remarked that " Our iron roads are the arteries and veins of the nation." Pursuing this very true and striking analogy farther, it may be said that the capillaries and smaller blood vessels are in their way just as important as the larger veins. Now while the railway arterial systems have developed enormously and satisfactorily, there is still much room in present modes of collecting and distributing goods for improvements, which would materially benefit the trade and commerce of the country. * Proceedings, 1898, page 149.

286

APRIL 1900.

ROAD

LOCOMOTION.

187

Railways are undoubtedly the cheapest system of land carriage for long distances, but there is a minimum distance below which the disproportion between haulage and terminal charges operates to their detriment. This feature of railway transport and the serious consequences of " breaking bulk " are exhaustively dealt with in an interesting and important report published in 1898, by the Special Light Eailways Committee of the Liverpool Incorporated Chamber of Commerce, in which report it is made evident that Liverpool in particular suffers from these causes, and what is true of Liverpool is probably true of other great commercial cities. Cartages and terminals exceed the haulage charges over short distances by rail, whilst they become only a very small percentage of the whole when the distances are considerable. It must be obvious that a motor vehicle, which can travel from any one point to any other, which absorbs the short cartages into one straightforward journey, and which absolutely eliminates railway terminal charges, has a wide and promising scope for application. Mr. Alfred Holt, one of the leading shipowners of the country, has for many years urged the necessity of obviating these terminal charges, concerning which he states that the matter of handling is a " giant, and the transport a dwarf, and the giant is daily growing larger, and the dwarf smaller." The relation between these two charges may be shown in a very striking manner by plotting the terminal and conveyance charges, worked out in pence per net ton-mile taken from the Government Blue Book of 1892, dealing with the rates and charges, and is given in Fig. 1 (page 188). This curve shows clearly that up to 40 miles there is a field for a system of conveyance in the working of which terminals are not incurred, and it will be at once appreciated by engineers. Apart from these considerations, there can be no doubt that with the rapid means for communication of ideas by telegraph and telephone, and of passengers and goods by means of the railway, the general want is felt of a more speedy means of transport by road. The great improvements which are needed in our road traffic have been set forth by Major E. E. B. Crompton, E.E., at present in South Africa, and one of our old Members, in a Paper read by

287

188

ROAD LOCOMOTION.

APRIL 1900,

him before the Automobile Club, in which he clearly shows the vast and beneficial changes that the general introduction of the motor vehicle would effect in the relief and expedition of traffic in our cities, and especially in the metropolis. Nor must the hygienic considerations be overlooked. This subject has been ably dealt with by Mr. E. Shrapnell Smith, Honorary Secretary of the Liverpool Self-Propelled Traffic Association, to whom the author is indebted for valuable assistance in many points dealt with in this Paper and particularly in connection with the final section. In his Paper, read before the Congress of the FIG. 1. Relation between Terminal and Conveyance Charges on Railways. (Classes 1 and 4.)

Sanitary Institute at Birmingham in 1898, he points out that with motor vehicles, not only will our streets be less offensive, especially in summer weather, but exposed food stuffs will less frequently afford a nidus for organisms conveyed by disseminated particles from the roads; and further, that the disintegrating effect of the horses' hoofs, which accounts for most of the dust of summer and the pasty slime of winter, will be to a great extent obviated. The sanitary advantages of the motor vehicle have also been recognised and strongly urged by many medical officers of health and surveyors to municipal and urban councils. This shows, then, what forces are at work urging us, both for light and heavy traffic, in the

288

APRIL 1900.

ROAD LOCOMOTION.

189

direction of utilizing more efficiently and with mechanical power the 100,000 miles of road which we possess in this country. From this side of the question we naturally turn to consider the difficulties of the problem, and it must at once be admitted that these difficulties are very great. The author has frequently seen the subject referred to as a question of mere mechanical detail, and the progress of the railway locomotive mentioned as a proof that these mechanical difficulties will be easily and rapidly overcome, Moreover, the whole blame for small progress made, and for previous failure, is often thrown upon restrictive Acts of Parliament. The truth is that the argument of the railway locomotive, so far from giving any grounds for the hope of an easy solution of the problem of road locomotion, really tends in the opposite direction. In the first place, railways are one of the most striking examples of the nature of mechanical progress first pointed out by Keuleaux, that machines became more and more perfect as their restraint by what he called " pairing " was more completely effected, i.e., as the mechanical boundaries compelled the parts to move with more certainty under required conditions. The provision of a suitable track, upon which the train moves and by which its motion is guided, is the real secret of railway development. Hence it is that with a steel wheel rolling upon a hard smooth track, a continuous increase of weight, and of tractive force, together with increase of speed is enabled to be obtained. The conditions of the historical " Rocket" were a weight of six tons, a speed of from 20 to 30 miles an hour, and a load of twenty tons, while the modern locomotive and tender together weigh one hundred tons, having a speed of over sixty miles an hour, and drawing a load of three hundred tons. Now this result has been obtained by increasing the number of wheels, until the locomotive and its tender may have the weight distributed over from sixteen to twenty wheels, each resting upon a hard smooth surface of contact, whereas the motor vehicle, at any rate at present, is limited to four wheels, which have to run upon an uneven surface which, if it is hard, intensifies the action of shocks and vibrations, and if it is soft, causes an enormous amount of resistance. The load thus being on four wheels, both this sinking and shock are magnified as the load is increased, and therefore

289

190

ROAD

LOCOMOTION.

APRIL 1900.

inventive effort has been naturally almost entirely directed to lightening the working parts for obtaining a given power, and this correspondingly diminishes the tractive adhesion which is a necessary feature for successful working. In short, the conditions of the problem are such as to involve improvement exactly in the opposite direction to that in which the railway locomotive has been successfully developed. No doubt the progress of invention will ever increasingly enable a greater amount of power from a given weight of motor to be obtained, but the surface to be moved over, which is the real difficulty of the road locomotive, will remain the chief factor of the problem. The First section of the Paper is therefore devoted to the mechanical problem of the behaviour of the wheel upon the road, and the progress which has been made in this direction. The Second section deals with Steering and Turning. The subject of Motive Power is treated in the Third section, which is divided into internal combustion motors or oil engines, external combustion motors (steam) and electrical motors. In this section Transmission and Gearing are briefly dealt with. Finally, a summary is given of the results which have been obtained, and certain general conclusions, together with Appendices containing some notes and Tables.

I. PROBLEM OF THE

WHEEL ROLLING UPON A ROAD.

When a wheel with a hard rim rolls upon a level hard surface, every point upon the wheel follows a curve as shown in Fig. 2, a, and since each point on the tyre comes in succession perpendicularly upon the surface beneath, there is no appreciable resistance to the

290

APRIL 1900.

ROAD LOCOMOTION.

191

motion. When, however, the surface beneath is either soft, Fig. 2, b, or! irregular, Fig. 2, c, the wheel no longer rolls in the same way, and the invaluable properties which it possesses are in a greater or less degree destroyed. FIG. 2.

Now it is difficult, from any data at present available, to separate the amount of resistance respectively due to each of the two FIG. 3. Resistance to Traction.

Asphalte Tram Rails Well laid stone

Best Macadam ................. Paris Streets Ordinary Macadam (a) Ordinary Macadarris (b) Dry Meadow. Hard Dry Clay Cobble Stones Ordinary Road & Gravel Ordinary Cobble Stones Fresh Earth __________ Sand if dry & loose

foregoing causes. But in Fig. 3 are plotted, from Telford and Babbage's data, the resistance to traction on roads of various kinds, and here it would appear as if a soft road involved greater resistance than an irregular one, and was more to be considered.

291

192

ROAD LOCOMOTION.

APRIL 1900.

The truth is that, when we are considering the question of the motor vehicle, the opposite is the case, for while the amount of the resistance due to the softness of the road remains practically thesame as the speed is increased, the resistance due to obstacles which FIG. 4. Horse-Power Velocity Curves.

cause shocks and vibrations rises rapidly. MM. Bovaine et Julien in their Tableaux Nurnerique Graphique, have investigated the horsepower up to speeds of 50 kilometres (31 miles) an hour for varying loads. It is not necessary to reproduce their numerical investigations, but the series of curves which the author has translated into English

292

APRIL 1900.

ROAD LOCOMOTION.

193

measures, Fig. 4, indicates clearly the great increase of resistance as the speed increases. Now it must be remembered that it is not merely the increase of resistance itself that becomes serious, but that the vibration which occurs tends greatly to destroy the structure of the motor vehicle, and makes the problem of keeping in working order the machine parts which it will be seen are necessary for automatic action, increasingly difficult. Mr. John Brown,* of Belfast has invented an instrument, which he has called the " Viagraph," with the object of obtaining autographic records of the surface of roads. The use of this instrument indicates in a remarkable way the vibratory effects produced according to different states of roads of nominally the same character, and the curves which he has given of macadam roads in various parts of the country show remarkable and instructive differences. The city engineer of Liverpool, Mr. J. A. Brodie —who is using one of these instruments—has kindly taken, specially for this Paper, a series of curves which are reproduced on Plate 29. From these diagrams will be seen the difference in the character of vibration on asphalte and wood pavement compared with that on sets or macadam road, and shows to what different influences a motor vehicle may be subjected when it has to run indiscriminately upon these various kinds of road, unless some special provision is taken to counteract such influences. Fig. 5 gives a good illustration of the severe effect produced by a bad crossing. Now obviously the remedy for shocks is a means of causing the vehicle and its load to ride over the obstacles without being lifted bodily; since a reference to Fig. 2, c (page 191), shows that a lifting of wheel and axle must take place unless the unevenness of the road is destroyed by being pulverized or removed, and this lifting, which amounts to a change of direction of the vehicle and its load, must be accompanied by shocks and consequent loss of power. Springs under the body of the car are the natural method of achieving this result, since even if the wheel rises, the springs give so as to allow the main body of the load to pass onwards over * Proceedings, Belfast Natural History and Philosophical Society, 1899.

293

194

ROAD

LOCOMOTION.

APRIL 1900.

the obstacles without being lifted bodily. But even with springs, the periphery of the wheel has to sustain a shock which causes noise and destruction of the wheel itself, and the best result is obtained by placing an elastic medium between the wheel and the road. More than fifty years ago, a brougham was running in London with pneumatic tyres which were the invention of an engineer, Mr. R. W. Thomson.* This invention did not attain a practical success at the time, for reasons so well understood by engineers, viz., that the mechanical appliances and materials available at that time did not enable the inventor successfully to cope with the practical difficulties in the way of its commercial construction. Since its revival in recent years, it has already played such an important part in road locomotion, and is probably destined to play a much greater part, that a few facts concerning it may well be brought forward. The action of the pneumatic tyre is really two-fold; it not only interposes the desired elastic cushion between the irregular road and the vehicle, but it does so by a continuous spring of compressed air, extending round the periphery of the wheel. This air, when once compressed by the load being placed upon it, absorbs—so to speak— the obstacle with only a temporary deformation of the elastic covering, so that no further work is done in the compression of the main spring itself. Photographs have been taken by the author of a pneumatic-tyred wheel passing over obstacles of various kinds, and it can be seen that, even allowing for a certain amount of preliminary compression upon the flat surface, how small a distance the load resting on the axle has been raised, and that the idea involved in the expression " absorbing an obstacle," when the obstacle is a small one, is practically correct. Many measurements have been made to compare the resistance to motion of a hard tyre with that of the pneumatic, but these are not very useful unless taken at varying speeds, since it is when the speed is increased that the differences become most marked. The best and most instructive results are those given by M. Michelin in a Paper * See Abridgments of Specifications, 1G25-1866, No. 10,990, 1845.

294

APRIL 1900.

11OAD LOCOMOTION.

195

read before the French Society of Civil Engineers a few years ago.* These trials were conducted for all kinds of loads, in all kinds of weather, upon varied kinds of roads and at various speeds, and though it would require a separate paper to deal with them adequately, the general conclusions may be briefly summarised as follows :— (1.) The force required for the pneumatic tyre increases very little from walking to trotting, while that absorbed by an iron tyre perceptibly increases with the speed. (2.) The advantage of the pneumatic over the iron increases rapidly as the speed increases. Taking 100 as the force required for the pneumatic tyre and iron tyre at a slow speed, the resistance of the iron tyre increased at an ordinary trot to 126, and at a quick trot to 164, and probably if the matter had been investigated for ordinary speeds of an autocar, the relative resistance of the iron tyre would have been found to rise even more rapidly. (3.) The solid india-rubber tyre is better than the iron-tyred wheel in certain cases, especially at the trot, if the surface be sticky, very irregular, or covered with snow ; but it becomes inferior to iron if the surface be hard and smooth. It never gives a much better result than the iron wheel, and it always remains vastly inferior to the pneumatic. On the other hand the pneumatic is fifty per cent, better than the iron tyre. As M. Michelin remarks, "It is a curious thing, but we have known many people who have only seen pneumatic tyres, and who have never tested them, affirm that these thick tyres must drag heavily. It is specially on bad ground, in mud, in snow, that the advantage of the pneumatic is made clear." It is interesting to note as confirming this statement, that it was recorded in the automotor journals that during the recent heavy snow and bad weather, automobilists were able, to their great delight, to use their vehicles freely, the pneumatic tyres appearing to be almost unaffected by the bad state of the roads. In addition to the numerical results obtained by M. Michelin, a series of graphic records were obtained from the * Translated and published in a series of articles in the "'Autocar," commencing 15th August 1896.

295

196

ROAD

LOCOMOTION.

APRIL 1900.

body of the vehicle, which explain to the eye at once better than any figures, the reason of the satisfactory results obtained by the pneumatic tyre. One or two of these are reproduced on Plate 30, in which the curve traced by the vehicle with iron tyres meeting an obstacle is shown by a full line, while the behaviour of the vehicle fitted with pneumatic tyres is shown by the dotted line. In these curves, instead of the sharp jump indicating a violent shock and loss of power which occurs when the iron tyre meets an obstacle, it is evident that the pneumatic carries the vehicle over with an easy and gliding motion. This, which is evident in the small obstacle, Fig. 14, is much more so in the case of the larger obstacle, Fig. 15, while when the three obstacles are interposed, Fig. 16, the beneficial effect of the pneumatic tyre is very remarkable. For heavy trafiic, when it is remembered that the load has to be concentrated upon the point of contact on the periphery of four wheels, it is no wonder that the wheels have hitherto been almost entirely made with iron tyres. At the first Liverpool trials of heavy motor vehicles great trouble was experienced with the wheels, which had been made in the best possible manner for ordinary vehicular trafiic. In almost all cases the wheels showed signs of the severe stresses and shocks to which they had been subjected, and some of the vehicles utterly broke down in consequence. The construction of the wheels themselves has latterly been the subject of much careful design on the part of makers of motor vehicles, the wheels having necessarily not only to bear the actual load, but to transmit tractive force from the motor. Plate 31 shows some of the wheels which have been specially designed by the makers of motor vehicles. These wheels show in each case the arrangements for driving, and this feature is a vital part of the design of the wheel. The iron portion has been blocked in to indicate which is metal and which is wood, and the nature of the construction will be evident without any detailed description. It may be pointed out that Messrs. Coulthard use a wheel entirely of iron, Fig. 20, which seemed to give very satisfactory results at the Liverpool trials, where it must be noted that part of the 40 miles run was over a road paved with cobbles of the sort well known in certain parts of

296

APRIL 1900.

ROAD

LOCOMOTION.

197

Lancashire, affording a test scarcely to be surpassed in severity. The last wheel shown on the series is that of Messrs. Bayley which differs from the other driving wheels in the important detail of being coned or dished, Fig. 22. A great deal might be said in considering whether the advantages of a coned wheel for heavy traffic are not more than counterbalanced by the constant tendency of such wheels to run outwards, since when the axles are not horizontal, they can only be made to run in a straight line by a certain amount of slipping constantly going on at the tyre, and a constant outward drag upon the axle. Messrs. Bay ley's wheel, however, worked very well at the Liverpool trials and subsequently, which is after all the main point to be considered. There is another important peculiarity, and that is that the spur wheel attached to the driving wheel is annular, and is driven by an internal pinion. This affords considerable protection from dust and dirt, and enables the outside of the annular wheel to be used very effectively as a brake wheel encircled by a band brake. The spokes are of oak with ash felloes, the iron tyre being 5 inches wide. In consequence of the new design and special construction of the wheels adopted by most of the makers, they were enabled to stand much better at the second Liverpool trials, but even in the second Report, the judges wrote in their special conclusions as follows:—" The wheels and tyres were generally efficient, but concentration of heavy loads upon the present small area of wheel contact is a serious difficulty in the problem of goods transport by motor vehicles, and constitutes the chief mechanical cause of the slow progress made." Quite recently one or two makers have been appreciating the great difficulties of this question, and have tried to adopt solid india-rubber for tyres, and Fig. 21, Plate 31, shows Messrs. Simpson and Bodman's wheel, in which the india-rubber is shown cross-hatched in section, and from which they have obtained very satisfactory results. The action of the solid tyre, however, differs in a most important respect from that of the pneumatic, namely, that although deadening shocks, there must always be a great loss of energy consequent upon the continual expansion of the different parts Q

297

198

ROAD

LOCOMOTION.

APRIL 1900.

of the substance itself, which is totally different from the behaviour of the air-cushion in the pneumatic tyre. In Michelin's experiments the solid india-rubber showed itself always greatly inferior to the pneumatic tyre (page 195), and under some circumstances inferior to the iron tyre itself. It was put into competition with the pneumatic tyre in the early days of the cycle, and in that competition has disappeared for ever. The same process of competition seems to be going on in the case of the light motor vehicle, though the difficulties are more than proportionally greater as the load upon each tyre increases. In this country the Dunlop Company have been for years at work improving the tyres for motor vehicles, and the section of their latest production is shown on Fig. 10, Plate 30. It will be seen that the tread has been made thicker where the greatest wear takes place and where there is the most liability to puncture, while the external circular form is retained. It is found necessary to use the very best india-rubber for the external portion, in order to admit of the necessary deformation, whilst the inner part of the body of the outer tube has to be strengthened by means of layers of woven canvas insertion. The steel rim on the wheels holds the two enlarged edges of the outer covering of the tyre so that when the inner tube is inflated the tyre is held in its place without the necessity for any internal wiring or fastenings. Special machinery is required to make this tyre at all cheaply. This Messrs. Dunlop are now putting down, and it is hoped that the pneumatic tyre for light motor vehicles will before long be obtained at prices which will enable it to be universally placed on such vehicles. Fig. 11 shows a section of the Michelin tyre, and it is interesting to note that M. Michelin, who, as he naively remarks, commenced his research on the subject " with the object of proving that the English did not now possess the monopoly for making pneumatic tyres," has now succeeded in producing tyres that, it is only right to say, are preferred by many users of motor vehicles in this country. This is not so surprising when it is remembered that the French have been far ahead of us in the production of light motor vehicles, and the demand for suitable tyres for such vehicles has been very keenly

298

APRIL 1900.

ROAD LOCOMOTION.

199

felt. The Callus, which is also a French tyre, is shown in Fig. 12, and is seen to be almost identical with the Dunlop tyre, except that in the important matter of tread there is not so much extra thickness allowed, and it is obvious that the Dimlop tyre would carry a much greater load as well as have a longer life. The Goodyear, Fig. 13, which is an American tyre, differs from the three preceding ones in that, like most American tyres, there is no inner tube, neither does it depend at all for attachment to the rim by inflation, but is held on by means of small screws, the nuts for which are inserted in the body of the tyre. There is not much appreciable difference of thickness for the tread of this tyre, which is circular in section, but it is sufficiently/ thick for all practical purposes. With heavy traffic, where noise and vibration are not so fatal to success as in the case of pleasure vehicles, the importance of having pneumatic tyres cannot be said to be so great. On the other hand it becomes of more importance to carry if possible a heavy load easily over an obstacle without shock than a light one, as the destructive effect on the vehicle of the inequalities of the road is naturally greater. This problem of spring wheels in connection with road locomotion is one which has exercised makers of traction engines for many years, but ten years ago it was stated,* " The exertions of inventors during the last quarter of a century seem to have boen inadequate to the production of a wheel with elastic tread, which will satisfy all the complex and most difficult conditions governing the use of traction engine wheels. Not a few of the most eminent and successful makers of traction engines have abandoned elastic wheels altogether, and resorted to springs between the main axle and the engine, and they have, on the whole, been successful." In spite of this, however, when one looks at the original design of Thomson's wheel in his patent, one cannot help feeling that he understood what was required, and that although the 'practical difficulties may be great, there are no mechanical impossibilities in the production of a pneumatic tyre for the very heaviest vehicular traffic. With any existing system, in which four wheels are used, the problem is a * " The Engineer," 12 December 1890, page 469. Q 2

299

200

ROAD LOCOMOTION.

APRIL 1900.

difficult one, because of the concentration of the loads upon such a limited area of support. The pneumatic tyre however extends this area of resistance by yielding, so that the area in contact is much greater than in the case of an iron-rimmed wheel, especially when running over sets or hard ground. Beyond this, it is quite conceivable that, just as in railways the number of wheels has been largely increased until a modern bogie carriage has commonly twelve wheels supporting it, it may be found economical to support a motor vehicle also upon a much greater number than at present.

II. STEERING AND TURNING.

Steering.—The steering of motor vehicles is evidently a very important part in their design, and it may-at once be said that with one or two exceptions, the great majority of motor vehicles are steered upon the principle which was invented by Ackermann as long ago as 1818, the leading feature as stated in his own words being that a vehicle " will turn within a small compass and with safety, because the wheels do not materially alter their bearing upon the ground when they are placed in the greatest degree of obliquity." The essential principle of the Ackermann system consists in replacing the pivoted fore carriage of an ordinary vehicle which has one axle for the two wheels, by two short pivoted axles each carrying one of the steering wheels. The conditions of correct running of the wheels are that when they are turned, their normals intersect on a point on the line of the axles of the driving wheels. In the Ackermann invention, this is obtained approximately by a pair of short levers inclined either inwards or outwards to the longitudinal direction of the vehicle. Fig. 23, Plate 32, shows a Daimler frame in plan in which the short levers are inclined away from each other, the connecting bar between the two being in front of the axles of the steering wheels, whereas Fig. 24 shows the plan of the Clarkson and Capel steering arrangement in which the levers are inclined towards each other, the connecting bar being behind the axles of the steering wheels. The other details of the steering arrangement are obvious without further

300

APRIL 1900.

ROAD

LOCOMOTION.

201

explanation. In the case of wheels steered on this plan, it is evidently of great importance to have the wheels coned so that the point of support of the wheel is as nearly as possible under the pivot about which each of the short axles turn. The reason for this is clear from the drawings, since the line of stress now comes more directly under the pivot which is vertical, and thereby largely reduces the bending effect or leverage upon the pivot. At the samo time this coning makes it very much more difficult to effect the steering when the vehicle is being manoeuvred, and especially when standing still, as the wheel tends to run round the point where the direction of the short axle meets the ground, and anyone who has watched the effort required on the part of the driver of a heavy motor wagon under the circumstances, even when the steering wheel transmits motion through gearing, will be at once convinced of the reality of this objection. Although the Ackermann invention is found to give sufficiently correct results in practice, it only approximately fulfils the required conditions, and in extreme positions of the wheels the axles may be very far from their correct positions. Mr. Davis has applied a geometrical principle in his steering gear, which gives absolutely accurate movement of the wheels for every position of the steering lever. The suitability of the two short pivoted axles for steering purposes is dealt with in the Appendix reports. Fig. 25, Plate 32, represents the plan of the fore wheels of the carriage, A1 A2 being the centre line of the driving wheels, B B the axes of the steering wheels in the initial position. If the directions of the short levers D D are always caused to intersect in the line C1 C2, which is an equal distance in front of B B that A1 A2 is behind it, then the normals of the wheels will themselves always intersect in the same point of the line A1 A2. Thus the directions of the dotted position of the levers are shown intersecting C1 and the normals of the wheels simultaneously intersect each other at the point A1. This condition can be effected by causing the two levers to be actuated by sliding straight guides, the bar connecting which is made to move parallel to itself, and at an equal distance from the line of the pivots upon which the short axles turn. Pig. 25 shows the plan of the actual arrangement in

301

202

ROAD

LOCOMOTION.

APRIL 1900.

which the levers D D are enclosed in the sliding tuhes E E, which are actuated by the guide bar F, the dotted lines showing the various parts in one extreme position of the steering wheel. Turning.—There is another feature of the motor vehicle to which care and attention has to be given, and that is the provision of some differential gearing for the driving wheels when any turning movement of the vehicle takes place. The operation of differential gearing, or jack-in-the-box as it is commonly called, has long been understood by engineers, but a number of forms have been designed to meet special requirements of the motor vehicle. The ordinary form of differential motion consists essentially of three bevel-wheels, the middle one being the driver actuated from the motor. Now when the vehicle is turning a corner the inner driving-wheel tends to lag behind the other one, and hence the resistance is not the same on each wheel. Now since the two bevel-wheels on either side are connected respectively at equal speed ratios with the two drivingwheels, it is clear that one will be allowed to run round more than the other, and so the two driving-wheels can accommodate themselves to the required conditions. The Beeston Company have a very ingenious arrangement by which an epicyclic train of spur-wheels and pinions is employed to effect the same result.

III. MOTIVE POWER AND TRANSMISSION.

Although the action of the wheel upon the road has been referred to as being at the root of the whole problem, the question of motive power attracts the most interest and gives the greatest scope for inventive talent as well as most trouble in actual use. The three kinds of motive power at present adopted, viz. oil, steam, and electricity, have fairly well-recognised spheres of operation, although this must be regarded as by no means a final condition of things, or giving a limitation to the employment of each type of motor. Thus, although at present oil-engines are used for light motor vehicles and steam for heavy traffic, there are very

302

APRIL 1900.

ROAD LOCOMOTION.

203

ingenious steam motor-cars both in this country and abroad, while light oil-engines have been applied in France and also in this country in connection with heavy traffic. Moreover, there are other kinds of motive power, such for instance as compressed-air, which is being tried in America. Oil-engines, or as we will call them internal-combustion engines, have by a process of the survival of the fittest been found so far best suited for light motors and pleasure vehicles. The cycle of the gasengine is really complex, but these motors have been brought to a high state of perfection, so that upon being started they are found to work for long distances without any attention. If really well designed and constructed, and used with a moderate amount of care, they need little repairs or adjustment, while the objection of smell, vibration and danger from the use of light petroleum spirit with a low flash point have all been much reduced, and each year sees an increasing number of places in town and country where petroleum spirit can be obtained. Still the objections above-mentioned must be admitted to exist, and, together with the great expense of pleasure vehicles, have to a certain extent prevented their introduction becoming general. Again, an oil-engine, which has little elasticity in regard to an increased demand for power when ascending a hill, requires elaborate gearing for change of speed, which may be after a time, if not at first when the car is new, a very noisy and objectionable feature. Heavy-oil engines for internal combustion have been tried for motor vehicles, but the difficulties of starting and smell have not yet been satisfactorily overcome. In connection with this matter some recent experiments by Mr. Henry Barcroft, of Newry, who has succeeded in maintaining constant mixture and compression under varying load, promise well for the future. Steam, or external-combustion motors, require not only a generator or boiler, but also a condenser, in addition to the steamengine itself. The author is aware that the latter is not used with all motors, but in winter the cloud of steam which must be visible in damp cold weather at a little distance from the exhaust, even if the steam is superheated, really contravenes the Act which states, "No smoke or visible vapour must be emitted, except from any

303

204:

ROAD LOCOMOTION.

APRIL 1900.

temporary or accidental cause." Steam introduces a more complicated array of pipes and fittings, and requires more attention and skill in working, but it is highly probable that such improvements will be made in connection with steam motors, that no skilled attendant will be necessary. There is already at least one steam system which is entirely automatic, whilst others are to a great extent so. It is not too early to speak as to the practical and commercial success of any of the systems using steam (vide Appendices), but if a condensing steam-engine, automatic in action, with a boiler which is perfectly safe from any fear of explosion, can be produced, it may safely be predicted for it that there is a great future before it, both for light and heavy traffic, as it would have the advantages of great power and elasticity, freedom from smell, and if using heavy oil or even coal or coke, would be free from the danger and trouble incidental on the employment of light oil, especially abroad. Moreover, the ease with which a steam motor can be started and stopped, and more particularly reversed, cannot be over-estimated. Fuels, other than coal, coke, or oil, have been the matter of careful consideration by motor-car designers. The most promising of these is acetylene, which, as derived from calcium carbide, enables a much greater quantity of energy to be obtained from a given weight of fuel; but although it only requires one-fourth of the weight of calcium carbide to produce a given amount of work as compared with coke, the expense at present makes its use commercially impossible. Electrical motors are clean, extremely convenient and simple, free from all vibration and danger, and altogether an ideal type of motor. The limitations in the use of electricity are however very serious, and will be discussed later on. Internal-Combustion Motors. There are six features of internal-combustion motors upon which their success more or less depends, and they are:— 1. Carburettion or Carburization; 2. Ignition; 3. Starting; 4. Governing; 5. Balancing; 6. Cooling. It is impossible even to enumerate the variety of inventions on these various parts of the oil-

304

APRIL 1900.

ROAD LOCOMOTION.

205

engine, especially adapted for the motor car, and these will be found to have been recorded from week to week in the journals specially devoted to motor vehicles. Some salient points however may be stated. 1. Carburettors, or equivalent.—Carburettors are of course the arrangement by which air is charged with a suitable portion of spirit vapour before entering the cylinder; and they are of two chief kinds:—(1) Vaporising carburettors, by which the spirit is heated so as to insure its giving off a certain amount of vapour, when a current of air is drawn through or over it. (2) Spray carburettors, such as a float feed, by which a fine jet of spirit is drawn into a current of air at ordinary temperatures. Of course there is some difference of terminology on this subject; thus the Daimler Company state that the vaporiser or carburettor has been entirely discarded, and a new system of automatic float-feed adopted, whereas types of float feed which are similar in general principle, if not in detail, are distinctly termed " Carburettors " by respective makers. The great object of many of these contrivances is automatically to adapt the charge of air and petroleum spirit to varying conditions of load, so as to attain the greatest economy with uniformity of speed. It is obviously a matter of great importance to separate completely or atomize the hydro-carbons and mix them with the air, and there are makers who go so far as to claim an economy of 25 per cent, by the special mixing arrangements they use. The Longuemare, Huzelstein, De Dion, Daimler, and Lucas are among the best known of these mixers. 2. Ignition.—There are two methods of ignition, which are of course common to all gas-engines, viz., tube and electric; but we have to consider these specially in their application to motor vehicles. For a very long time tube-ignition maintained its popularity, owing to its great simplicity and comparative trustworthiness. In strong winds however, as the author can testify from experience, it was a frequent occurrence for the flames to be blown out in running, and this fact particularly led to the

305

206

ROAD

LOCOMOTION.

APRIL 1900.

invention of the Lyon and Whitmore system of ignition, which was awarded a diploma at the Richmond Show, in which, by means of spirit, a small bunch of platinum was kept white hot (inside a tube) which projected into the cylinder of the motor. Now however there seems to be a steady and certain tendency in the direction of electricignition. As long as the electric-ignition was limited to either primary or secondary cells, which require charging and attention, this was not the case, because, in spite of the claims made as to the number of miles the cell would run, there was no means of being sure as to the amount of the charge until it had ceased to work. Now however the magneto-electric ignition, in which the current is obtained by the revolution of the motor itself, or by turning a handle previous to starting, has removed these objections. The Simms-Bosch is one of the best known of these, and is being fitted by the Motor Carriage Supply Company to all their engines at the present time. There is an advantage however in electric ignition above all others, and that is the fact that it can be practically used as a governor, for, by advancing the time of ignition merely by turning the point of contact on the dial, the combustion of the charge can be made to take place at any point of the stroke, and hence a most effective means is found of regulating the speed. 3. Starting.—All starting arrangements of the oil-engine require vigorous turning of a handle in order to get an initial compression, and therefore instead of, as in the case of steam, stopping the motor with the temporary stopping of the vehicle, it is found more convenient to throw the engine out of gear. This is at present one of the most objectionable features of the oil-engine used for the motorcar, since on such occasions, owing to there being no load, it is very difficult indeed to insure perfect combustion, the impulse taking place only once in several revolutions, giving a consequent vibration to the engine, which when unloaded is very marked. It will be easily understood that when stopping for a few minutes in the street, or outside a house or shop, and particularly when only one of the passengers requires to dismount, this is an annoyance which probably tells more than everything else against the popularity of

306

APRIL 1900.

ROAD LOCOMOTION.

207

the motor vehicle. Hence it is that various attempts have been made under the same general lines as the ordinary gas-engine, to obtain self-starters under a charge of compressed-air. Opinions of course differ as to the importance of such self-starters, and it yet remains to be seen whether their general introduction will take place, adding as it does another complication to the vehicle. 4. Governors.—The use of a governor of some kind for the oilengine, which is frequently thrown out of gear with the driving wheels, becomes an important feature of the engine. The governor of the Daimler motor operates on the exhaust valve. When the engine overruns the valve-rods they are prevented from being operated on by cams; the valves are thus kept shut, which prevents a fresh charge being drawn in through the inlet valve. The peculiarity of the engine is that the valve-rods for both engines, which are always in pairs, are not each operated on at the same time, one valve-rod being first operated upon and then the other, which enables a much more regular action to be obtained. Usually, where a governor in a hydro-carbon motor is employed, the speed of the engine is regulated by either throttling the inlet valve or not lifting up the exhaust-valve. The Simms-Bosch ignition-gear already alluded to.is usually actuated by a lever, thus enabling the speed of the motor to be adjusted by hand. In some cases however this invention is combined with a centrifugal governor; thus the governor can be set to any required speed, and the timing gear automatically actuated not by the more or less crude way of throttling, but by actually decreasing the number of explosions by the retardation of the time of ignition. This seems a very scientific method of governing, and is rapidly coming into favour. 5. Balancing.—The satisfactory balancing of the gas-engine is a much more difficult problem than that of the steam-engine. In the first place, owing to the explosions in the cylinder of the former, the piston receives a violent impulse even under the best conditions of mixture and compression, which is quite different from the behaviour of steam; secondly, steam can be regulated by admitting only a

307

208

ROAD

LOCOMOTION.

APRIL 1900.

small quantity at a reduced pressure every stroke, whereas the Otto cycle, which is practically universal, only allows the impulse to take place once in two revolutions. This necessarily introduces an irregularity in the motion, the effect of which cannot be appreciably modified on a motor-car by a fly-wheel, and accentuates any want of balance in the working parts. When the engine is in gear and the vehicle travelling at a fair speed, these irregularities are absorbed by the mass of the vehicle, but at slow speeds or when the engine is disconnected want of balance makes itself felt. The original Daimler placed the angles of the two engines nearly at 180°, and thereby mechanically effected the balancing of the reciprocating parts, although of course there was a turning couple, which was however not of great importance, owing to the two cylinders being placed quite close together. But strange to say, as progress has been made, this mechanical balance has been entirely departed from, and the two cranks of the similar engines are placed side by side, on account of the fact above-mentioned of the Otto cycle being employed, and there is found to be less vibration and generally better results, since the cranks would in the original plan be separated by angles of 180° and 540°, whereas if placed side by side they are only really separated by angles of 360° when measured by the period at which the successive impulses occur. Plate 33 shows the typical ways in which balancing is effected. Fig. 32 represents the original Daimler in which the cranks are separated, and in which the pistons are always moving in opposite directions. Fig. 26 represents the Pretot and Koch, with a single cylinder and having an Otto cycle, in which two pistons are employed moving in opposite directions (as also do the levers) connecting rods which are attached to the crank. This is in some respects a very effective method of balancing, but the increase in the number of rods is of course a serious drawback to its general introduction. In this of course, as the cranks must be separate, the twisting action on the couples is unavoidable. Fig. 27, which has been termed the " Gobron" method, has been very frequently adopted by other inventors. This effects the same end as in the previous case without requiring a number of additional levers, and is probably one of the only single-cylinder methods of balancing

308

APRIL 1900.

ROAD LOCOMOTION.

209

which entirely avoids the action of a twisting couple, the only other one known to the author is that of Hyler-White, shown in Fig. 28, which however, as will be seen from the drawing, obliges the use of spur gearing. The Henriod system, Fig. 30, really corresponds to the Daimler method ; although it is not so compact, it has a theoretical advantage which becomes obvious when the crank effort diagrams are drawn out, from the connecting rods operating at different sides of the crank shaft. Fig. 29 is that of Lanchester No. 1. In order to understand the action of this it must be remembered that the two cranks are disconnected, the two fly-wheels moving in opposite directions; of course, with only a single-cylinder engine, the piston is not balanced, but as there are two connecting rods which move outwards and inwards in opposite directions, these balance each other perfectly, and by the addition of another cycle a perfect balance of the system can be obtained. Fig. 31 gives another form of Lanchester engine. Fig. 33 gives the "Monarch," and method of balancing. 6. Cooling.—For small oil-engines, such as are used on the motor tricycle, the cylinder can be kept sufficiently cool by radiating plates cast on the cylinder, as in the well-known De Dion motor. When the engine is increased to three-horse power and upwards, water-jackets become an essential feature of the engine. At first there were no special arrangements made to cool this water, and with severe running in warm weather a considerable quantity was evaporated in this way. In recent years more attention has been given to this subject, and by means of forcing air round circulating tanks, and the use of special coolers or radiators consisting of pipes through which the water is circulated, the pipes having a special contrivance either in the shape of very fine wire attachments so as to dissipate the heat, or of small rosettes of thin metal threaded on the pipe, a great improvement has been effected both in efficiency and in reducing the consumption of water. In the case of one of the trials at Richmond in 1899, a wagonette, carrying three passengers besides the driver, ran 50 miles in 43/4hours with a consumption of only 21/2ounces of water. The Daimler Company are now fitting water coolers upon all their vehicles.

309

210

ROAD LOCOMOTION.

APRIL 1900.

External Combwtion (Steam). The external-combustion motors for vehicles are up to the present entirely steam-engines, and so much has been done in recent years in developing and improving steam-engines upon a small scale, that it is scarcely possible to have any absolutely new features in the engines themselves. Thus, although there are various special details which are of interest, it is not worth while to take up time in describing the varieties of the engines themselves that are employed. On the other hand, oil-burners and boilers, as well as condensers, have in many cases been invented entirely with the object of their application in this direction, and must be, however briefly, touched upon. As before mentioned, the use of steam-engines in this country has been to a great extent limited to heavy motor vehicles, and it is a point of great interest to compare how the more important makers have arranged the distribution of the boiler, motor, gearing, tanks, condensers, &c. The principal systems of steam-motor vehicles have been arranged for convenience of comparison upon one diagram, Plate 34, and further for convenience a uniform system of lettering (explained in the Plate) has been adopted throughout. A glance at the diagrams will show the very varied methods in which makers have distributed the essential features of the heavy motor vehicle. It will be noticed at once that all the boilers except one are placed in the front and above the car. The exception is the Musker system, Fig. 36, and this, owing to its having a horizontal boiler and a special fan or draught for the burner, by which it is enabled to do without a funnel, is placed transversely under the middle of the car. The Musker system really differs from the others in the essential feature of employing a separate auxiliary engine, which supplies air and oil for the burner in proper proportions, and also water to the boiler. The fan, as will be seen from the diagram, takes the air through the condenser in whieh it is partially warmed. It is obviously important to have the boiler and engine as near as possible to the main driving wheels, which are in every case the rear wheels, because, although when

310

APRIL 1900.

RO

AD LOCOMOTION.

loaded the weight of the load is in most vehicles to a great extent distributed over the driving wheels, yet when running light, if the boiler and a fair proportion of the engine are carried upon the steering wheels, there may not be sufficient weight upon the driving wheels to provide tractive effort. A large platform area is provided by the Musker system, but at the same time it must be pointed out that when one of these vehicles is carrying its full load, the weight is not so much concentrated over the driving wheels as in the other systems, which is a point decidedly in their favour. This system has recently been fully described in a Paper before the Liverpool Engineering Society. The next important feature of difference between the systems is in the position of the engines. In the Thornycroft & Lifu systems, Fig. 34, the engines are placed horizontally in the middle of the wagon, and the main driving wheel is driven by means of toothed gearing. This is also the case in the Musker system. The Coulthard, Leyland, Fig. 37, and Glarkson and Capel systems, Fig. 39, all have vertical engines which, by means of chain gearing operating through a counter-shaft, transmit the motion to the main driving wheel. In the Bayley system, Fig. 38, the engine is also vertical, but transmits the motion by means of a horizontal shaft placed longitudinally with the wagon, and drives a counter-shaft by means of bevelled gearing, which counter-shaft in turn drives tbe main driving wheel by a pinion and spur-wheel. In the SimpsonBodman system, Fig. 35, the distribution of parts for some reasons is the best of all, and the whole arrangement is extremely neat and ingenious. In this case there are a pair of small three-cylinder engines which work separately and independently the two main driving wheels. These engines are placed at the rear of the vehicle in a convenient and accessible position, and their weight, together with the weight of the gearing, tends to increase the tractive effort of the main driving wheels when the wagon is running light. By using separate engines, the necessity for a jack-in-the-box or differential gearing is avoided. Interchangeuble spur-wheels are used to transmit the motion from the engine to a counter-shaft, by means of which change of speed gear can be effected in a few

311

212

ROAD

LOCOMOTION.

APRIL 1900.

minutes, and from them the power is transmitted to the main driving wheels by a powerful chain. The engines of Simpson-Bodman are very well balanced, and run so smoothly at several hundred revolutions a minute that a coin will stand upon its edge on the top of the cylinders. A full description of the Simpson-Bodman system has been recently given in the Automotor Journal. Another important feature of difference between the various systems is to be found in the fact that the Musker, Leyland, Coulthard, and Clarkson and Capel, all use condensers, the location of which can be seen by an inspection of the various diagrams, whereas in the others the effect of superheating the steam is relied on in order to avoid the emission of visible vapour. A full description, giving all the important details of the Thornycroft, Lifu, Coulthard, Leyland, Bayley, and Clarkson and Capel systems, will be found in the two volumes of the Keport by the Self-Propelled Traffic Association on the heavy motor trials in Liverpool, copies of which the author has presented to the library. Burners for Liquid Fuel.—The number of burners for liquid fuel which have been invented in recent years is very great, and Plate 35 represents the types of those in successful operation. An important distinction must be made between burners for light and heavy oil. For the latter, which are the burners chiefly illustrated, special means must be taken to heat in each case the petroleum in order to vaporise it, whereas with a light spirit, vaporization, although necessarily effected on the same principle, is a far less difficult matter, and may be said not to involve the risk of carbonising the products. It is well worth the application of considerable effort and ingenuity in devising means for the use of heavy oil, as it is relatively much cheaper and safer to use. For internal-combustion engines, with one or two exceptions, light oils are invariably used on account of their cleanliness and ease of vaporization, while, on the other hand, for external combustion, in which petroleum is employed to heat the boiler, there are only one or two examples in which light spirit is employed. The Longuemare burner, which is shown in plan and elevation in Fig. 40, Plate 35, is largely used

312

APRIL 1900.

ROAD

LOCOMOTION.

213

In France, and consists of a row of coils through which the spirit is brought, from which it afterwards passes down by a pipe B, through a needle-valve which is regulated by a wheel C, that can be operated by the driver. The Lifu burner, Fig. 42, of the Liquid Fuel Company, late of East Cowes, has worked very successfully, and consists of a casting D, in the tortuous passages of which the petroleum is made to circulate ; it thus becomes thoroughly vaporised, since the casting is placed in the body of the flame which issues at E. F is an air cone, allowing the proper proportion of air to mix with the vapour issuing from the needle-valve, shown in section, and is self-regulating. A peculiar feature of the contrivance is an igniter G, filled with fire-brick, which is also maintained in a red-hot condition by the flame, so that in the event of the flame being extinguished suddenly, it is immediately re-lighted from the whitehot fire-clay acting as a temporary reservoir of heat. In neither of the foregoing is any attempt made to regulate the air-supply. This is an important matter in order to insure perfect combustion, and arrangements are made for doing so in both the Clarkson and Capel and Musker burners. In the former, which is shown in Fig. 44, this is in a sense done automatically. The air can be regulated in quantity by altering the amount of opening of the diaphragm at L L. It mixes there thoroughly with the vapour which has been generated in the coil H, round which the flame circulates. This vapour enters the mixing chamber J, through a small needle-valve M, at the orifice K. The needle-valve is opened and closed by a lever P, which at the same time raises and lowers the larger valve N, so as to regulate the outflow of combined mixture of oil and air underneath at Q Q, the flame being baffled en the inside of a hollow nickel cone. The whole arrangement worked very satisfactorily in the Liverpool heavy motor trials. For the burner of Messrs. Musker, which is shown in Fig. 52, Plate 37, the air is supplied by a fan which is driven by the same auxiliary engine which supplies both the water for the boiler and oil for the burner, the right proportion being thus automatically regulated. The air passes inwards as shown, through the passage J J, which is kept at a high temperature, by means of cylindrical B

313

214

ROAD

LOCOMOTION.

APRIL 1900.

projecting ribs, which form part of the solid.'ignition-chamber K. The oil, which is admitted by drops at the point L, falling upon the heated iron surface is thus vaporised and immediately mixed with the heated air. The mixing is further insured by passing through a number of holes in a perforated block M, and ignition takes place in the chamber K. The actual working of this burner is very striking, since the air being regulated in relation to the oil, the flame, instead of as in many cases varying in quality according to the oil supply and sometimes shooting forward in a long flame with a perceptible smell, is always, under the various conditions in which the author has examined it, of the same character and intensity. The conditions of ignition are always the same, although the actual size of the flame varies with the supply from the auxiliary engine. The Leyland burner is shown in section and plan, Fig. 41, Plate 35. The burner, Fig. 43, Plate 35, is given merely as an example of one for the use of light petroleum spirit or benzine, being that employed in what is known as the Stanley Motor Car. In this case the diaphragm at S has one side exposed to the pressure of steam from the boiler by means of a steam-pipe T. The spring V is thereby caused to regulate the needle-valve at X, by which the spirit enters by the pipe Y, and passes to the furnace under the boiler; thus the supply of spirit is regulated according to the pressure of steam in the boiler. Steam Generators. The two most important considerations in connection with the design of a steam-boiler for motor vehicles are, first, to have a boiler as light as possible consistent with high pressure, and secondly, to have a type of boiler which is capable of being forced so as to meet a sudden demand for an increased quantity of steam at a higher pressure. The latter point really constitutes, as has already been pointed out, one of the great advantages of steam for a motor vehicle, and in many designs has enabled change of speed gearing to be dispensed with, since in engines working under the compound system, arrangements are made to use high-pressure

314

APRIL 1900.

ROAD LOCOMOTION.

215

steam in both cylinders in order to surmount a hill or to transport a heavy load over a piece of bad road. Hence it is that water-tube boilers of various types have been found particularly suitable, while as far as the author is aware one type of fire-tube boiler is used for motor wagons. Water-tube Boilers.—One of the most successful water-tube boilers is that of the Thornycroft Steam Wagon Company, Chiswick. This boiler is represented in section in Fig. 45, Plate 36, and really consists of two separate annular portions A and B, almost rectangular in section, connected by a number of cylindrical straight water-tubes which form the walls of a slightly tapered hollow cone C C. The furnace D is contained in the hollow of the lower annulus B, being fed through the opening in the upper annulus A, through the cover E, which can be removed for the purpose. The flame has to find its way on all sides through the narrow spaces left between the water-tubes, the products of combustion escaping by means of the funnel F. Particulars of evaporation of this and other boilers are given in Table 1 at the end of the Paper (pages 228, 229). The boiler next illustrated, as shown in half sectional elevation, Fig. 46, is in some respects similar to the Thornycroft, Fig. 45. The products of combustion, however, in this case escape entirely round the upper portion, thereby adding a feature of economy to this type of boiler, for perfectly dry steam is increased by the steam-dome H being contained in the smoke-box. The heating is effected by means of a petroleum burner J, the supply of which can be regulated by means of the screw-valve at K. Messrs. Merryweather and Co. have a boiler specially suitable for motor vehicles represented in Fig. 47. The feature of this boiler is the large size of fire-box L, which is entirely surrounded by the water space. The flames pass upwards encountering a double set of water-tubes, one being straight and slightly inclined from the horizontal, and the other being curved and vertical, and from their position insuring a very complete circulation of the water above the furnace. The De Dion boiler shown in Fig. 48, is better known in France than in this country, R 2

315

216

ROAD LOCOMOTION.

APRIL 1900.

but at the Liverpool trials on the Bayley wagon it proved itself a very efficient steam generator. It consists, like the Thornycroft, of a double annulus of rectangular section, connected by water-tubes, the essential difference between these two boilers being that one annulus M is much smaller than the other, being partly contained in it and the water-tubes connected with them, instead of being vertical as in the Thornycroft, are slightly inclined from the horizontal. The furnace N, as in the case of the Thornycroft, is fed through a cover P. It is obvious that the heating-surface is disposed to the best advantage, as the results given on pages 228-229 show. Flash Boilers.—This type of boiler in which a small quantity of water is injected at each stroke of the engine into a heated coil of metal to be flashed into steam and superheated, is by no means new. More than twenty years ago a small engine of this type, the invention of Mr. Henry Davey of Leeds, was working in the engineering laboratories of University College, London, and gave about 3/4 H.P. Professor Kennedy, who showed it to the author, at that time expressed a favourable opinion of its future possibilities, but it cannot be said that this type of boiler has really come much to the front until it was revived by Serpollet a few years ago in connection with motor vehicles. The boiler of M. Serpollet has undergone a considerable modification during the past year. It origiually consisted of a battery of thick steel tubes, jointed together by bends outside the furnace, the thick steel tubes which were originally circular being squeezed together, and finally indented so as to give a kidney-shape section, the concave side being towards the flame, and a very narrow space left for the water to pass through. M. Serpollet has now modified his boiler so that it consists of two portions, the lower being thick steel tubes twisted into a helical form, and placed so as to intercept the flame as much as possible, as shown in plan and elevation A A, Pig. 49, Plate 37. The upper portion consists of a coil B, of cylindrical tube of lighter section and not twisted as in the lower portion which is exposed more directly to the flame. The heating, which was originally effected by

316

APRIL 1900.

ROAD LOCOMOTION.

217

coke or coal furnace, is now done by means of a petroleum burner. Messrs. Simpson and Bodman have a very strong and effective flash boiler, which is shown in sectional elevation and side elevation in Fig. 50. This boiler consists of a series of heavy steel tubes C, indented after the Row condenser pattern and connected outside the furnace by a Haythorn joint E E, which is shown in section. The Row indentations alternate about 168 times in the generator, and any fluid passing round them must encounter an amount of baffling that would expose it in the most effective way to the action of the heating surface. The steam is made to pass through a drum D, which is found necessary to prevent the superheated steam having too high a temperature. The boiler is heated by a coal furnace F F, and, in about 40 minutes from lighting the fire, steam is generated. There are many boilers of course in which steam can be generated more quickly, but it must be remembered that the success of this type of boiler depends upon a reasonable mass of metal in which heat can be stored. Tangye's boiler, Fig. 51, is convenient and compact in form, and consists of a single coil of steel tube in conical form, the lower part of which is shown in plan. One of the newest forms of combined boiler and burner is that of Messrs. 0. and A. Musker of Liverpool, Fig. 52. This is placed in a horizontal position underneath the vehicle, requiring no chimney for the escape of the waste products of combustion. It consists of three cylindrical coils H H of strong steel tubes, and the flame is made to circulate in the annular space between them. The point at which the water enters is shown on the drawing, and likewise that at which the steam is supplied to the engine. In a test made by the author on one of these boilers it was found that in a period of 401/2minutes, 371/2gallons of water were evaporated, the steam-gauge remaining during this period almost absolutely steady at 300 Ibs. per square inch. To evaporate this quantity of water, 31/2gallons of commercial petroleum was used. As the weight of this boiler and contents, including the burner, was only 4 cwt. this represents roughly 15 Ibs. per H.P., and indicates the great steaming capacity of this type of boiler anl its suitability for motor vehicles.

317

218

ROAD LOCOMOTION.

APRIL 1900.

Electricity. The whole problem of the use of electricity is determined by the life and capacity of the battery, and the merits of any accumulator should really be judged to a great extent by the condition of its batteries after six months' daily use in a motor vehicle. The fact is that the sudden heavy discharges, so often taking place with a motor vehicle, expand the grids and release paste, which is washed out by the movement of the surrounding acid, and these detached particles cause internal short-circuiting by remaining in suspension between the plates. If a cell is short of " acid room," excessive heating expands the grids to such an extent that they never contract sufficiently to make good their connection with the remaining paste. Moreover splashes of acid are the cause of much more loss than is usually suspected. Again, the practice of grouping cells in parallel is open to the serious objection that if a cell on one side becomes dead or is reversed, those on the other expend energy in re-establishing equilibrium. English, French, and American tests prove that after six months' running, even under the most careful supervision, practically all secondary cells must have the positive plates re-pasted or renewed at a cost not below one-fifth of the original outlay, while in many cases as commonly used they are practically worthless at the end of this period or even sooner. So long as a range of 40 miles per charge, at speeds not exceeding 10 miles per hour, meets the requirements of an automobilist, electricity, at a cost of not more than 2d. per B.T.U., is at least on a par with steam or oil even for heavy traffic. Where these limits are exceeded, electricity is inadmissible. Distances greater than 40 miles, and speeds greater than 10 miles an hour, involve prohibitive dead-weight and excessive discharge rates. It is a pity that no electrical motor vehicles for heavy traffic have yet been brought to the practical stage, as this is to be regretted for several reasons. The ease with which each of the four wheels of a lorry can be driven is of particular importance in respect of available propulsive effort, whilst the motors and batteries can readily be disposed so as to leave the entire platform free for

318

APRIL 1900.

ROAD

LOCOMOTION.

219

merchandise. The weight of accumulators is not included in the legal tare of three tons, and the slower rates of speed demanded in this class of traffic should permit of a low ratio of battery weight to total moving weight. Experience has shown that, with a total moving weight of one ton, one-third each being batteries, vehicle and load, the effective distance is about 48 miles at a speed of 8 miles per hour, and only about 24 miles at a speed of 15 miles per hour on average roads. It must be borne in mind that the cells deteriorate sadly when used for high speeds; nevertheless, results have been attained on a motor vehicle by electricity which surpasses anything by any other kind of motive power. An account of these results has been given by the Marquis de Chasseloup-Laubat in the "North American Review " for September 1899, from which it is interesting to note that on 27th January 1899, Jenatzy, mounting a carriage not specially constructed for great speed, left the starting point (a straight, level road near Paris) and covered 11/4mile (2 kilometres) in 1 minute and 414/5seconds; the first 3-5 ths mile (1 kilometre) with a standing start in 57 seconds, that is, at the rate of 39 • 2 miles (63 • 166 kilometres) an hour, and the second kilometre with a running start in 444/5seconds, that is, at the rate of 50 miles (80.357 kilometres) an hour. On 4th March 1899, Comte de Chasseloup-Laubat, on a Jantard carriage, not specially built for this style of test, but with certain most important modifications, such as running the front and rear of the carriage into sharp points, so as to offer least resistance to air, covered the same course of 11/4mile (2 kilometres) in 1 minute 27f seconds, the first kilometre with a standing start in 48f seconds, or at the rate of 461/2miles (75 kilometres) an hour, and the second kilometre with a running start in 384/5seconds, or at the rate of 573/5miles (92.783 kilometres) an hour. Finally, on 29th April 1899, Jenatzy, riding a carriage specially built to break all records, and which has the shape of a large cigar mounted on four small wheels, made 11/4mile (2 kilometres) in 1 minute 214/5seconds, the first kilometre with a standing start in 474/5seconds, and the second kilometre with a running start in 34 seconds, or 65 3-5ths miles (105.852 kilometres) an hour. The extraordinary

319

220

ROAD

LOCOMOTION.

APRIL 1900.

nature of these results lies not so much in the fact of a high speed of 65 miles an hour by a motor vehicle, but in the fact that starting from rest the average speed for the first kilometre was 461/2miles an hour. It is safe to say that no locomotive has ever achieved this result, that is, a distance of a mile covered from rest in less than 80 seconds, and those who witnessed the trials say that the start of the electrical carriages under these conditions left the same kind of impression on the observers as the start of a rocket. There is little doubt that these results could be beaten if the consequent expense were to be faced, but, according to report, the batteries were themselves practically destroyed in the run, and at any rate the contesting vehicles were not only towed out to the scene of the trial, but towed home again afterwards, and the trial, though interesting as showing the possibilities and use of electricity, must not at all be taken as giving results which could be of use for practical purposes. Gearing or Transmission.

The ordinary modes of transmitting power, viz., by friction gearing, toothed-wheels, belts and chains, have been specially adapted and employed for transmitting the power from the motor to the driving wheels of a motor vehicle. The requirements for this particular purpose are in many respects of a special character, and the author had collected material for a special section of the Paper devoted to this subject. It has become evident, however, that it would be impossible to treat the question in a satisfactory manner within the limits of this Paper, and the subject well deserves a special Paper before the Institution. Take for instance the transmission of power by chain-gearing. This mode of transmission, from being in a very crude state a few years ago, has received so much attention that chain-gearing is as efficient, if not more so, than any other mode of transmitting power; whereas the wearing of the links of the chain, which as it occurs gives so much trouble and annoyance, has been met by the special provision of large bearing surfaces which are hardened, so that the mechanism of the chain-gear may be said now to compare favourably with the other working parts of the engine,

320

APRIL 1900.

ROAD LOCOMOTION.

221

such as cross-head pins and main bearings. The efficient lubrication of the chain is a matter of the greatest importance and difficulty. The grease chiefly requires to be inside the joints, and Major E. E. Crompton showed the author a most ingenious way of effecting this which he has most successfully employed with bicycle chains. He places the chain in a bath of melted grease, in which graphite is mixed. The air being expelled from the joints of the chain by the heat, the mixture naturally finds its way between the pins and rollers, and forms a complete internal solid lubricant, when the whole is allowed to cool together. The Eenold chain, which is largely used in heavy motor traffic, is a beautiful invention which meets in a most ingenious manner the change of alteration of pitch due to wear, and the latest improvements in this chain are also designed to obviate as far as possible the wear upon the pin itself, and reduce it to a minimum. Again, one special feature of the oil-engine is the necessity for change of speed-gearing of some form or other, and this has led to a large number of arrangements of speed-gearing which of themselves are worthy of detailed consideration. Besides the arrangements for changing the speed-ratio by means of combination of toothed-wheels are inventions such as the expanding and contracting pulley of Mr. Lucas, the arrangement for altering the throw of an eccentric through which the power is transmitted such as in the invention of Mr. Newton, and last, but not least, the most ingenious hydraulic variable speed-gear of Mr. Hall. These inventions are all in actual operation with apparent success, but unless actually examined they could not really be understood, except by means of drawings with detailed descriptions. It is obvious therefore that this part of the subject must be left for future treatment.

IV.

RESULTS AND CONCLUSIONS.

In considering the actual results which have been obtained by motor vehicles, we must make a distinction between pleasure vehicles and those for the conveyance of goods. For the former, the actual cost of working is not by any means the first consideration; in a

321

222

ROAD LOCOMOTION.

APRIL 1900.

large number of cases in fact the cost is comparatively of small importance. Questions of comfort, durability and safety, as well as freedom from liability to break down, are the chief points to be considered. These matters can scarcely be summed up except as the result of lengthy experience, and now undoubtedly that experience is gradually being acquired. Before this Paper is read, the most important trial that has taken place in this country of pleasure vehicles will have begun, in the form of a 1,000 miles tour, undertaken under the auspices of the Automobile Club. More than eighty vehicles have entered in this competition, which is to take place over a large part of England and Scotland. These vehicles, which will run over all kinds of roads, and in all weathers, will be all under the same conditions, and the careful observations and records which will be taken of their behaviour will doubtless give the most valuable opportunity which has yet been afforded for intending purchasers to form an opinion of their relative merits. When we come to the question of goods traffic, the matter is of course entirely one of cost, including not merely the outlay, working, and upkeep, but deterioration, which in road vehicles is exceptionally heavy. Extended trials of actual working are necessary for any final opinion of the relative merits of different types of heavy motor vehicles, and the author has fortunately been able to secure much valuable testimony of this sort on the subject (see Appendices). A great deal, however, can be ascertained by careful trials such as those which have been undertaken on two occasions at Liverpool (1898 and 1899), since measurements and data can be obtained with a staff of observers for a limited period, which could scarcely be secured in continuous working. The results of these trials have been given in the reports upon them, which, for the purpose of this Paper, have been summarised and thrown into an entirely new form in Table 1 (pages 228-229), a study of which will enable an opinion to be formed on many points of interest. Take, for instance, the fuel consumption of the Thornycroft vehicle on the two successive trials, and it will be noticed under the heading " Fuel," the consumption fell from 3 • 64 Ibs. of Welsh coal per net ton-mile of freight to 2.38 Ibs. of Welsh coal in the

322

APRIL 1900.

223

ROAD LOCOMOTION.

following year. Again, take the improvement in the design of the same vehicle as shown by the ratio of dead mean weight to freight, which fell from 1 • 29 in 1898 to 0 • 96 in 1899. In the same way the performance of the various heavy motor vehicles can be compared by studying this Table, since the results were all obtained under exactly similar conditions of road and weather, and over exactly the same course of nearly 40 miles. The figures form a series of facts the truth of which cannot be disputed, and of the practical value of which engineers will doubtless be able to form their own opinion. Table 2 (page 230) is an attempt to give the various items of expense for two motor vehicles, one carrying four tons and the other eight tons, the data being arrived at from a consideration of all the best types of motor vehicles at present in operation. This Table cannot of course pretend to absolute accuracy, and many of the items will vary considerably in different localities; but anyone can alter for himself any of the details, such as the wages or repairs, and make the necessary corrections in order to form some conclusions of a practical nature. These results have been used in arriving at the motor curve, Fig. 53 (page 224). Estimated Cost of Transport by a Motor Wagon carrying Four Tons of Freight. Miles

20

25

30

35

40

45

50

3.90

3.21

2.76

2.43

2.18

1.99

1.84

In the data for obtaining the results for the above Table £75 per annum has been taken as the minimum amount for repairs, with an increment of £7.5 for each 5 miles above 20 per diem. It has been assumed that the boiler is under fire for 70 hours per week, at a minimum fuel consumption averaging 30 Ibs. of gas-coke per hour, with an increment of 2 • 5 Ibs. per hour for each 5 miles above 20 per diem. Saturdays are regarded as half-days, and only 50 working weeks are reckoned per annum. If a single-trailer carrying another 4 tons of freight be added, the costs per net ton-mile are practically

323

224

ROAD LOCOMOTION.

APRIL 1900.

reduced by one-half; if only half the net ton-mileage is obtained, that is, equivalent to 2 tons of freight throughout, the costs per net ton-mile are practically doubled. The next way of estimating the cost is naturally to try and arrive at some comparison of the cost of the motor wagon with that of the railway. Fig. 53. Relation between cost of transit by Railway and by Motor Vehicle. Miles

The following are the general figures which have been abstracted from the Blue Book of 1892,* and from this the railway curve, Fig. 53, has been plotted, and this must be considered in connection with the motor curve on the same figure. Now it must be remembered that, in considering this comparison, the costs per net ton-mile are assumed to take the place of * Analysis of the Railway Eates and Charges Order Confirmation Acts, 1891 and 1892.

324

APRIL 1900.

225

ROAD LOCOMOTION.

Comparison between Railways and Motors. Miles

10

20

30

40

50

11.54

7.18

5.62

4.84

4.37

5.90

3.90

2.76

2.18

1.84

Kail way (Pence per netf tonmile) Motor (Pence per net ton-mile)

)

railway charges added to cartages, and not to include the receiving and delivering at distances common to both. This point must be borne in mind, as a failure to appreciate it has led to more than one occasion of misunderstanding of the advantages of the motor vehicle, which it undoubtedly has over short distances. In one direction there certainly does appear a field for profitable construction, namely, to make the motive portion detachable from the main body of the lorry or wagon, and capable of transporting itself for purposes of moving a similar wagon or lorry, which is waiting elsewhere. This would meet one of the greatest objections to motor vehicles which occurs in the detention of the expensive part of the vehicle (namely, the motor and gearing) for long periods, during which the lorry is waiting to either receive or discharge its load. This is analogous to the cause which has led certain shipowners to abandon steambarges in favour of the use of steam-tugs. Something in this direction is apparently urgently needed in connection with motor traffic. Whatever value may be attached to "the statistics of trials, the independent experience of those actually using motor wagons must have even greater weight. The author has been fortunate in securing from Mr. A. G. Lyster, the docks engineer of Liverpool, and Mr. John A. Brodie, the city engineer of Liverpool, and Mr. H. A. Hoy, the chief mechanical engineer of the Lancashire and Yorkshire Railway, statements concerning the use of the motor vehicles which they have now for some time had in operation. These will be found in the form of Appendices, as it is far better that the statements of gentlemen so

325

226

ROAD LOCOMOTION.

APRIL 1900.

well able to express a sound and at the same time impartial opinion should be given in their own words. The two former have Ley land wagons, and the latter a Thornycroft wagon, and it will suffice to remark here that their testimony is unanimous in favour of these vehicles as compared with horse-traction. The value of such testimony lies in the fact that these motor wagons are in each case not really employed to do a new class of work, but to take the place of a certain number of ordinary wagons. There is thus every reason to believe that the heavy motor vehicle, at any rate, has before it a very important future, and it is gratifying to think that English engineers are well to the front in the design and manufacture of such vehicles. For instance, the details (taken from the report of the French trials) of one of the best of the French heavy motor vehicles is given in Table 1 (page 228), namely the De Dion, and the point specially to be noted is the very great tare weight of this vehicle, which exceeds so largely any of those of English design and manufacture. This is no doubt to a great extent owing, not to any superior virtue on the part of English makers, but to the legislative restrictions of the Highways Act in this country, which compel a maximum tare weight of 3 tons. The fact remains however that this limit is too low in view of the loads which are now expected to be transported on wagons, 8 tons (and frequently 10 tons) being a common load on Liverpool lorries. [Representations have been already made to the Local Government Board on this subject, which it is hoped will bear some fruit in at any rate a moderate relaxation of this tare limit, since it is found extremely difficult to obtain the necessary strength and durability with the lightness of structure required by the 3-ton limit, even with the aid of such an expensive metal as aluminium. Whilst the special improvements and inventions of the various makers engaged are properly and legitimately protected by patents, all the essential features of the motor vehicle are absolutely free from restrictions as to manufacture and sale, nor has there been any attempt at company promoting on a large scale with this class of vehicles. In this the history of the recent development of the subject bears a remarkable contrast to that of lighter motor vehicles

326

APRIL 1900.

ROAD LOCOMOTION.

227

for pleasure purposes. The position of affairs in regard to the latter has been brought forward in a remarkable Paper * by the Hon. John Scott Montagu, M.P., on " The General Aspects of British Automobile Manufacture." In this Paper, a thorough exposure of the iniquitous system by which the general public is defrauded by company promoters has been made in a most courageous and emphatic manner. Contrary to general belief the engineering profession suffers much by such practices. When there is a prospect of the healthy development of any engineering enterprise which has been gradually and carefully led up to by the successive efforts of many inventors, who have probably obtained nothing but loss and disappointment from their labours, an invention, more or less successful, which can be boomed, is made use of by the company promoter in such a way as to cause a promising industry to be looked upon for many years with suspicion and distrust by the general public. It is however satisfactory to know that the light motor car, including the oil engine, is as free to all engineers to develop as the corresponding branch of the heavy motor industry. Looking at the whole question, it may be safely said that the motor vehicle has come to stay, and that its uses both in peace and war will rapidly and enormously develop. The public interest, which is now seen partly by the enormous number of patents taken out in connection with the industry, partly by the great growth of literature on the subject, and by the formation of automotor clubs, is not a mere transient thing, and although the motor vehicle is at present still somewhat of a rara avis upon our roads, it may not be going too far to think that the coming century will see a development of locomotion upon roads comparable with the development of locomotion of the railway in the century which according to our individual views of chronology is either past or so very nearly past. In concluding this Paper the author desires to acknowledge the services of Mr. Bibby, B.Sc., who has drawn nearly all the Figures, which are in many cases entirely new. * Automotor Journal, February 1900.;

327

TABLE 1.—Liverpool Trials of Motor Vehicles for heavy traffic—(continued

on next page).

Results of the years 1898 and 1899 compared. Proportion Boiler. Ratio of mean of mean total moving Dead Weight Weight Average Heating to Freight. on Driving- of observed Surface. Wheels. Pressures.

Vehicle.

Tare.

Freight.

Mean total moving Weight.

1898 1899

Thornycroft (4- wheeler) Do. do.

Tons. 2.83 3.00

Tons. 2.53 3.73

Tons. 5.91 7.24

1.29 0.96

1898 1899

Thornycroft (6-wheeler) Do. (8-wheeler)

3.85 3.90

4.73 6.65

9.25 11.28

0.93 0.71

1898 1899

Leyland Do

2.86 2.85

4.06 4.44

7.29 7.64

1898 1899

Lifu Clarkson

2.39 3.00

2.20 3.35

1898 1899

DeDion (French Trials) Bayley

4.72 2.97

3.25 3.67

Year.

0.77 0.68

Lbs. per sq. inch. 152 147

Square feet. 65 83

0.46 0.54

125 176

65 83

0.78 0.71

0.62

178 167

110 110

4.94 6.68

1.21 1.00

0.77 0.51

207 193

80 80

8.60 7.13

1.65 0.95

0.60 0.67

? 174

56 70

TABLE 1—(concluded from opposite page). Ratios of gearing between Engine-Shaft per and Compound Revs. min. at DrivingCylinders. full speed. Wheels. Engine.

Year.

1898

Vehicle.

Thorn ycrof t (4-wheeler)

1899 1898

Do.

do.

Thornycroft (6- wheeler)

1899

Do.

1898

Leyland

1899

Do

(8-wheeler)

Inches. 4 and 7 X5 do. 4 and 7 X 5 do. 3 and 5 X 6 2f and 5 X 6

1898

Lifu

1899

Clarkson

1898

De Dion (French Trials) 4 and 7f X6 Bavlev 4 and7 X 5

1899

. . . .

3 and 6 X 5 2f and 6 X 4

Consumption per net ton-mile of freight. Fuel.

Water.

Speed. Miles per hour. Actual running. Commercial.

Prime cost.

500

8 to 1

Lbs. Welsh Coal. 3.64

Gals. *l.46

5.98

5.22

£ 630

770

10-land 17-7

2.38

1.88

5.94

5.31

590

500

9 and 12

4.03

*1.93

3.41

2.79

750

770

10.landl7-7

1.87

1.33

6.48

5.67

640

500

Gals. Kerosene. 0.130

*0.84

5.25

4.45

375

400

8, 13.5 and 28 8.5, 15.25 and 35

0.121

*0.91

6.17

5.02

450

600

8 to1

0.298

2.06

8.29

7.02

525

600

12 and 36

0.216

5.59

4.94

450

600

7.4 and 12.5

Lbs. Gas Coke. 1.81

1.68

6.33

5.47

760

500

8.4 and 13.7

1.84

1.27

5.54

4.93

600

* Surface Condenser.

0 .64

Wired Tube-Condenser and Fan.

230

ROAD LOCOMOTION.

APRIL 1900.

TABLE 2.—Estimated Annual Expenditure.

Weight of Freight carried.

Class of Work.

4 tons on self-contained wagon.

8 tons on motor-wagon and trailer.

PBIME COST

£500

£600

Interest at 5 per cent, per annum .

25.0

30.0

Depreciation at 15 per cent. „

75.0

90.0

50.0

75.0

3.0

4.0

91.0

110.5

105.0

125.0

Oil, grease, and waste

10.0

12.0

Insurances

10.0

12.0

5.0

7.0

£374.0

£4655

70 hours under steam, Per week 22Q miles fully laden 50 weeks per annum.

„

Fuel—Gas Coke at 16s. per ton .

Wages Kepairs (labour and materials) .

Rent, rates, and taxes Total per annum, exclusive of establishment charges Vehicle-miles p e r annum

. . . .

Cost per vehicle-mile . . . . Net ton-miles per annum Cost per net ton-mile.

.

.

pence

8.16

11,000

10.15

. . . .

44,000

88,000

. pence

2.04

1.27

d. 2.59

d. 1.68

Cost, using liquid fuel — kerosene at 4d. per gallon, being three times as costly as coke at 16s. per ton.

330

11,000

Fig. 5. Deep depression shows crossing.

Fig. 6.

Fig. 7.

Fig. 8.

Fig. 9.

Fig. 10. Dunlop (English). Fig. 11. Michelin (French).

Fig. 12. Callus (French).

Fig. 13. Goodyear (American)

Fig. 14.

Fig. 15.

Fig. 16.

Fig. 17. Ley land.

Fig. 20, Coulthard.

Fig. 18.

Fig. 21.

Thorny croft.

Simpson-Bodman.

Fig. 19

Simpson-Bodman.

Fig. 22. Bay ley.

Fig. '23.

Fig. 24.

Daimler and others.

Clark son & Capel and others.

Fig. 25. Davis.

Fig. 26. Pretot and others.

Fig. 28.

Fig. 27. Gobron and others.

Hyler-White.

Fig. 29.

Fig. 30.

Henriod.

Fig. 32. Daimler.

Lanchester,

Fig. 31. Lanchester, No. 2.

Fig 33. Monarch and others.

335

Fig. 34. Thorny croft (and Lifn).

Fig- 37Coulthard (and Ley land).

Fig. 35Simpson -Bodnian.

Fig. 38. Bay ley.

Fig. 36.

Musker.

Fig. 39Clarkson and Capel.

Fig. 40.

Fig. 42.

Longuemare.

Fig. 41.

Leyland.

Lifu.

Fig. 43. Locomobile or Stanley.

Fig. 44. Clarkson and Capel.

Fig. 46. Fig. 45.

Thorny croft.

Fig. 47. Merry weather and Clark son & Capel.

Fig. 48. De Dion.

338

Fig. 49.

Serpollet.

Fig. 50.

Simpson-Bodman.

Fig. 52.

Musker.

Fig. 51.

Tangye.

Please note that Plate 38 has been intentionally omitted, as have the Appendices.

340

Rudolf Christian Karl Diesel (1858-1913) Diesel was born and spent his childhood in Paris, where he was fortunate to live close to the Conservatoire des Arts et Metiers. Here he saw engines by Lenoir and Cugnot. By 1872, Diesel had decided upon a career in engineering, despite family opposition. He studied at Augsburg and Munich, and on graduation he sought experience in the Sulzer factory in Winterhur, Switzerland. Sulzer's were then manufacturing Linde ice-making machines, which used ammonia gas as a refrigerant. The young Diesel proposed a method of producing potable table ice, because the Linde machine made industrial ice only. This idea bore fruit as a practical, industrial process in Paris. More importantly, he began to think of using superheated ammonia, as an alternative to steam, for engine power. This was the first step towards the engine which was to bear his name. Diesel moved to Berlin in 1890. Here, he reconsidered his proposed engine and in a key paper written in 1892 Theorie und Konstruction eines mtionellen Wtirmemotors... noted the possibility of using highly compressed air to ignite fuel. A series of patents developed the idea through to 1893, when experiments were embarked upon to realise a practical and thermally efficient internal combustion engine. There followed an epic of building and testing at the Maschinenfabrik Augsburg, managed by Heinrich von Buz. From 1893 to 1897 evolving Diesel test engines were run with increasing success, from outright explosion to full acceptance trials. Widespread licensing agreements brought both fame and success. The first non-German, manufacturers were Mirrlees, Watson & Yaryan of Glasgow, who acted on Lord Kelvin's personal endorsement of Diesel's designs. New applications of the engine came quickly during the inventor's lifetime, in ships, railway trains and lorries. However, the 20th century saw Rudolf Diesel dogged by health problems and by running arguments over the priority and importance of his contributions to engine development. Charles Parsons noted too that Diesel was "concerned about the breath of war...blowing across Europe". His loss at sea, from the Dresden ferry en route to Harwich in 1913, may have been suicide. A tragic end to a great engineer.

341

MARCH 1912.

179

THE DIESEL OIL-ENGINE,

AND ITS INDUSTRIAL IMPORTANCE, PARTICULARLY FOR GREAT BRITAIN. BY Dr. RUDOLPH DIESEL, OP MUNICH.

[Translated from the German.] There have been so many publications recently, and especially during the past year, in technical periodicals of all languages, on the construction of the Diesel engine and its various types, that it is hardly possible to give any fresh information on the subject. Moreover, an excellent Paper was read only last summer at the Zurich Meeting of the Institution by Mr. J. F. Schubeler. The author proposes, therefore, to discuss only questions of general importance concerning the Diesel engine, especially those questions which are brought into prominence in the title of the Paper, and to consider them only as the starting point of a thorough and stimulating discussion. He desires, further, to be exonerated if statements are made in the Paper on points which are common knowledge. Since its first appearance about fourteen years ago the Diesel engine has been built by the thousand in the best factories of all industrial countries, and has been set up in the most remote corners of the world. It has been proved to be a most reliable engine when properly built, the working of which is quite as safe as that of any p 2

343

180

DIESEL OIL-ENGINE.

MARCH 1912.

other system of prime mover; and in general it is even more simple, since it does not require any auxiliary apparatus, and since the fuel in its natural and original form, without having previously undergone any transforming process, is directly converted into work in the cylinder of the engine. As early as 1897, when, after four years of difficult experiments, the author had put the first engine into working order in the factory of the Augsburg Works, numerous engineering representatives and experts who came from various countries to examine this engine expressed the opinion that it gave better heat utilization than any known kind of heat engine. From experience gained subsequently by working many engines, by gradual improvements in the construction and manufacture, and by increasing the size, the results have been still further improved, and to-day the thermal or indicated efficiency reaches 48 per cent.* in this engine, and the effective or brake efficiency reaches, in some cases, 35 per cent, of the heat value of the fuel. Fig. 1 shows the heat utilization for 1 b.h.p.-hour in the different kinds of prime movers known to-day. Science and technical knowledge are making continuous progress, and the time will come when even these figures will be exceeded; but, with our present scientific knowledge, any considerably higher efficiency in the process of transforming heat into mechanical work is not obtainable. Further progress seems only possible with some other method for transforming heat into work; this would mean an entirely new principle, which, however, one cannot speculate upon in the present state of science. The Diesel engine is therefore the engine which converts the heat of the natural fuel into work in the cylinder itself, without any previous transforming process, and which utilizes it as far as the present standard of science permits; it is therefore the simplest and, at the same time, the most economical prime mover. These two facts explain its success; it lies in the new principle of the internal working process, and not in constructional improvements or alterations of older types of engines. There is * See Mr. Dugald Clerk's remarks (page 236).

344

MARCH 1912.

DIESEL OIL-ENGINE.

181

no doubt that the careful working out of all the constructional details also plays a great part in the practical success of the Diesel FIG. 1.—Heat Consumption of different Heat Engines per B.H.P.-hour.

engine, as in any other; but they are not the essential points, and above all they do not constitute the great importance of this engine to the world's industry.

345

182

DIESEL OIL-ENGINE.

MAROH 1912.

A further reason for this importance is that the Diesel engine has broken the monopoly of coal, and has solved the problem of using liquid fuel for power production in its simplest and most general form. It has become for all liquid fuels what the steam-engine and gas-engine are for coal, but in a much simpler and more economical way. The truth of this statement was strikingly proved at the Turin Exhibition of last year. At this Exhibition, in the large Machinery Hall, a steam-turbine and a large Diesel engine, both made by Franco Tosi, of Milan, and set up on the same stand, were worked together with the same liquid fuel. The boilers belonging to the plant were fitted with Korting nozzles for burning crude oil. The difference between the two plants was therefore this: for the working of the steam-engine the whole boiler plant with its chimney, fuel supply apparatus, purification plant for feed-water with feed-pumps, extensive steam-pipes, condensation plant with water-pumps and an enormous water consumption, had to be provided, with the final result of consuming 21/2or more times the fuel per horse-power required by the Diesel engine standing beside it. The latter, being an entirely independent engine without any auxiliary plant, took up its crude fuel automatically and consumed it direct in its cylinders without any residue or smoke. A better proof can hardly be imagined, even for the non-technical man, that except in special cases the steam-engine cannot compete economically with the oilengine, and from this point of view the power-plant in the Machinery Hall at Turin must be looked upon as marking a historical event. It is hardly possible for a country which produces no coal, like Italy, to develop a great industry based on the steam-engine, and this is one reason for the exhibition at Turin of about thirty Diesel engines of various types and sizes, and made in different countries. Thus the Diesel engine has doubled the resources of mankind as regards power-production, and has made new and hitherto unutilized products of nature available for motor power. The Diesel engine has thereby exercised a far-reaching influence on the liquid-fuel industry, which is at the present time improving

346

FIG. 2.—Map showing Petroleum Fields of the World in the Year 1908.

184

DIESEL OIL-ENGINE.

MARCH 1912.

more rapidly than was previously conceivable. This is not the place to discuss the matter in detail, but the author wishes to mention that, owing to the interest which petroleum producers have taken in this important question, new petroleum sources are continually being developed, and new oil districts discovered. Moreover, it has been proved by recent geological researches not only that there is probably on the globe as much, or perhaps even more, liquid fuel than coal, but also that it is more conveniently distributed as regards its geographical position, Fig. 2. These facts, which are indisputable nowadays, have gradually silenced those who objected to too great a development of the Diesel engine for fear of insufficient stores of liquid fuel. Any such anxiety may be relieved by the fact that the world's production of crude oil increases at present 31/2times more quickly than the production of coal, and that the ratio of increase itself is steadily getting higher. Further, that 40 per cent, of the present production of mineral oil is already sufficient to supply the whole of the naval and mercantile fleets of the world with power, if they were worked by Diesel engines; also that with the world's present production the number of Diesel engines now working could be increased about a hundredfold. It may thus safely be asserted that, with the continual development of new oil-districts, the production of mineral oil will increase much more quickly than the demand for newly-built engines. It is therefore not surprising that, in the last yearly report of the Shell Transport Co., attention was called to the fact that the oil consumption in these engines did not nearly equal the production, and that the company has to look out for other markets to dispose of their superfluous stock. That the auxiliary industries of petroleum production are also considerably influenced is shown by the great increase which the transport industry for liquid fuel has experienced in recent times, especially the great development of tank-vessels which are, or will be, mostly driven by Diesel engines. But with all this, the influence of the Diesel engine in the world's industry is not exhausted. As early as the year 1899 the

348

MARCH 1912.

DIESEL OIL-ENGINE.

185

author utilized in his engine the by-products of coal distillation and coke plants, such as tar and creosote oils, with the same satisfactory results as with natural liquid fuels, but at that time the quality of these oils was generally too inferior for their use in the Diesel engine, and it was, moreover, subject to continual variations. The difficulties were then chiefly the following:— (1) Muddy deposits of solid hydrocarbons, especially naphthalenes, which made the working of the fuel pumps difficult, filled up the pipes and nozzles, and formed a hard crust at the nozzle mouths. These solid hydrocarbons also made higher ignition temperatures necessary. (2) Continuous change in quality and composition of the crude, uncleaned tar-oils; with each cask fresh variations appeared even in cases where the works guaranteed the use of the same coal, and the carrying out of the same distilling process, so that it was impossible to make scientific observations for drawing any definite conclusions, or making logical experimental arrangements. The characteristics of crude tar-oils were not then exactly known even to the producers ; for instance, nobody imagined that differences in the distilling temperature and variations in the nature and position of the retorts gave entirely different tar-products even when the same coal was used. It is only in recent years that the chemical industries interested in the matter have, by improved methods of fractioning and refining, combined with more careful selection of the material, succeeded in supplying fuel of a constant, and regular quality, without the drawbacks of the crude tar-oils used previously. These products—the tar and tar-oils—are thus to-day definitely brought into the sphere of activity of the Diesel engine. In Appendix I (page 208) is given a Table showing the application of various fuels for Diesel engines, while Appendix II (page 210) contains specifications for tar-oils for Diesel engines, and Appendix III (pages 212-5) gives the properties of tar. From what has been just stated, it will be seen that the Diesel engine is having an increasing influence on two other industries—the manufacture of gas and coke—the by-products of which have become so important for power production that an enormous business is at present connected with them.

349

186

DIESEL OIL-ENGINE.

MARCH 1912.

It is especially noteworthy that every town gas-works with a modern installation, and every coke-works, can be completed with an electric power generating plant by using its tars. This will have an excellent effect on many municipal and national works. It would take too much time to enter into the details of this question, but one fact stands out clearly in this connection, namely, that coal, which seemed to be most threatened by the liquid fuels, will, on the contrary, gain a new and wider ground of application through the Diesel engine. As tar and tar-oils are from 3 to 5 times better utilized in the Diesel engine than coal in the steam-engine, a much better and more economical utilization of coal is obtained if, instead of being burned under boilers on grates in a wasteful way, it is first transformed into coke and tar by distillation. Coke is used in metallurgical and other general heating purposes; from a part of the tar the valuable by-products are first extracted, and undergo further processes in the chemical industry, whilst the tar-oils and combustible by-products, and a great part of the tar itself, are burned in the Diesel engine under extraordinarily favourable conditions. The proper development of the utilization of fuel, which has already been started, and is now making rapid progress, is therefore the following: On the one hand liquid fuel in Diesel engines, and, on the other hand, gas fuel also in the form of gasified coke in the gas-engines; solid fuel as little as possible for steam-power generation, but only in the refined form of coke for all other heating and metallurgical purposes. The list of fuels applicable to the Diesel engine is not, however, exhausted with these liquid fuels mentioned previously. It is known that brown coal or lignite, the production of which is about 10 per cent, of the pit-coal production, also yields tars by distillation, and these tars when distilled, and worked up on paraffin, produce the so-called paraffin oils as by-products. Not all kinds of brown coal, however, are suitable for this process; but, in any case, these oils have been produced in such quantities that so far they have satisfied a good part of the German demand for liquid fuels for Diesel engines. To these must be added other products, such as shale and similar

350

MARCH 1912.

DIESEL OIL-ENGINE.

187

oils—produced not in great but in sufficient quantities to be of importance for power generation. Some countries—France and Scotland for instance—possess them in considerable quantities, and they are used in a good many Diesel engine plants. It is not generally known that it is also possible to burn fat vegetable oils and animal oils in the Diesel engine without any difficulty. At the Paris Exhibition in 1900 there was shown by the Otto Company a small Diesel engine which, at the request of the French Government, ran on Arachis-oil,* and worked so smoothly that only very few people were aware of it. The engine was constructed for using mineral oil, and was then worked on vegetable oil without any alterations being made. The French Government at the time thought of testing the applicability to power production of the Arachide or earth-nut, which grows in considerable quantities in their African colonies, and which can be easily cultivated there; because in this way the colonies could be supplied with power and industry from their own resources, without being compelled to buy and import coal or liquid fuels. This question has not been further developed in France owing to changes in the Ministry, but the author resumed the trials a few months ago. It has been proved that Diesel engines can be worked on earth-nut oil without any difficulty, and the author is in a position to publish, on this occasion for the first time, reliable figures obtained by tests:—Consumption of earth-nut oil, 240 grammes (0.53 Ib.) per b.h.p.-hour; calorific power of the oil, 8,600 calories per kg. (15,480 B.Th.TJ. per Ib.), thus fully equal to tar oils ; hydrogen, 11.8 per cent. This oil is almost as effective as the natural mineral oils, and as it can also be used for lubricating oil, the whole work can be carried out with a single kind of oil produced directly on the spot. Thus this engine becomes a really independent engine for the tropics. Similar successful experiments have also been made in St. Petersburg with castor-oil; and animal oils, such as train-oil, have been used with excellent results. The fact that fat oils from * Earth-nut or pea-nut—botanical name Arachis hypogaea.

351

FIG. 3.—Four-stroke Cycle.

FIG. 4.—Two-stroke Cycle.

MARCH 1912.

DIESEL OIL-ENGINE.

189

vegetable sources can be used may seem insignificant to-day, but such oils may perhaps become in course of time of the same importance as some natural mineral oils and the tar-products are now. Twelve years ago the latter were not more developed than the fat oils are to-day, and yet how important they have since become. One cannot at present predict what part these oils will play in the colonies in the future. In any case, they make it certain that motor power can still be produced from the heat of the sun, which is always available for agricultural purposes, even when all our natural stores of solid and liquid fuels are exhausted.

HISTORICAL

SUMMARY.

The author thinks that a summary of the whole development of the Diesel engine, and of the general points connected therewith, with illustrations of a few engines which mark stages in its evolution, may be of interest to the members. Several of these have already been published separately in the technical Press, but the series, as a whole, in its historical connection is quite new, and a certain number of the photographs have not been previously published. Figs. 3 and 4 show small illustrations of the principal movements in the four-stroke and two-stroke cycle engines, with the corresponding indicator-diagrams in Figs. 5 and 6 (page 190), because these will constantly be referred to in the Paper.*

FOUR-STROKE CYCLE ENGINE. Vertical Stationary Engines.—The first experimental Diesel engine, Fig. 7, Plate 4, constructed in 1893, had the piston fitted with a piston-rod and external cross-head, the cylinder having no waterjacket ; the cam-shaft was arranged very low, and the valves were actuated by means of long rods. The starting storage-chamber * These diagrams are taken from a publication by the Maschinenfabrik Augsburg und Maschinenbaugesellschaft Nurnberg A.G. (M.A.N.).

353

190

DIESEL OIL-ENGINE.

MARCH 1912.

consisted of a wrought-iron pipe with riveted flanges, and there was no air-supply pump, the fuel being injected directly. A later pattern, Fig. 8, Plate 4, built in 1895-96, had a similar base to that shown in Fig. 7, but it had a waterjacketed cylinder and the cam-shaft was placed higher up. But the most important difference from the old pattern was in the air-supply pump, the necessity for which was only recognized after several years' experimenting, as without it a smokeless combustion could not be effected. This air-pump is single acting, but the author previously used a special vertical compound air-pump driven from a transmitting shaft. FIGS. 5 and 6. Indicator Diagrams of Single-acting Diesel Engines. (Taken from original diagrams.) Four-stroke Cycle. Two-stroke Cycle.

The first French and Belgian engines were nearly of the same type as that shown in Fig. 8, but had no air-pump; they were also of better and more compact construction. The first reliable Diesel engine, Fig. 9, Plate 4, of 18 h.p. was finished in 1897 at Augsburg, after about four years' laborious experimenting. It was a vertical engine having the piston connected to an external cross-head and worked on the fourstroke cycle. The illustration shows the engine with the testingbrake attached, and with the other testing apparatus exactly in the state in which it was used by the numerous Commissions of engineers and experts who came from different countries to examine the engine, as mentioned earlier in this Paper. This type was for

354

MARCH 1912.

DIESEL OIL-ENGINE.

191

about ten years the exclusive and almost stereotyped pattern for all Diesel engines, which were built in various countries. In the following year, 1898, the first single-cylinder engine of 20 to 25 h.p. was built at the Augsburg Works. This engine had almost all the characteristic details of the experimental engine just mentioned, the only difference being that the experimental engine had the cylinder connected to the base by an inclined column in front, whilst the later pattern had the well-known A-frame which the author employed on his first experimental engines. No further alterations have been made. The engine had still the external cross-head and guides, and the petroleum-pump was actuated by the cam-shaft in exactly the same way as in the experimental engine, Fig. 9. The air-pump was cast on the base in exactly the same way, and was driven in both cases by rocking beams from the cross-head. The lubrication of the cylinders was effected by means of Mollerup appliances, and the valve-rods, the regulator, and all the details were identical in both engines. Except for alterations in unimportant details of construction, the only changes since made were that the dimensions and the numbers of cylinders were enlarged. The two-cylinder engine of 60—76 h.p. was made in 1899, in which all the details of Fig. 9, Plate 4, are still to be recognized. The only alteration which was made in the year 1901 was the abandonment of the external cross-head and adoption of the trunk piston shown in Fig. 10, Plate 5. A comparison of this engine with that illustrated in Fig. 9 shows that, with the exception of the omission of the cross-head, no alterations of any importance have been made. Vertical four-stroke cycle engines of from 10 to 250 h.p. per cylinder were constructed after this pattern, and units up to 1,000 h.p. were obtained by combining several cylinders. These engines ran at comparatively low speeds, from 160-200 revolutions, according to their size, and were of very heavy construction. Their weight was originally from 280 to 350 kg. (617-771 Ib.) per h.p., and later from 240-300 kg. (529-661 Ib.) per h.p. This type of engine was used exclusively as a stationary plant for various industrial purposes. The two-cylinder M.A.N. engine of this type of 250 h.p. or 125 h.p. per cylinder was built in 1902.

355

192

DIESEL OIL-ENGINE.

MARCH 1912.

A three-cylinder engine of the same type, Fig. 11, Plate 5, was made by Sulzer Bros, in 1906. The two latter engines show a slight alteration; the petroleum-pump is driven from the vertical instead of from the horizontal cam-shaft, as was the case in the previous engines. The engines built by Sulzer Bros, and by Carels have also a rotating stuffing-box for the fuel-needle. This arrangement was first built in Sweden on the author's instructions, and worked successfully. The well-known 500 h.p. three-cylinder engine of Carels was exhibited at Liege in 1905. The author has purposely referred to this type of engine to show that these engines, which have been built in various factories and in various countries, still remain almost an exact copy of the old experimental engine, Fig. 8. Only in America was the design simplified, or rather cheapened, from the commencement by the director of the American Diesel Engine Co., Colonel E. D. Meier. In America the engines were built without cross-heads from the beginning, an idea which, as already mentioned, was followed in the year 1901 by the European firms after the American engines with trunk-pistons had proved successful. The Americans also built from the commencement a closed base frame, and this construction, as will be seen later, has also been recently adopted in the European high-speed engines, but in a more refined and better form. Moreover, the American engines had no valves in the cylinder-covers, Fig. 12, Plate 6, but they were placed in a chamber cast at the side of the cylinder, which necessitated the fuel-needle being placed horizontally between the suction- and the exhaust-valves. Finally, the Americans, instead of driving the air-supply pump direct from the engine, always set it up independently and drove it either by a small extra engine, by a transmission shaft, or by an electric motor, in the manner in which air-pumps are now set up in many Diesel engine plants on board ship. All these alterations were made with the object of cheapening the manufacture, which is the cardinal feature in American practice. A remarkable fact is that the first Diesel engines, built in 18971898, are still working, without any change in their fuel consumption;

356

MARCH 1912.

DIESEL OIL-ENGINE.

193

also the first English engine built by the Mirrlees, "Watson and Yaryan Co. at Glasgow, according to the author's design of fourteen years ago, is still working. As the central electric stations took up the Diesel engine very early, the necessity for quicker running engines arose. This need, and the improvement in methods of construction and utilization of materials, caused the gradual introduction of the new quicker running four-stroke cycle engines, with speeds of from 300 up to 600 revolutions. These, however, were still exclusively vertical. The main difference in construction as compared with the first type was that the bearings of the crankshaft were connected with the cylinders by means of light steel columns instead of by heavy castiron A-shaped frames, so that the cast-iron pedestal of the machine became a light crank-case, relieved from great strain; in addition, the thickness of all the castings was diminished. By this means the weight of the engines was reduced to about one-fourth to one-fifth of the weight of the old types, or to about 50 kg. (110 Ib.) per h.p. Engines of this kind are now built up to about 700 h.p. and are especially suitable for driving dynamos, blowers, and centrifugal pumps, and also as auxiliary engines on board large vessels, etc. The first of these high-speed four-stroke cycle engines made by the M.A.N. had no alterations in the valves, the needle, or the gearing, nor in the driving and the position of the petroleum and air-supply pumps, etc. In a four-stroke cycle high-speed engine, Fig. 13, Plate 6, made by Messrs. Sulzer Bros, in the year 1909 the only difference, except the box-pattern frame, between it and the arrangement of the old type consisted in the position of the air-supply pump, which was in this case fitted to one end of the engine and driven direct from the crankshaft. In a later four-stroke cycle high-speed engine of 350 h.p. made by Messrs. Sulzer Bros, in 1911 the air-supply pump was also driven from the crankshaft, but was fitted between the cylinders on the box-pattern base in a neater way. In this case also no radical alteration has been made. These latter kinds of engines may be regarded as the final and permanent type of the vertical four-stroke cycle engine for Q

357

194

DIESEL OIL-ENGINE.

MARCH 1912.

stationary purposes, both for high and low speeds. With this and similar types the development of the four-stroke cycle engine reached a definite state of development. When in the last decade, through rapid development of the French submarines, an urgent need for a reliable submarine engine was felt, these four-stroke cycle engines were further reduced in weight by using steel and brass castings, with still thinner walls, and they have also recently been fitted with reversing gear. The author will return to this point later when discussing marine engines. Small Engines.—This summary of the development of the vertical four-stroke cycle engine would not be complete without a reference to the small engines which have recently been built in accordance with the author's designs. Fig. 14, Plate 7, shows a complete 5-h.p. one-cylinder plant, designed in 1909, for 600 revolutions per minute with petroleum tank, starting and air admission chambers. The officially recognized consumption for this small engine is 240 grammes (0.53 Ib.) per b.h.p., which is therefore not much more than with the old large engines of medium horse-power. At present the author is endeavouring to simplify and strengthen this small engine, which will then be suitable for small manufacturers and for farmers, who are not especially skilled in mechanical work. Fig. 15, Plate 7, shows a 10-h.p. plant which is composed of two 5-h.p. cylinders of the above kind; the air-pump is driven direct by the main shaft. The cam-shaft is arranged in the lower part of the base as in automobile engines. For this construction many hints have been taken from,automobile engine designs. Horizontal Stationary Engines.—After vertical engines had solely been used for about twelve years, horizontal four-stroke cycle engines were built. The author is uncertain whether this type was a real necessity, or whether it was originally only constructed for purposes of competition to bring out something new ; it is, however, not his intention to compare the merits of the two types. The first horizontal engines were practically only vertical engines laid on their sides without any independent structural innovations ; all the

358

MARCH 1912.

DIESEL OIL-ENGINE.

195

valves were fitted in the cylinder-cover, in exactly the same way as was done in the old vertical engine, shown in Fig. 9, Plate 4. The valves were actuated by a small cross-shaft, which was itself driven by means of gears from another shaft parallel to the cylinder axis. The air-pump was fitted in exactly the same way as in the old vertical engine. Gradually the designers freed themselves from the tradition of the vertical engine, and some details were altered in such a way that they were more suitable for the horizontal position, and a type of engine was thus obtained which is hardly distinguishable from the horizontal gas-engines, as Fig. 16, Plate 7, shows. In this engine, made by the Swiss Locomotive Works, Winterthur, the inlet-valves are no longer placed in the cover, but on the side of the cylinder as in gas-engines, and are directly driven from the longitudinal camshaft. A cross cam-shaft is no longer used. Only the fuel and exhaust valves are left in the cover, while the air-compressor is here arranged in another way. These designs are to-day very often used for smaller plants of 20 h.p. and more, especially by gas-engine manufacturers, who took up the construction of Diesel engines on their own account on the expiration of the patents, and who preferred to keep to the old types of horizontal gas-engines. But the M.A.N. built such horizontal Diesel engines for very high horse-powers as doubleacting four-stroke cycle engines with two or four cylinders arranged tandem. The largest engine of this kind so far is a double-acting four-stroke cycle tandem twin engine, of 1,600-2,000 h.p. or 400—500 h.p. per cylinder, with a speed of 150 revolutions per minute; this engine is working in the corporation gasworks at Halle, using water-gas tar as fuel. It was constructed on the lines of the well-known Niirnberger large gas-engine for blastfurnace gras. TWO-STROKE CYCLE ENGINES. As very often stated by the author, the Diesel principle is essentially suitable as a two-stroke cycle engine, because the Q 2

359

196

DIESEL OIL-ENGINE.

MARCH 1912.

scavenging is not done with a fuel-air mixture, but with pure air, so that not only untimely ignitions but also fuel losses are avoided, and the scavenging can be done more effectively, and with almost any quantity of air desired. The first two-stroke cycle engines on the Diesel principle were built in 1900 and 1901 in Germany and England, after drawings made by Giildner, but without success, because these drawings still followed too closely the two-stroke cycle gas-engines, and because the constructional arrangements were unsuited to the Diesel engine. Successful attempts to construct a two-stroke cycle Diesel engine on entirely new lines have, however, been made recently by Messrs. Sulzer Bros., of Winterthur, so that to-day this type is on a nearly equal footing with the old four-stroke cycle engine. This has been effected by working entirely on the original Diesel principle. The author says " on a nearly equal footing," because the four-stroke cycle engine still has a better combustion and a more economical fuel consumption, and is, above all, simpler in its method of working. It thus remains the standard perfect engine, and still predominates for medium-sized stationary plants up to 500 or 600 h.p. (no exact limit can be given) wherever the highest perfection and the greatest economy are desired; but engineers are now doubtful whether this supremacy will last much longer. On the other hand, the two-stroke cycle engine with its smaller cylinders has now come into favour for stationary plants of higher horse-power, and, as a marine engine, is likely to become the standard type. Two considerably different fundamental types of two-stroke cycle engines have so far been competing. To explain the principal difference, the author shows some sectional drawings. The first fundamental type is the engine made by Messrs. Sulzer Bros., Fig. 17 (page 197), with separate scavenging pump. The second—the M.A.N. engine, Fig. 18 (page 198)—was brought out much later; in it the scavenging pump, which has an annular piston, is placed underneath each combustion-cylinder. Both engines are single-acting. The author does not wish to comment upon the advantages and disadvantages of the different types, but would leave them to be discussed. Their relative merits can only be settled by experience.

360

FIG. 17.

FIG. 18. Single-acting Two-stroke Cycle Engine with Scavenging Pump underneath Combustion-Cylinder.

(M.A.N.)

MARCH 1912.

DIESEL OIL-ENGINE.

199

A three-cylinder 750 h.p. Sulzer-Diesel two-stroke cycle engine, and a still larger Sulzer-Diesel four-cylinder two-stroke cycle engine on the same system of 2,000-2,400 h.p., were illustrated in Mr. Schubeler's Paper read at the Zurich Meeting of the Institution last year. For such large engines two scavenging

FIG. 19.

pumps are necessary. This engine has therefore an even greater power than the horizontal M.A.N. engine at Halle. An addition to the two-stroke cycle engines of an entirely new type has been recently made. It was built by Prof. Junkers on the lines of the old Oechelhauser gas-engine, with two pistons working in opposite directions in one cylinder but acting on the Diesel

363

200

DIESEL OIL-ENGINE.

MARCH 1912.

principle. This two-stroke cycle process is explained in the diagram, Fig. 19 (page 199). A horizontal 1,000-h.p. engine of this kind is at present being tested at the laboratory of Prof. Junkers at Aix-la-Chapelle. The merits of this type of engine are left for discussion. MARINE ENGINES. The first marine Diesel engine of 20 h.p., Fig. 20, Plate 7, was constructed in 1902-3 in France, for use on a canal-boat, by the French engineers, Adrien Bochet and Frederic Dyckhoff, in conjunction with the author. This engine had, like the Junkers engines already mentioned, two pistons working in opposite directions in one cylinder, but the fly-wheel shaft was not at one end of the cylinder as in Junkers' engine, but traversed a cooled chamber passing straight through the combustion chamber. The engine worked on a four-stroke cycle. The great feature of this arrangement was the very high speed which was made possible by the perfect balance. This small engine worked quite satisfactorily. Others were also built in various sizes up to several hundreds of horse-power for some French submarines by Sautter, Harle and Co., Paris. This type of engine is of no further practical interest to-day; but while its first application to a canal-boat is of no importance in itself, it has at least the historical interest of being the first Diesel engine to be used on a boat. Since the date named the evolution of the Diesel marine engine has steadily continued, chiefly on the demand of the French submarines and Russian river-boats. The author has already mentioned that later on the high-speed fourstroke cycle engines, built for electric power stations, were made even lighter than before, and used for French submarines and for Russian river vessels. These engines were not originally reversible; on the contrary they were used to generate electricity by means of which the propellers were driven indirectly for manoeuvring. In the most favourable case (Delproposto) the propulsion of the vessel was performed directly by the engine,

364

MARCH 1912.

DIESEL OIL-ENGINE.

201

whilst the manoeuvring and slow driving were done by means of electricity. Thus Fig. 13, Plate 6, represents an engine which has been worked originally, not only as a stationary, but also as a marine engine, although it was not really designed for marine purposes. The first reversing marine two-stroke cycle Diesel engine was built in 1905 by Messrs. Sulzer Bros, at Winterthur; it was exhibited in 1906 at Milan and fitted to a vessel on Lake Geneva in the same year. At that time engineers were not quite clear as to the importance and value of the two-stroke cycle principle, and many firms went on trying for years to make the four-stroke cycle engine reversible. The first engine of this kind was built by Messrs. Nobel Bros, at St. Petersburg in the year 1908, and was fitted to a Russian submarine. Fig 21, Plate 8, shows this 120-h.p. three-cylinder engine. It is even visible from the outside view what great mechanical complications were at first caused by the reversing of the four-stroke cycle engine. This problem has recently been solved in a much more simple and neater way, in a six-cylinder 150-h.p. reversible four-stroke cycle engine of 350 revolutions, constructed in the year 1911 by the French firm, Messrs. .Delaunay-Belleville. This engine is fitted with two air-pumps, of which a spare one is for manoeuvring. In many factories reversible four-stroke cycle marine engines are still built, but, on the whole, engineers are, for navigation purposes, inclined to abandon the four-stroke cycle engine entirely and to replace it by the two-stroke cycle engine. The small four-cylinder engine of 30 h.p. and 600 revolutions per minute, illustrated in Fig. 22, Plate 8, is also a reversible four-stroke cycle engine. It was built for experimental purposes in the year 1909, after designs by the author, as an automobile engine for heavy loads, but it can also easily work as a marine engine. The cam-shaft is mounted on the cylinder-cover, and the illustration shows the engine with the cover lifted. The illustration above referred to is again of historical value in so far as it illustrates the first attempts to construct the Diesel engine as an automobile engine for traction wagons, and no doubt

365

202

DIESEL OIL-ENGINE.

MARCH 1912.

in future years these experiments, carried out in some different way, will lead to satisfactory results. Fig. 23, Plate 8, shows a quite new Sulzer six-cylinder marine engine working at 300 revolutions, in which an innovation may be noted. The scavenging valves are not fitted on the top within the cover but below in the scavenging air-reservoir. Quite recently very large cylinder-units, single- and double-acting vertical twostroke cycle Diesel engines, have been built. Fig. 24, Plate 8, shows one of these single-acting two-stroke cycle engines with one cylinder of 1,200 h.p., made by Messrs. Carels Freres, the results of the tests of which are as yet little known to the public. It is generally known, however, that in the Nurnberg works of the M.A.N. important experiments with large double-acting two-stroke cycle engines are being carried out. In these works prolonged official tests of a three-cylinder doubleacting two-cycle engine of 850 h.p. were made in August 1911, and the results proved highly satisfactory. At the present time a double-acting two-cycle three-cylinder engine of 2,000 h.p. per cylinder is being tested at Nurnberg. The dimensions of these cylinders are as follows :— Diameter . . . Stroke Revolutions per minute

.

.

.

.

.

8 0 0 m m . ( 3 11/2inches). 1,060 mm. (41| inches). . 160.

The air-supply pump for fuel injection is driven by a special Diesel engine, whilst the scavenging pumps are driven direct from the crankshaft. If, as seems probable, these tests also give satisfactory results, the era of very large Diesel engines has begun, especially low-speed marine engines suitable for driving propellers. This last engine has already yielded considerably more than 2,500 h.p. per cylinder, so that an engine unit of this kind with six cylinders would give 15,000 h.p. or 45,000 h.p. for a vessel with three propellers. This kind of marine engine requires six cylinders to ensure a regular turning moment and balancing, so that the number of cylinders cannot be considered abnormally high ; on the contrary, it must be accepted as the most suitable and proper number. The opinion

366

MARCH 1912.

DIESEL OIL-ENGINE.

203

which was still prevalent some months ago, that about 1,000 h.p. per cylinder is the maximum for Diesel engines, has therefore been quickly overthrown by these facts, and it may be safely assumed that the day of the large marine engine is already very near at hand. At Messrs. Krupps' Germania Works cylinder units of 2,000 h.p. double-acting two-cycle are being tested at present; also at Sulzer's Works a single-acting two-cycle cylinder of 2,000 h.p., and at Messrs. Yickers' Works a 3-cylinder single-acting, two-cycle, of 2,000 h.p. per cylinder, are being constructed. Junkers' system, illustrated in Fig. 19 (page 199), is used for vertical marine engines. The author regrets that he has not a photograph of a finished engine. From motives of prudence, the various navies which are now fitting some large warships with Diesel engines started with one Diesel only out of the two or three engines on board; the Diesel works alone when the ship is running at normal speed, but for high speed, steam is used as an auxiliary. It is evident that larger warships will not be fitted solely with Diesel engines until practical tests on the high seas have proved to be completely successful. Conclusions.—As will be seen from this historical summary, countless different types of Diesel engines have been so developed that it has become very difficult for even the expert to choose between them. If one looks through the technical Press, numerous other schemes will be found on which the author will not dwell, as they have never been actually carried out. To-day, if a new firm is about to manufacture Diesel engines, it is almost impossible to give them sound advice as to the type and size they should choose from this bewildering variety. The fads, habits and tastes of the purchasers, and the kinds of machine-tools in use in the works, have to be taken into consideration rather than technical points of the engine itself. Development is proceeding so rapidly at present that, within a few months, opinions even on important points may easily be changed. Still, the author believes that this period of chaotic production will soon be over. At present it is generally agreed that the

367

204

DIESEL OIL-ENGINE.

MARCH 1912.

four-stroke cycle engine from 5 up to 600 h.p. may be regarded as the exclusive type for stationary plants; but it will probably not remain so much longer in spite of its perfection, in view of the development of the two-stroke cycle engine, especially that of the double-acting type. The use of the two-stroke cycle engine for stationary work has increased, but it is anticipated that this will be still further extended. Although this engine may never equal the four-stroke cycle engine as regards thermal efficiency, its initial cost is so much lower that its slightly higher fuel consumption will be more than counterbalanced by the greater interest and amortization on the higher-priced fourstroke cycle engine. When this stage is reached, the question is simply one of economy. In the author's opinion the two-stroke cycle engine will thus soon make headway for stationary plants. It will therefore be necessary to produce from the various systems of two-stroke cycle engines a simple standard type, with which the more complicated types will not be able to compete. It is the author's belief that this simple standard type will make its appearance very soon, and that thus the Diesel engine movement will leave the unsettled stage and enter on a period of quiet expansion. In Appendix IY (page 216) is given a list of vessels propelled by Diesel engines, which is as complete as the author has been able to make it from private information and from publications. As many firms do not publish the details of their work, and as very little information can be obtained about anything connected with warships, such a list cannot claim to be complete. It gives only a general idea of the immense amount of work which is being done at present in this direction. In Appendix Y (page 219) some brief results of vessels propelled by Diesel engines are given, as far as it has been possible to obtain them.

Please note: Appendices IV and V have been intentionally omitted as they do not add significantly to the paper.

368

MARCH 1912.

DIESEL OIL-ENGINE.

205

SPECIAL IMPORTANCE OF THE DIESEL ENGINE FOR GREAT BRITAIN. After this short summary about the importance of the Diesel engine for the world's industry in general and the historical review of Diesel types, the author desires to add a few words on the importance of the Diesel engine, especially for Great Britain. In this connection the three following facts must be borne in mind:— (1) Great Britain is an exclusively Coal-producing Country ; (2) Great Britain has the largest Colonial Empire in the world; and (3) Great Britain is the greatest Shipping Nation in the world. Dealing with (1), Great Britain has had, at least until to-day, no natural liquid fuels of its own ; it is an exclusively coal-producing country. Based on these statements, it has often been seriously put forward in recent times that England has no interest in the Diesel engine, and that it is against her most vital interest to participate in the development of this engine, as she would neglect her wealth in coal, and make herself dependent on foreign markets when using liquid fuels. The Exhibition at Turin has shown clearly that England has still very little interest in the Diesel engine as far as stationary plants are concerned owing to her abundance of coal; for in the English Section, with the exception of some very small oil-engines, only generator-gas engines were exhibited; but these were numerous, of first-class workmanship, and were built by the best firms in the country. England has therefore made great progress in power generating with gas, whilst she has up to the present given relatively little attention to power generating with liquid fuels. But the foregoing arguments in favour of coal are not correct; the contrary is true. Great Britain has the greatest interest in replacing the coal-wasting steam-engine by the more economical Diesel engine, and this, firstly, because she can therewith effect enormous savings in her most valuable treasure—coal, and thus defer the exhaustion of her stock; and, secondly, because she can run

369

206

DIESEL OIL-ENGINE.

MARCH 1912.

her coal industry and the independent chemical industries on more economical lines, when using the coal in the more rational and refined way, as already mentioned. Finally, because she will also make herself free and independent of foreign markets for the supply of liquid fuels by using coal in this icay, that is, by working the tars and tar-oils in the Diesel engine. It is not intended to imply that the whole demand of fuel for England could be produced in this way in the country itself, but through the inland production, increase of prices for foreign oils and the establishment of trusts and monopoly companies will be prevented. In this sense independence is meant.* (2) Great Britain has the largest Colonial Empire in the world.— It is not possible to foresee to-day what England can obtain for her Colonies from the Diesel engine; even when using only the natural mineral oils, the Diesel engine is the predestined colonial engine, because only about the fourth or sixth part of weight in fuel has to be transported for it to the Colonies and their hinterlands, as compared with the steam-engine. For a colonial engine the cost of freightage for fuel is generally the determining factor. Further, the transport of these liquid fuels is considerably easier and more convenient than the transport of coal, especially when tank-vessels and pipe-lines are used; and finally the difficulty of working a boiler plant will only occasionally come into consideration in the Colonies, especially in the interior, except, of course, for small plants using wood, straw, and the like. It may be mentioned, in this connection, that a pipe-line of 400 km. (about 250 miles) in length for crude petroleum is being * In this connection it may be of some interest to state that the tar production of Germany is sufficient for more than 5 milliards of horse-powerhours per year, which means about 1 3/4 millions of horse-power running 300 days of ten hours each all the year. In case of war and cutting of! the supply of foreign fuel, this quantity would be entirely sufficient for running the whole fleet, war and mercantile, and for providing in the meantime the power for the inland industry as far as necessary. The author has not the figures for England, but he presumes they are of similar significance.

370

MARCH 1912.

DIESEL OIL-ENGINE.

207

built on the River Congo from Matadi to Leopoldville, by means of which this extensive district will be supplied, in the simplest and cheapest way, with a constant flowing fuel source, from which navigation and the railways, agriculture as well as other industries, will get their element of life, namely, motor power. This magnificent example ought to be followed in the English Colonies; it is not necessary to specify in detail the great importance of such an undertaking for them. If one also considers that the Diesel engine can utilize vegetable oils, entirely new prospects are brought to light for the cultivation and expansion of industry in the Colonies, which are for no other country of such eminent importance as for Great Britain; and this is where she ought to start as soon as possible. The Diesel engine can be worked with the Colonies' own resources, and thus again can influence to a great extent the further expansion of agriculture in districts where it is predominant. This sounds to-day somewhat like a dream, but the author ventures to prophesy, with full conviction, that this way of using the Diesel engine will one day be of the utmost importance. (3) Finally, Great Britain is the greatest Shipping Nation in the world.—When the first successes of the Diesel engine, as a marine engine, were heard of in England recently, and it was published that already numerous small mercantile and war vessels were fitted with Diesel engines, and the possibility of more important plants was mentioned, and when it was realized that larger ocean vessels were destined to be fitted with Diesel engines, and that even a warship with a very large Diesel engine was in construction, it provoked a great movement and excitement in Great Britain which is still fresh in our minds. Moreover, the reports about successful voyages with Diesel vessels under very difficult weather conditions are already increasing in number. The captains who have had Diesel engines on their vessels report on the great security and comfort in working; shipbuilders publish the figures of their savings. It is unquestionable that one of the greatest evolutions of modern industry will be connected with this

371

208

DIESEL OIL-ENGINE.

MARCH 1912.

development of the Diesel engine, and that Great Britain as the greatest shipping nation of the world will derive the greatest advantage from it. The Paper is illustrated by Plates 4 to 8 and 9 Figs, in the letterpress.

APPENDIX I. SUITABLE OILS. The Swiss Fuel-testing Laboratory at the University of Zurich, under the direction of Professor Constam, decided to undertake the examination of the qualities and composition of all liquid fuels which can be used for Diesel engines. These researches included the investigation of the following points :— (1) On the physical properties, such as (a) Properties when cold, (b) Properties on heating (boiling-analysis). (2) Chemical properties, such as (a) Chemical constituents, (b) Percentage of H2O and Ash, (c) Calorific power. This Laboratory is destined to become a centre for the investigation of fuels for Diesel engines on account of the perfection of its equipment, the accuracy of its work, and the excellence of its management. The Laboratory intends to publish from time to time exhaustive reports of its researches. From tests and examinations already made, power oils have been divided into the following three classes :— 1. Normal oils which can always be used :

Hydrogen over 10 per cent. (a) Mineral oils freed Calorific power over 10,000 cal. from benzene (gas per kg. (18,000 B.Th.U. oils) . . . perlb.). No solid impurities.

372

MARCH 1912.

DIESEL OIL-ENGINE.

209

Hydrogen over 10 per cent. Calorific power over 9,700 cal. (6) Lignite tar oils per kg. (17,460 B.Th.U. per lb.). Scarcely any researches have (c) Fat oils from vegebeen made on these. Earthtable or animal nut oil has 11-8 per cent, sources, such as hydrogen, and calorific power earthnut oil, castor 8,600 cal. per kg. (15,480 oil, fish oils, etc.* . B.Th.U. per lb.). 2. Oils which can be used only with the aid of special apparatus:— (a) Pit coal-tar oil. (b) Yertical-oven, water-gas and oil-gas tars, probably also coke-oven tars, the tests on which have not yet been completed. General characteristics: Hydrogen not over 3 per cent.; Amount of free carbon not over 3 per cent.; Residue on coking not over 3 per cent.; Calorific power not under 8,600 cal. per kg. (15,480 B.Th.U. per lb.). 3. Oils which cannot be used: Tars from horizontal or inclined retorts. It must not be understood that these will not be used in Diesel engines under special conditions; but, on the whole, the above classification is accurate in the present state of development of the Diesel engine. It is evident that for estimating the value of power oils, not only the above qualities, but all their chemical and physical properties must be considered, which is only possible after a thorough investigation of each kind of oil. * This class of oil has been added by the author from his own investigations of earthnut oil.

R

373

210

DIESEL OIL-ENGINE.

MARCH 1912.

APPENDIX II. SPECIFICATIONS OF TAR-OIL SUITABLE FOR DIESEL ENGINES. (From the German Tar Production Trust at Essen-Ruhr.) (1) Tar-oils should not contain more than a trace of constituents insoluble in xylol. The test on this is performed as follows:—25 grammes (0.88 oz. av.) of oil are mixed with 25 cm3 (1.525 cub. in.) of xylol, shaken and filtered. The filter-paper before being used is dried and weighed, and after filtration has taken place it is thoroughly washed with hot xylol. After re-drying, the weight should not be increased by more than 0 • 1 gr. (2) The water contents should not exceed 1 per cent. The testing of the water contents is made by the wellknown xylol method. (3) The residue of the coke should not exceed 3 per cent. (4) When performing the boiling analysis, at least 60 per cent, by volume of the oil should be distilled on heating up to 300° 0. (572° F.). The boiling and analysis should be carried out according to the rules laid down by the Trust. (5) The minimum calorific power must not be less than 8,800 cal. per kg. For oils of less calorific power the purchaser has the right of deducting 2 per cent, of the net price of the delivered oil, for each 100 cal. below this minimum. (6) The flash-point, as determined in an open crucible by Yon Holde's method for lubricating oils, must not be below 65° C. (149° F.). (7) The oil must be quite fluid at 15° C. (59° F.). The purchaser has not the right to reject oils on the ground that emulsions appear after five minutes' stirring when the oil is cooled to 8° (46° F.).

374

MARCH 1912.

DIESEL OIL-ENGINE.

211

Purchasers should be urged to fit their oil-storing tanks and oil-pipes with warming arrangements to redissolve emulsions caused by the temperature falling below 15° C. (8) If emulsions have been caused by the cooling of the oils in the tank during transport, the purchaser must redissolve them by means of this apparatus. Insoluble residues may be deducted from the weight of oil supplied.

R 2 375

Fig. 7.' First Experimental Diesel Engine, 1893.

Fig. 8. Later Experimental Pattern, 1895-96.

Fig. 9. First Complete Engine, 18 H.P., 1897. Showing brake and other testing apparatus.

pl.

4,

Fig. 10. Trunk Piston Type. 70 to 90 H.P. 1901. Cyl. 15.75 inch diam. 26.62 inch stroke. 150 r.p.m.

Pl. Fig. 11. Three-cylinder 300 H.P. Engine. 1906.

5.

DIESEL O I L - E N G I N E .

Plate 6.

Fig. 12. Valves in Chamber cast on side of Cylinder.

Fig. 13. Four-stroke Cycle High-speed Engine, 1909. Showing Box-frame and Air-pump.

378

Mechanical Engineers 1912.

DIESEL

Fig. 14. Fourstroke Cycle. 5 H.P. 1909.

OIL-ENGINE.

Plate 7.

Fig. 15. Double cyl. 4-stroke Cycle. 10 H.P.

Fig. 16. Horizontal Type. 50 H.P.

Fig. 20. First Marine Engine for Canal-boat. 20 H.P. 1902-3.

379

Mechanical Engineers 1912.

DIESEL

OIL-ENGINE.

Fig. 21. Four-stroke Cycle Engine. 120 H.P. 1908.

Fig. 22. Experimental reversible 4-cyl. 4-stroke cycle. 30 H.P. 600 r. p.m. 1909.

Fig. 23. Six-cyl. Marine Engine. 300 r. p.m. 191L Scavenging Valves below.

380 Mechanical Engineers 1912.

Plate 8.

Fig. 24. Single-cyl., Single-acting Two-stroke Cycle. 1,200 H.P. 1911.

Frederick William Lanchester (1868-1946) Educated at various establishments, including Finsbury Technical College, the bright and impatient Lanchester began his professional life without qualifications. A brief spell at the Patent Office was followed by employment at the Forward Gas Engine Company in Birmingham. Here, he devised a pendulum governor and was appointed Works Manager in 1890. When Lanchester invented a gas engine starter this brought him into contact with Dugald Clerk (1854-1932) who became a lifelong friend. It also brought income, particularly from the Crossley Brothers Company, allowing Lanchester to consider beginning his own family business. After a false start as makers of bicycle parts, the Lanchester brothers, Frederick, George, and Frank, started building automobiles to Frederick's original designs. A prototype was completed in 1895 - the first all British car. The results were sufficiently encouraging to allow the formation of the Lanchester Engine Company in 1899. Lanchester provided many individually successful innovations, including torsional vibration dampers and worm gearing. However, his vision was in conceiving the vehicle as a precision-engineered, purpose-built machine, not as an adapted horseless carriage. This same scientific approach informed his work on flight, which was a parallel interest. From the early 1890s, Lanchester had experimented with model aeroplanes. On his formal return to the subject in the book Aerial Flight (1907-08) Lanchester produced, a definitive study of aviation. He was practically the originator of the modern automobile, but in theoretical 'terms he was father of the aeroplane. Lanchester's later years were active and honoured. He contributed numerous papers to the Proceedings of the Institution of Automobile Engineers, where he served as President. His book publications ranged from Aircraft in Warfare (1916) to Relativity (1935), a manifestation of his active and creative mind. Aged 70, Lanchester delivered the IMechE's Thomas Hawksley lecture on The gas engine and after, marking a return to his most youthful enterprises.

381

THE FLYING MACHINE.

171

THE FLYING MACHINE: THE AEEOFOIL IN THE LIGHT OF THEORY AND EXPERIMENT.

By F. W. LANCHESTER. M.Inst.C.E. (MEMBER OF COUNCIL.)

THE design of the wing member, or aerofoil, of a flying machine, may be based either on the theoretical consideration of the function it has to perform, or alternatively on the ascertained results of scale model or full scale experiment. These two methods or points of view by which the problem may be approached are to be regarded as not of necessity overlapping; thus, on the one hand a designer may base his work entirely and absolutely on the result of experiment, without worrying himself in the least degree as to the why and wherefore; or on the other hand, he may work entirely from theoretical considerations and produce an aerofoil whose lift and lift/ drift* ratio he will know within certain limits without actual trial. If he elect to proceed on an empirical basis, he observes, for example, that a certain aerofoil model, submitted to trial in a wind channel (or on a whirling arm), at a certain angle and speed gives some definite lifting reaction and lift/drift ratio, and from this figure he may determine the area and calculate the other data required for a full scale aerofoil to lift the required load; his calculations in this respect are based on the V2 law and on corrections to take account of the laws of dynamic similarity. Allowances also are usually made for structural members, etc., not represented in the scale model. There is, evidently, nothing to prevent, on the one hand, the designer who is working from pure theory from verifying his results experimentally, or, on the othe* * The word "drift" in its present usage made its first appearance as an example of aeronautical slang, and has been widely adopted ; it would be better to say resistance. 383

172

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

hand, the designer who is basing his work on scale model experiment from endeavouring to effect such improvements as may be suggested by his theoretical knowledge: the point which the author desires to make at the present juncture is that there are two methods, each of which is, within reasonable limits, capable of giving unaided the data required. Unfortunately, those who at present have at command facilities for experimental work tend to rely too much on pure empiricism, not making a sufficient use of theoretical knowledge, and so experimental programmes are drawn up which by no means represent the best or most economical mode of procedure. In order that there shall be no doubt as to the foregoing, i.e. the present-day neglect of theory, the author would cite the fact that an aerofoil constructed by him in 1894,* and used almost exclusively for his experimental models of that period, was mainly designed on the basis of the theoretical investigations subsequently published in his work "Aerial F light, " The aerofoil in question, tested with modern appliances, has shown a lift/drift ratio somewhat greater than 17 : 1, which to-day (in spite of the experimental work which has since been done in the wind channels of the National Physical Laboratory at Teddington, the Aeronautical Laboratory at Gottingen, in Germany, and in the laboratory of Mr. .Eiffel, in Paris) stands as a record which has not been beaten. The author gives this fact at the outset by way of negative encouragement (and by way of a challenge) to those who affect to believe that direct experiment is the only method by which problems in aeronautics can be successfully tackled. The modern tendency has been to establish data for each arid every trial form of aerofoil by wind channel experiment, and to search more or less empirically for forms which show better results than those already obtained. The author is of opinion that the theoretical methods and lines of reasoning commonly followed in the work in question are at fault, especially, for example, when it is sought to assign independent values to the upper and lower containing surfaces of the foil, or when, again, the question of aspect ratio is studied apart from that of camber. On the other * At the date in question scarcely a sing-le experimental determination with other than flat planes was available. Constable, London. J Tested both at the National Physical Laboratory and at Gottingeu (Comp. James Forrest Lecture Inst. 0. E. 1914, Appendix III.).

384

THE FLYING M A C H I N E .

173

hand, it must be remembered that the questions entering- into an investigation are highly complex. There are factors of importance, such as those which tend to permit of flying over the greatest possible range of speed (rather than those which conduce to the highest economy at some particular speed); also such questions as position of centre of pressure—variations of centre of pressure with change of attitude, etc.—these are matters in the investigation of which purely empirical methods may be justified. The empirical method is, in brief, to take some form—we may say any form—of aerofoil, and ascertain its behaviour, namely, its pressure constants, etc. in the wind channel or on the arm of a whirling table. The early experimenters, as Hutton, Vince, Dines and Langley, confined their attention mainly to actual planes, that is to say, to aerofoils of flat form, in which the main variants were angle and the proportion of breadth to length (fore and aft)—in modern phraseology, the aspect ratio. More recent experiments of a similar character have been carried out by the National Physical Laboratory in this country and by Mr. Eiffel in Paris, aerofoils of wing-like or pterygoid section having been investigated in addition to planes, some of the said sections being based on the aerofoils of actual machines of well-known type; others were expressly devised with a view either to demonstrating some particular point, or to obtaining improved results in one direction or another. In its essence, the empirical method is an entirely simple matter.* The subject of experiment, the aerofoil, is mounted in such a way as to permit, for instance, measurements to be made of the lifting component and resistance component, commonly called for short lift and drift, for variations of attitude, results being tabulated or plotted as graphs. A plotting may, for example, be given as a graph in which the angle of attitude is denoted as abscissa, and the lift or drift, or other function of the angle, is shown as ordinate, each different form of plane or aerofoil, or different aspect ratio, being plotted as a separate curve. Examples of this method of plotting are given in Fig. 1. According to another method of plotting, the total reaction is plotted as a vector quantity, the magnitude of the lift and drift being obtained by an ordinary * The simplicity of the matter in theory is in great contrast to the elaboration of apparatus sometimes found necessary to obtain consistent and accurate results. The attitude is that which defines the angular position of the aerofoil about a horizontal axis transverse to the line of flight.

385

174

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

parallelogram of forces as co-ordinates, Fig. 2. According to this method, the angle of attitude of the plane appears to an irregular scale on the graph itself. Briefly, the method shown in Fig. 1 is the most generally serviceable where exact information is wanted, as round about the region of least resistance, or greatest lift, whereas the polar method (Fig. 2) has some advantage as being a multum in parvo by which the whole properties of an aerofoil from zero angle to 90 degrees may be represented in a compact form: if desired, the locus of the centre of

FIG. 1. pressure may conveniently be indicated on the same diagram. This method is that mainly used by Mr. Eiffel in the publication of his work. The National Physical Laboratory results, as given in the Reports of the Advisory Committee for Aeronautics, are, for the reasons stated, usually in the form shown in Fig. 1. This also is the method more commonly adopted. From our standpoint as engineers, the empirical method reduces the matter to its utmost simplicity. We leave the onus of the experimental work to the Aeronautical Laboratory, and 386

THE FLYING MACHINE.

175

take "ready made," for the purposes of construction, a form denned in advance; we may establish at once from the graph, for each angle of attack, exactly the speed necessary to sustentatiou and the resistance to be faced. We do not concern ourselves at all

FIG. 2. with theoretical questions of the form of the air flow round the foil, but accept established experiment and make full use of it, just as the designer of a steamship may take without question a known form of hull, either one for which experience has already provided the necessary data, or one which has been made the subject of 387

176

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

investigation, in the experimental tank. For the commercial constructor of aeroplanes the above may be deemed sufficient, but it is not a progressive method; as engineers we cannot (at present at least) afford to relinquish entirely the theoretical study of aerofoil design. Given the necessary apparatus, working empirically without real theoretical [understanding, it is always possible to accomplish a great deal, but it is evident that results cannot be achieved in this way either in a minimum of time, or at least cost, and if in these respects experimental efficiency is being considered, it is necessary to have some sound and satisfactory theoretical basis on which to work. By a sound and satisfactory basis, the author does not mean of necessity a complete treatment or method based, like electrical theory, on the exact concepts of the mathematician, but rather a theory whose fundamental ideas are in accordance with the facts of the problem, in the main qualitatively right and capable of pointing the direction, even if not always adequate to measuring the distance. A theory which is properly founded does not require to be quantitatively complete in order to constitute an adequate guide to scientific laboratory investigation. § 1. THE THEORY OF SUSTENTATION BASED ON VORTEX MOTION. The foundation of the theory of sustentation in flight, and of the supporting member, the aerofoil, is, as laid down in the author's "Aerodynamics," properly to be sought in the study of the vortex system set up in the air by the distribution of pressure on the aerofoil in motion as constituting the reaction by which the load is sustained; it has, however, proved difficult to attack the problem of sustentation directly from our knowledge of vortex theory. The author is not acquainted with any theoretical treatment giving results which accord as closely with experiment as his own. This is, in effect, an application and extension of the Newtonian method as applied by Kankine and Froude to the theory of marine propulsion; it rests directly on the third law of motion, and involves that which has been sometimes termed the doctrine of the continuous communication of momentum; the fundamental equation is that the force sustained by the fluid reaction is numerically equivalent to the momentum communicated per second to the fluid: in the case of the aeronautical problem, the air. 388

THE FLYING M A C H I N E .

177

In the author's work it is shown, firstly, that the immediate application of this method does not lead to conclusions which accord with experience; and secondly, that by the aid of certain auxiliary concepts (introduced into the theory as based on a study of vortex motion) it has been found possible to frame a regime in connection, with which the Newtonian method is made to give results which are in close agreement with experiment. The first step, therefore, in the preliminary consideration of the subject is the detailed study, from the point of view of vortex theory, of the motions generated in the air by the aerofoil as due to its function in sustaining the load, and which otherwise accompany and surround the aerofoil in its flight.

FIG. 3. § 2. THE DISTRIBUTION OF PRESSURE ON THE AEROFOIL AND ITS CONSEQUENCE. We will now examine the distribution of pressure on an aerofoil as we know it by experiment to exist;* in Pig. 3 it is presumed that we are viewing the aerofoil along (in the direction of) the axis of flight. Now, it is an experimental fact that in any ordinary design of aerofoil, such as used in actual construction, the pressure difference is maximum in the central region and tails off towards the extremities, and that this applies both to the positive pressure below and the negative pressure, or suction, above; the relation * The author finds here the option of approaching the present subject either in the manner of treatment already adopted in his Aerodynamics, that is to say, by a process of cold-blooded deduction with the least possible assumption as to experimental fact, or to deal, as he is here doing, with certain experimentally ascertained facts, as involving of necessity certain consequences in the behaviour of the air. LANCHESTER.

M

389

178

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

of the pressure to the suction, therefore, has no immediate bearing on the present argument. A result of this distribution of pressure is that any small unit mass of air passing beneath the aerofoil receives (in addition to downward acceleration) an acceleration component outwardly towards the aerofoil extremities, as shown in the figure by the arrows. Likewise, air passing above the foil receives an acceleration inwards towards the central region, since the vacuum there is greatest; thus, looking at the aerofoil in plan, Fig. 4, we see that the path of the air stream passingover the aerofoil is convergent, as shown by solid lines, and the air passing under the foil is divergent, as shown by the dotted lines, the curvature of the path in the immediate vicinity to the

FIG. 4. aerofoil being an indication of the acceleration applied. When the two adjacent layers of air, one from the upper surface and the other from the lower surface, rejoin at the after edge of the aerofoil, they are thus found to have relative motion impressed upon them, as already given by the arrows in Fig. 3. Nqw this condition denotes that there is a surface (or stratum) of discontinuity existing in the track of the aerofoil behind the right and left hand wings, otherwise known as a vortex sheet (the surface of gyration of Helmholtz), the air forming which contains rotation. We also know that such a vortex sheet or stratum rapidly splits up into a number of vortex filaments belonging to a common vortex system involving the surrounding fluid. That this phenomenon is no purely local manifestation becomes quite evident when we consider the fact that the pressure region under each

390

THE FLYING MACHINE.

179

aerofoil wing and partial vacuum above it give rise everywhere to the air in the vicinity of a circulatory movement, as indicated in Fig. 5, the said circulatory motion being superposed, from the point of view of an observer moving with the machine, on the motion of translation. The whole of the foregoing is deduced in the author's " Aerodynamics," apart from any experimental evidence such as taken here as the basis, the motions being there depicted more particularly in Figs. 83, 85, 86. Thus, the disturbance left in the wake of an advancing aerofoil consists of two equal and opposite vortices—a vortex pair in fact—whose " core " consists of a number of distributed vortex filaments all bound up in the common cyclic motion, in accordance with well established principles, having a downward precession, the movement of each vortex trunk as a

FIG. 5. whole being that proper to the motion of its fellow: it will be demonstrated that it is the reaction due to the continuous production of this vortex pair tvhich is the eventual source of sustentation in flight. § 3. INDEPENDENT EVIDENCES OF THE TRAILING VORTEX PAIR. Before dealing with the whole consequences which follow from the established existence of this vortex pair, certain comments will not be out of place as to the independent evidences, such as they are, for the existence of these vortices, and some of the indications that the designer of the aerofoil may gather by taking them into account. Firstly, it is not difficult by passing an inclined plane through air charged with smoke to see with one's own eyes the vortex motion produced. Secondly, the extent of the " wash " left behind M 2 391

180

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

by the aerofoil, and acting on the tail member astern of it, can be correctly estimated on the basis of this residuary vortex motion (Appendix I.). Thirdly, as stated in the author's ''Aerodynamics" (p. 174, Fig. 80), if an inclined plane be moved under water in the vicinity of the surface, indications of vortex motion are found in the dimples which can be seen in the vicinity of the wing terminations. To the designer, the immediate importance of considering the function of the aerofoil in producing these trailing vortices lies in his realisation of the need for the pressure diminution towards the aerofoil extremities; manifestly, if the pressure is carried too uniformly into the terminal regions, the rotational core of the vortices will be concentrated and the energy left behind in the vortices themselves becomes disproportionately great; that is to say, the motion will tend to become a strong local vortex at the wing extremities containing great energy, instead of being more diffuse and consequently more economical. Probably, there is yet considerable work to be done in the direction of improvement from this point of view. § 4. THE CYCLIC COMPONENT IN THE MOTION AROUND THE AEROFOIL. We will now proceed to consider a very important consequence of the vortex theory of sustentation. It may be taken that it is impossible that two vortices of opposite hand should be attached, or attach themselves, to any material body without the vortex motion extending continuously around the body from one to the other. From the strict standpoint of hydrodynamic theory, we know that before cyclic motion can be regarded as possible, the region involved must be doubly or multiply connected, and we know that in the case of an aerofoil (under actual conditions in three dimensions), this condition is not complied with. But we also know perfectly well that in the real fluid the conditions of double connectivity can be and are simulated, in so far as the central bar to connectivity is concerned, by a filament or column of fluid in rotation. In some cases this filament or column forms a complete loop as in the ordinary smoke ring; in other easeis its extremities attach to some boundary surface, such as the earth or sea in the case of a whirlwind or waterspout, or to the surface of the water (from beneath) in the case of the half vortex hoop 392

THE FLYING MACHINE.

181

which is generated by a stroke of a paddle in water. A certain difficulty presents itself in the case of vortices trailing away indefinitely into the distant air (or water) without apparently any direct boundary connection, as in the case of the lateral vortices we have been considering; probably, it will be found when the theory of the real fluid is sufficiently understood that this condition is dynamically possible, and is in some way related to a rate of decay or degeneration; in any case, it is definitely a condition which we know may exist if only from optical demonstration, hence even if we cannot for the time being satisfy ourselves as to the

FIG. 6. theoretical aspect of the subject, we must at least be prepared to accept it as a fact. * Thus each of the lateral trailing vortices has one of its extremities dying away in the distant air astern, where its motion is undergoing decay, and its other extremity, its forward end, attached to the aerofoil, where it is continuously being renewed; thus, the aerofoil forms, as it were, a bridge connecting two vortices which (viewed from their points of attachment') are * Compare Appendix II.

393

182

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

of opposite hand. Now it is quite clear when we view the aerofoil and vortex system as a whole, that the interruption in the simple connectivity of the region introduced by the rotation in the vortex cores is maintained between them by this bridge, i.e., by the aerofoil itself, and moreover that the vortex trunks on the two flanks are really part of a single system whose cyclic component surrounds the aerofoil in at least the same strength as it exists at the region of attachment, Fig. 6. The validity of this view can scarcely be disputed; the sudden termination of the vortices at the wing extremities would be quite inconsistent with hydrodynamic theory. Beyond this, all the analogies which are known to exist between electro-magnetic and hydrodynamic theory point in the same direction: a line of magnetic force, for instance, passing through a particle of iron is not and cannot be regarded as two separate lines terminating on the surface of the iron and unconnected—they are part of a continuous unit. Thus, the author regards the two trailed vortices as a definite proof of the existence of a cyclic component of equal strength in the motion surrounding the aerofoil itself. § 5. CONSEQUENCES or THE CYCLIC COMPONENT. STREAM LINES.

RESULTANT

Now the vortex motion, or, more correctly, cyclic motion, round the aerofoil is under somewhat different conditions from those which obtain in the flank or trailing continuations; firstly, its core is solid and is subject to a powerful force or reaction (i.e., the load sustained) at right angles to the direction of flight; secondly, its extremities are co-terminous with and continuous with those of the flank vortices, so that there is no necessary dissipation of the vortex energy; thirdly, the axis of the cyclic system around the aerofoil is at right angles to the superposed motion of translation (the line of flight) instead of being in line with it as in the case of the trailing vortices. This latter condition results in an entire change in the geometrical form of the stream lines, sufficient almost to disguise the nature of the motion; thus, in the trailing vortices, the motions due to the vortex are in planes normal to the superposed translation, and so the translation and the circulatory motions are clearly separable by observation, whereas in the motion around the aerofoil the circulatory motion

394

THE FLYING MACHINE.

188

is in the same (vertical) plane as the translation, and the stream lines undergo a transformation which entirely alters their appearance. This is just as motion in a circle when superposed on a translation may become either a screw or a curve of the cycloid family; in the former case, the circular character of the motion is quite self-evident, in the latter it may be disguised almost beyond recognition. It is evidently a necessary step to learn to recognise and study the properties of cyclic motion superposed on translation in the same plane. The solution in certain cases is well known, and constitutes a problem in two dimensions. Thus, in the simplest case, where the cyclic motion takes place round a filament or cylinder of negligible size, the solution is the same as that of a conductor of small diameter carrying an electric current in a magnetic field, Fig. 7. Where the "core " of the cyclic system is a plane parallel to the direction of translation, the solution is given in Fig. 8, and Fig. 9 similarly represents the case of an elliptic cylinder.* Analysis shows that in all these cases, indeed in every similar case of cyclic superposition, the dynamic conditions can only be satisfied provided that a reaction be applied from without to the " core " (whether it be a filament, a plane, or whatsoever its form) at right angles to the direction of flight (that is to say, at right1 angles to the superposed translation), and as shown in the figures* this applied force is downwards, as it is in fact under the conditions of motion already deduced from the assumption of the continuity of the trailing vortices. This downward force applied from without is applied by gravity—it is load sustained. The matter may be put another way; the motions shown in Figs. 7, 8, and 9, result from their nature in a lifting reaction on the filament, plane, or cylinder, and this in the actual problem definitely constitutes a lifting reaction on the aerofoil. On examining the three figures given, we see that none of them represent a type of flow which could exist in its entirety in any real fluid such as air; the stream lines undergo many sharp turns, which in practice would give rise to eddies and motion of the discontinuous type, just as when a body of not properly "faired" form is * Figs. 7, 8, and 9 are taken from the author's "Aerodynamics," where the plotting* are also given for two superposed planes, also for a plane in the vicinity of a boundary surface.

395

FIG. 7.

FIG. 8.

FIG. 9.

THE FLYING M A C H I N E .

187

drawn or propelled through air or water; and we are led to seek for indications of the features required in order that the sectional form of the aerofoil may be conformable to the lines of flow. One of the most important features to note is that the air in the region being entered by the aerofoil has an upward motion or trend: this betokens that in order to meet the air conformably the aerofoil must have a dipping front edge. Here we have at once an immediate and powerful confirmation of the cyclic or vortex theory as set forth: it had long been known that the dipping front edge was a feature commonly found in the wings of birds before the present theory had been formulated, and certain experimenters, as Horatio Phillips in England and Lilienthal in Germany, had discovered its value in connection with artificial or mechanical flight. The explanations offered as alternative to that of the author, are either a " paraphrase " of the cyclic theory, or demonstrably unsound. It is shown subsequently in the present paper that vortex theory leads to direct quantitative results of value, but at present the exact interpretation is in some respects by no means easy or certain. It is difficult, for example, to define the precise role played by the cyclic motion round the aerofoil, which neither gains nor loses energy; yet we know it to be essential, for it is not possible to frame a regime which accords with experimental fact without taking it into account.* In a section of his work devoted to the screw propeller, the author has suggested that the cyclic component of the periptery (= around the wing) in effect constitutes a kind of dynamic tool or device by which the deeper layers of air are involved in a motion initially confined to the immediate neighbourhood of the aerofoil (the conception being that the direct action of the aerofoil concerns more particularly motions of the air which may be represented quantitatively by a layer of limited thickness, defined as the " sweep " of the foil, and represented roughly by the distance apart at which two aerofoils may be superposed without material interference), and that a more extended mass of the air, defined as that within the "peripteral area" or "peripteral zone," eventually takes up the downward momentum and in reality is the measure of the uniform * "Aerodynamics," § 160 et seq "Aerodynamics," § 210.

399

188

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

motion equivalent of the total air movement to which the sustaining- reaction is due. More recently, the author has found new lines on which to explain the relation of the two portions of the vortex system, namely, the portion surrounding the aerofoil and the two flank trailing continuations. The new point of view has the advantage of disclosing more exactly the working of the two parts of the system as a dynamic whole, and so is of interest both as leading to new results and as confirming the validity of the original work. § G. A NEW HYPOTHESIS. Let us conceive an artificial atmosphere in which the constituent particles have restricted motion—they are only permitted freedom in two dimensions; briefly, they are constrained to move only in planes at right angles to the line of flight. We cannot literally imagine space to be cut up into thin vertical slices by partitions, since this would restrict also the motion of the aero-

FIG. 10. foil, which must be permitted its ordinary freedom of movement, but with this reservation the kind of restraint which would be imposed by numerous vertical walls is identically that contemplated. It will be seen that in each vertical stratum the motion of the air will be in two dimensions, and the conditions of vortex motion will be more readily satisfied, since the terminations of the rotary cores cease to present any difficulty. As the aerofoil traverses the system of strata normally, it deals with the contents of each in turn, and under the conditions postulated, the history of one will in due course be the history of all: thus, the problem is reduced to a single case of two-dimensional vortex motion. Now, in order to generate a vortex pair about any two points representing the vortex cores, x and x', Fig. 10, hydrodynamic theory tells us that it is necessary to apply an impulse to the fluid along and normal to the line x, x', as shown by the arrow heads. An "impulse" is, of course, understood to be an infinite force applied for an infinitesimal time, but we suppose in reality the quantities to be finite, and we have to 400

THE FLYING MACHINE.

189

imagine a pressure difference for a certain brief interval of time established by external agency at the line x, x'. But a pressure difference of uniform value will result in much more rapid motion in the vicinity of the terminal points x and x' than in the intervening- region, this difference being in accordance with the wellestablished laws of cyclic motion. The essential character of the generating impulse is that the momentum represented by its action is constant per unit area. Now the forces on X X' are brought to bear by the action of the aerofoil, which we may represent in side elevation as in Fig. 11 ; we may take it for the present purpose to be of constant sectional form from end to end. Also,

FIG. 11. since we have now no possibility of disturbance in front of the advancing aerofoil, we may assume that its correct form of " entrance " is parallel to the line of motion—horizontal in the figure—and that in order to exert a uniform acceleration on the air during its passage, its form will be parabolic, or, as drawn, approximately a circular arc; the trail angle r1 will then be the factor which determines the downward velocity v2 imparted to the air, which will be given by the expression V tan where V is the velocity of flight. But this (under the conditions specified) will be constant throughout the length of the foil, and therefore will not give rise to a simple vortex pair; we may, however, suppose this constant velocity to represent a number of superposed force distributions over narrower and narrower bases as indi401

190

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

eated diagrammatically in Fig. 12, which will give rise to a number of superposed cyclic systems whose core will be a vortex sheet or surface of gyration, and we recognise at once the whole system as already described and depicted in Figs. 4 and 5. Under our restricted hypothesis it is clear that the fore-andaft dimension of the aerofoil is a matter of no immediate importance; the greater the aspect ratio (i.e., the less the fore and aft dimension), the more nearly impulsive will be the action, that is to say, the shorter the time and the higher the pressure system, but so long as the value of w is constant the variation in the aspect ratio is of no effect on the velocity given to the fluid in its different parts. It is quite true that with great fore-and-aft length (= lower aspect ratio) the aerofoil would give a greater eddy at its extremities, and in practice doubtless this effect is not

FIG. 12. negligible; but for the purpose of the present argument it can be ignored: it is referred to later. § 7. REAL CONDITIONS RESTORED. Now, when we take awajr the restriction as to the motion of the fluid, what is the result and what inferences can we draw? Firstly, it may be noted that the circulatory motion around the aerofoil will be far greater for one of high aspect ratio than for one of low, and if this represented lost motion or leakage the foil of high aspect ratio would be at an immediate and overwhelming disadvantage. But on the theory of the continuity of the vortex system the circulation in the vertical plane of flight around the aerofoil is a cyclic motion, which, compounded on one of translation (relatively to the motion of the aerofoil) constitutes a conservative system which neither gains nor loses energy, and results, in sum, in no permanent change in the air velocity; so the 402

THE FLYING MACHINE.

191

conditions as to the generation of the flank trailing vortices are not in effect ultimately different from those of the two-dimensional motion postulated as the basis of the argument of the preceding paragraphs. In other words, the air although receiving motions in the third dimension (the line of flight) has these motions so bound and regulated by the peripteral cyclic system, and the energy of thess motions so strictly conserved, that the result is in effect the same as if the additional degree of freedom were denied to it. In practice we know that it is not merely sufficient that the aerofoil should act as an obstacle to connectivity; it is also necessary, in order that the cyclic motion should conserve its energy, that the sectional form of the aerofoil should be conformable to the lines of flow. Thus, the forms shown in Figs. 7, 8, and 9 would not comply with this condition in a real fluid, they

FIG. 13. would set up eddies or generate vortices foreign to the regime. It is evidently necessary, as we already know, to adopt an arched form with a dipping front edge in order to meet the up current component conformably, and since this current is part of a system superposed on that proper to the flank vortices, this arched section must be regarded as obtained by the bending of the aerofoil beyond the curvature indicated by Fig. 11, as shown in Fig. 13. § 8. THE ASPECT RATIO OF LEAST RESISTANCE. If we imagine the fore-and-aft dimension to be progressively reduced, with a corresponding increase in the aspect ratio, the velocity of the up current will locally become greater, for although the strength of the vortex motion will remain unchanged the velocity will be greater owing to the reduction in the circuit of the "core." Hence, the necessary steepness of the camber will become greater the less the " chord," and manifestly 403

192

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

a point will be reached at which the steepness of the camber will itself set up eddies, when the resistance will be rapidly augmented with any further reduction of the chord dimension. In addition to the above, structural considerations also begin to tell; the aerofoil section has, so to speak, to "swallow" or contain the spar sections necessary for its strength, and the vertical depth of these spars cannot be reduced beyond a certain point; consequently, when the chord dimension is unduly limited the solid conformation of the aerofoil section is liable to become too bluff, and to offer excessive resistance. On two counts, therefore, Ave have it that there is a practical limit to the reduction of the chord and the increase of aspect ratio, but there is still another. In designing a machine for variable flight speed, and in providing for slow flying, it is desirable to allow for the aerofoil being employed at other than its most desirable angle of attitude; when this is the case the supporting reaction is found to fall off more rapidly for an aerofoil of high aspect ratio than for one of moderate value. On the other hand, the advantage of the chord reduction is mainly that, with the reduction of surface ("wetted" surface in the language of the Naval Architect), the skin-frictional resistance is lowered, not quite in the relation of the reduction of surface, but nearly so. Thus the point to be sought in fixing the aspect ratio is that at which the increase of resistance as due to the two causes mentioned (steepness of camber and "form"), is equal to that saved on the score of skin-friction: this is the ordinary condition for maximum or minimum value. We are unable to give any mathematical solution owing to the fact that the law connecting resistance with the two items steepness and form is unknown; in the absence of information on this point the solution is a matter for the Laboratory. In discussing the hypothetical atmosphere confined to twodimensional motion it was mentioned that an aerofoil of low aspect ratio, i.e., having considerable length in the line of flight, would not comply very closely with the ideal conditions as represented by an impulse, and that its lateral extremities will become the seat of motion of a character foreign to the regime. We may regard the lateral ends of the aerofoil as giving rise to two vortex sheets, Fig. 14 (a) (otherwise known as surfaces of discontinuity, or surfaces of gyration), which ultimately break up and become part of the trailing vortex system. Now the points 404

THE FLYING MACHINE.

193

of origin of these vortices are more widely separated than the span, and so the aerofoil of low aspect ratio (Fig. 14 (b)) may bo expected to have a somewhat wider " seat " in the air than the high aspect ratio foil (c) of the same span. Were this not the case, for any given load sustained by an aerofoil, the energy aerodynamically necessary would depend upon the span alone,* and be independent of the chord dimension. An. examination of the forms of aerofoil found in nature is instructive; a few examples are given in Fig. 15. Thus, whatever the aspect ratio may be we never find anything resembling the

FIG. 14. square extremities of Fig. 14, and we may take it that the type of flow therein depicted, as at (a), is to be regarded as undesirable; either the extremities are found to taper off as in the Hawk, Fig. 15 (a) (and in the great majority of other birds), or to become ctenoid or serrated as in the Rook (b) or the Heron (c). The meaning of this is that in nature the " grading " down of the pressure reaction towards the wing extremities is universal; in the language of the aeronautical designer the aerofoil section is given a decided "wash-out." This feature has been imitated to a very slio-ht extent in actual aeroplane construction; it corresponds according to the present theory to a decided reduction in the vortex strength towards the outer margins of the gyration surface. * Inversely proportional to the square of the span. LANCHESTEK.

N

405

194

THE INSTITUTION OF AUTOMOBILE E N G I N E E R S .

FlG.

406

15.

THE FLYING MACHINE.

195

Another peculiarity or feature of less evident signification but worthy of note in the present connection, is that the widest part of the wing spread is by no means 'always to be found in the vicinity of the centre line, or rather adjacent to the body of the bird; the maximum width or chord dimension is often some considerable distance from the point of origin of the wing; this is illustrated in the three examples given. The feature in question does not appear to be related to any anatomical consideration, it may possibly denote that, as indicated in Fig. 14 (b), the surfaces of gyration only extend along the outer portion of the wings. Alternatively, the feature may be quite unconnected with any real variation in the vortex distribution; it may be merely an indication of the higher velocity of the air motion in the vicinity of the convexity of the body, as adapting the wing locally more accurately to its correct P/V2 value. § 9. AUTHOR'S PRESENT AND FORMER TREATMENT COMPARED. SUBSTANTIAL IDENTITY DEMONSTRATED. To those who are familiar with the author's " Aerodynamics" the fact will at once be apparent that although the whole line of reasoning adopted in the present paper differs from that previously employed, the general result is the same. On looking into the matter more closely it will also be observed how some of the conclusions given in the previous work as the outcome of lengthy investigations arise immediately and directly as the natural outcome of the new line of treatment. Thus the conception of the peripteral area and peripteral zone introduced into the author's " Aerodynamics" in connection with a discussion on the screw propeller, § 210 (Ch. IX.), and allowed to remain rather abstract in character, may now be identified as an area representing by its content the equivalent mass of the trailing vortices; it is here that we find the augmented mass of fluid in downward motion, which, according to the regular application of the Newtonian method, represents by its downward momentum the reaction of the sustained load. We know that in mathematical theory any free vortex ring or (in two-dimensional motion) vortex pair may be regarded as carrying the momentum communicated by the impulse by which it is generated,* and so is * That any such vortex ring or pair does not as a whole contain momentum the author has definitely proved ; however, we may nevertheless legitimately regard it as carrying a definite quantity of momentum just as though it were N 2

407

196

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

the equivalent of some appropriate mass of fluid in motion; it is, strictly speaking, the vertical cross section of this hypothetical mass which, in the author's former work, constitutes the peripteral area. It will now be shown that the results given by the present line of reasoning are in substance identical with those obtained in the author's previous work. In both methods of treatment we have the aerofoil dealing directly with a cyclic component in the motion of the air, and ultimately leaving in the wake downward momentum whose value per second is the equivalent of the load sustained. Thus the direct action of the aerofoil is measured by the sum of the momenta of an up current received and a down current discharged; it is in fact the downward momentum represented by the continuous reversal in the vertical component of the flow of a mass of fluid, just as, for example, in the Pelton wheel. The energy spent or work done by the aerofoil is thus the difference between that received from the up current and given or imparted to the down current, or, in the symbols employed in the author's " Aerodvnamics," and If the mass per second representing the trailing vortex pair, that passing through the peripteral area, be denoted by M and v be its downward velocity, we also have. and or, existent. The author suggests that this quantity should be termed the " pseudomomentum " of the vortex ; it is ever equal to that of the impulse by which it is generated. The position may be illustrated by an analogy; the actual momentum of a hoop in motion is just the same for any given velocity, whether it be in rotation without skidding, or skidding without rotation. In the former case, however, its effective momentum is double its actual momentum ; thus, when it is brought to rest by the application of a force, the rolling hoop requires twice the force applied for a given time that it would require if not in rotation. It is understood, of course, that in the case of the rolling hoop when brought to rest, no skidding takes place. In the vortex ring or pair, as in the case of the hoop when it is brought to rest, there is a reaction in the opposite direction ; in the first case this is exerted on the confines of the fluid region ; in the second case, as a tangental force on the surface with which the hoop is in contact.

408

THE F L Y I N G MACHINE.

197

and

by(i) or (2)

or, also by (1) a

In the author's "Aerodynamics" a is represented by a constant s; substituting we have,

Now the mass per second m is that coming within the " sweep " of the aerofoil as denned by an area K A (A being the area of the foil), hence the peripteral area is,

The above is the result and in substance the reasoning given in §'210 of "Aerodynamics." Since our present basis is that of an aerofoil of fixed span and variable chord, we will take the span I as basis, and express the peripteral area in terms of l2; thus, employing the values of the constants K and e as given in the author's " Aerodynamics,"* we obtain the results given in Table I. Referring to this Table, we see that the side of the square representing the peripteral area varies from being approximately equal to the span l in the case of an aerofoil of aspect ratio =3, to 0'83l in the case of aspect ratio = 12. Now the side of this square may be taken as roughly representing the base of the vortex pair, since the equivalent mass of a vortex pair is approximately equal to that of a mass of fluid represented by the content of the square on its base, so that we * Table III., p. 261. In the case of a simple vortex pair generated by a uniform impulse as in Fig. 10, the area representing the equivalent mass isp/3times the square on the base X X', or approximately 1.05. See Appendix III.

409

198

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

have definitely a quantitative confirmation of the theory presented in the present communication based upon the constants determined by totally different and independent methods. Moreover, we find that so minute a matter as the influence of the end effect or eddy, already noted, as tending to widen the base in the case of aerofoils of low aspect ratio, is accurately reflected in the figures obtained from the equation. Again, we find complete harmony between the two lines of treatment in the interpretation of equation (5) in conjunction with Fig. 13. We have already seen that in accordance with our present hypothesis the velocity v2 is (with reference to Fig. 11) TABLE I.* Aspect Ratio, n

Constants. K

1+e

£

K —••• 1-8

K/nX l + e V n- x 1 -e

7 8

1.03 1.064 1.10 1.12 1.14

0.48 0.54 0.59 0.62 0.65 0.68

2.85 3.45 4.13 4.70 5.30 6.0

0.975 0.93 0.91 0.88 0.87 0.86

10

1.175

0.72

7.2

0.85

12

1.195

0.75

8.4

0.83

3 4 5 6

1.00

given by the expression V tan n. When, as in Fig. 13, we add the arched section representing the path of flow of the cyclic component as a superposed system we have the angle of dip, the a of the previous investigation ("Aerodynamics") and an equal angle superposed on the angle 73, making a total trail angle and we have

* In the expression heading Column V., l2 = An ; hence

410

12

= * n

THE FLYING MACHINE.

199

Since we are working on the "small angie hypothesis"* the v2 of the present investigation is identical with the v of equation (5), and we see that the present treatment and that of the author'ts previous work give identical results. The agreement between the results of the author's previous and present work is not to be regarded as an independent confirmation

FIG. 16. but rather as a justification of the original theory; the present investigation may be looked upon as a development directed to elucidate much that might otherwise be regarded as obscure in the regime. Put concisely the earlier investigation and the present deal with the same main problem, the mode of support, on the same founda* The author's method, as employed in his "Aerodynamics," is based on the assumption that the angles concerned in the sectional form of the aerofoi (relatively to the time of flight) come within the definition of a small angle, i.e., j (in circular measure)—sin j=tan j within permissible limits of error.

411

200

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

tion theory, but they begin at opposite ends; in the earlier work the cyclic motion around the aerofoil was taken as a basis and the remainder of the system was deduced as a corollary, the present line of argument begins with the ultimate or final step in the communication of momentum to the air in the trailing vortices, and works backwards to the motions and behaviour of the air in the more immediate vicinity of the aerofoil itself. § 10. QUANTITATIVE THEORETICAL TREATMENT. The cyclic or vortex theory of the aerofoil is capable of yielding

FIG. 17. quantitative results quite apart from any experimentally determined pressure values or constants whatever. There are difficulties at present, due to the limitations of mathematical analysis, but these will be without doubt overcome: if the mathematician fails us we can fall back upon graphic methods. Reverting to the hypothesis of § 6, we have seen that the simple " bifocal" vortex pair is impossible owing to the high velocity in the vicinity of the foci, or vortex filaments; this would 412

THE FLYING MACHINE.

201

betoken an aerofoil whose camber and angle /3 increase towards the lateral extremities to an indefinite degree, an altogether absurd and impracticable feature even from a theoretical standpoint. It has further been pointed out with reference to Fig. 11 that the more usual condition is that of uniform camber from end to end, but no attempt has been made so far either to carry the study of this condition to its logical conclusion, -or more generally to examine the relation which exists between the appropriate aerofoil camber and the vortex distribution. If we take as our basis for the purpose of illustration the case of uniform camber, at least if we suppose that which we term the primary camber (Fig. 11) to be uniform, we have the appropriate two-dimensional vortex solution ready to hand, for the linear motion impressed on the fluid will be constant throughout the length of the span, and the resulting system will be identical with that due to the normal motion of a plane the equations of which have been solved and the curves of flow plotted in Fig. 16. In the interpretation of this figure it is necessary to suppose an impulsive force applied to the plane, the latter then being withdrawn: in reality, the plane must be supposed to move through a short finite distance, and in our problem as it stands the line ordinarily understood to be the plane is actually a section of the aerofoil, and its brief motion represents the change taking place as the aerofoil traverses the vertical stratum represented by the plane of the paper. When the aerofoil has passed through the stratum under consideration, the latter is left without any break in its continuity, and with the motion as defined by its stream lines intact. The motion then becomes a vortex system, the core being constituted by the wake of the aerofoil as a vortex sheet or surface of gyration, as described with reference to Figs. 4 and 5. The equivalent mass of this particular system is known,* it is equal to that of the fluid content of a cylinder of circular section of diameter equal to the width of the plane, i.e., the span of the aerofoil. In other pl2 words the peripteral area = — whore I is the span. It may be

p

otherwise expressed as -An where A is the area of the foil and n the .aspect ratio. * "Aerodynamics," § 83. Also Hydrodynamics; Lamb, § 71. The stream line system is the same as for any elliptical cylinder whose foci are situated at the extremities of the line representing the impulse surface in Fig. 16.

413

202

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

§ 11. QUANTITATIVE THEORETICAL TREATMENT—continued. AN EXAMPLE. Let us be quite clear as to the position: we have Fig. 17, an aerofoil of span l, and of any aspect ratio we wish within reason, dealing with a mass of air represented by the volume swept by a circle (shown dotted in the figure), of a diameter equal to the span. This is the equivalent mass of air to which downward velocity is continuously imparted, representing a quantity of momentum per second corresponding to (in absolute units equal to) the load sustained. Thus, if as a numerical example we take the case of an aerofoil of 40ft. span, with a flight velocity of 100 ft./sec., sustaining a load = 1.800 pounds, we have

Hence v2 (velocity of downward discharge), .-. trail angle of primary camber,

The resistance may be calculated either from the mean angle of the camber, =n2 or from the independent calculation of the energy /sec. Thus:—

Alternatively,

or,

resistance

=

= 53 pounds (as before).

The identity of the results by the two methods merely serves to show (as, indeed, is otherwise obvious) that, under the conditions assumed, the horizontal component of the pressure reaction is correctly given by a resolution of forces based on an angle with the vertical equal to -xj/2. In other words, the resistance in question may be reckoned as due to the angle of the hypothetical chord.* * The chord, as measured from an actual aerofoil, is commonly very different from the chord with which we are concerned in theory.

414

THE FLYING MACHINE.

203

It is to be understood that the resistance now in question is only that portion of the total known as the aerodynamic resistance, the y-component of the author's "Aerodynamics." We may make our knowledge at the present juncture the basis of more general calculation; we will take for example the case of the aerofoil R.A.F. 6, of which tests have been made by the National Physical Laboratory. (Adv. Committee for Aeronautics, 1912-13, Eeport No. 72.) Since we have so far made no provision in the present quantitative exposition for correlating the angle of attack with the motions of the air, we shall deal only with the drift I lift ratio as a function of the absolute lift coefficient, the

FIG. 18. former being plotted in Fig. 18 as ordinates and the latter as abscissae. In Fig. 18, curve No. 1 is plotted from the same data as Fig. 8 of Report 72, to which reference has already been given ; the determinations relate to a model of the aerofoil known as "R.A.F. 6" about l/30th full size, namely, loin, by 2.5 i n . , aspect ratio = 6. The velocity is given as 30 ft./sec. Curve No. 2 is the calculated drift/lift as due to aerodynamic expenditure of work on the basis of circular column of air diameter = l in accordance with the present hypothesis. Curve No. 3 is the drift /lift as due to the skin-frictional or direct resistance of the aerofoil, based on a coefficient x —0.017 computed in the manner laid down in the author's James Forrest: Lecture, 1914, Appendix I. Curve No. 4 is the sum 415

204

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

of curves Nos. 2 and 3, and purports to represent on our present basis of theory the total drift/lift as given by curve No. 1 from direct experiment. It will be observed that the agreement between curves Nos. 1 and 4 is only exact in the immediate vicinity of minimum value; incidentally it may be remarked that this is the most important portion of the curve. The reason for the want of agreement at other points in the two curves is undoubtedly due to the fact that the camber of the model used in the experiments was particularly adapted to work with a lift coefficient (absolute units) round about 0'3, and when employed at greater coefficients (corresponding to steeper angles) its camber would evidently be insufficient and the resistance would be augmented by eddy making; when, on the other hand, the lift coefficient is lower the camber is too great and eddies again arise, resulting in increased resistance. In order to obtain a curve experimentally to agree with No. 4, the camber and angle ought to be varied in sympathy; that is to say, as the lift coefficient is increased both angle and camber should be increased pro raid. The arithmetic in connection with the plottings in Fig. 18 is given in Appendix IV. Curves Nos. 5 and 6 are based on a lower coefficient of skin-friction (x = 0.0105) as appropriate to the full-sized machine at 100 ft./sees, flight velocity; these, although of no immediate bearing on the discussion, are not without interest. § 12. QUANTITATIVE THEORETICAL TREATMENT—continued. THE ANGLE OF DIP. We have already seen that the angle of dip a (Fig. 13) is the factor which determines (in conjunction with the angle n) the angle of trail b, and thus also the theoretical camber of the aerofoil. It remains to be demonstrated how the angle a may be deduced or computed from our knowledge of the peripteral system. We will take the most rudimentary system as a starting point. In Fig. 19 we have the aerofoil span = l, and aspect ratio = n, generating a simple vortex pair about its extremities x, x'; then the cyclic motion around the aerofoil is of the same strength as that in the trailing vortices. Now confining ourselves to the middle section of the aerofoil, in the line of flight, we presume to know the angle n (Fig. 13) as given by our main calculation 416

THE FLYING MACHINE.

205

for sustentation, hence V tan n is the combined velocity of the two trailing vortices at a distance l/2 from the vortex foci. But the velocity of the cyclic component varies inversely as its distance from the focus; and for the single vortex the velocity at l/2 is7 tan n/2, or at a distance of half the chord the velocity is n V tan n/2 which is the value of V tan a required ; or tan a = n- ,tan n. 2

FIG, 19. On this basis the constant e of the author's "Aerodynamics" becomes (assuming always the hypothesis of the small angle) a result which is of the right order but some 15 per cent to 25 per cent too high. We know, however, that the system, consisting as it does of a single vortex pair, is in reality impossible; we will now apply the same method to the case of distributed vortex motion already considered, as generated by 417

206

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

the impulse system of Fig. 16. Without going fully into detail, it may be pointed out that the equivalent base of the trailing vortices is now very much less than the span of the aerofoil, the actual relation is p — 0.78, or the velocity V tan 73 corresponds to a distance from the equivalent vortex centre = 0.39 l, or, working as before for the single vortex at a distance equal to half the chord, we have 0.39nV tan n = V tan a, or On this basis the constant e —

The resulting values

are now within 10 per cent of those given in the table.

FIG. 20. In the case in point, taking the condition of least resistance of Fig. 18, the angle n works out as 0.07 (radian) and since n = 6 we have and these values being represented by a surface of simple arc section in Fig. 20. The above defines the theoretical camber for the mid-section of the aerofoil, but under the conditions assumed this will not be constant; n is constant, this being the basis on which we are working, but the added camber required to deal with the cyclic component will depend upon the local strength of the vortex, which varies from point to point. The distribution is actually that of the velocity potential field of Fig. 16, and from this it appears that the correct grading for the addendum camber as represented by the angle a will be proportional to the ordinates of a semi-circle whose base is the span of the aerofoil.

418

THE FLYING MACHINE.

207

§ 13. CONCLUSION. It has not been found possible in the present paper to do more than give an outline of the theory of sustentation, with sufficient examples and references to practice and experiment to illustrate the importance of the theoretical aspect of the subject as bearing1 on the experimental treatment; the latter has hitherto been dealt with almost without regard to considerations of theory, and has degenerated into empiricism pure and simple. For example, mouldings are made to "stock sections" and cut off to different lengths to determine the effect of change of aspect ratio; or again, aerofoils of some definite camber are tested under unfair conditions, i.e., those to which their particular camber is ill adapted, and inferences are drawn as to best values of pressure constant, etc., which are in no sense justified. The attempt is obviously that one condition only shall be changed at a time* but owing to lack of consideration the result is exactly the contrary. We arc acquainted with many parallels in engineering practice. If it were desired to find the speed of maximum torque of an internal combustion engine, for example, it would of course be absurd to religiously change the one condition of speed; the ignition timing must follow the speed, and the carburettor may need several trial adjustments. Scores of similar examples could be quoted. The latter portion of the paper dealing with the quantitative treatment on vortex theory only purports to be a sketch of a very promising development, and may be taken as an indication merely of the lines on which the author is at present at work. So far, the cases taken are those for which the main material for solution happens to be at hand; it must not be imagined that the examples of vortex motion utilised have been selected for any other reason,f also the results have in some cases been given without the whole of the reasoning or proofs on which they rest. Before the method can be considered complete, we require to possess the means of specifying our vortex distribution at will, and of * An axiom instilled into the mind of the elementary science student, representing an ideal which can be acted upon only when dealing with independent variables. + Compare Appendix III.

419

208

THE INSTITUTION OF AUTOMOBILE ENGINEERS.

solving the same to give the primary and secondary camber for all points along the span of the aerofoil; the author is at present engaged in endeavouring to find a practical solution of this general problem in such form as will give the designer full control over his product—the difficulties do not appear to be insurmountable.

Please note: Appendices I to VI have been intentionally omitted, as has the discussion.

420

William Henry Bragg FRS (1862-1942) The founder of the science of X-ray crystallography was a Cumberlandborn, Cambridge-educated mathematician who moved into the realm of physics on appointment to the University of Adelaide in 1886. William Konrad Rontgen's discovery of X-rays was announced during Bragg's Australian tenure. He began to perform simple experiments with X-ray tubes and ponder the nature of the new phenomenon. Bragg moved successively to Leeds University (1909) and to University College, London (1915) developing lines of research as he went. First, he looked at the rates of absorption and scattering of radium-source particles in solid and gaseous matter, becoming an expert in studying the ionisation of X-rays. Results obtained by Max von Laue in 1912, outlining X-ray diffraction effects observed in crystals, changed Bragg's tack. With the aid of his son, William Lawrence Bragg, von Laue's methodology was improved and used to map the exact atomic structure of crystalline substances, such as diamond. The Braggs were jointly awarded a Nobel Prize in 1915. The wartime emergency caused by the success of the submarine saw Bragg engaged in developing countermeasures, notably the hydrophone, projects which had already employed the skills of Ernest Rutherford. On returning to his primary scientific work he took up the Royal Institution's Fullerian Professorship, directing the work of the Davy-Faraday Laboratory. From this position his techniques were widely disseminated. The crystal structure method became used for the study of organic material and (crucially for engineers) metals and minerals. War broke out again during his Presidency of the Royal Society (1935-1940) and, despite his advanced years, Bragg chaired the key Cabinet Committee on Scientific Policy.

421

THE FOURTEENTH

THOMAS HAWKSLEY LECTURE. THE APPLICATION OF X-KAYS TO THE STUDY OF THE CRYSTALLINE STRUCTURE OF MATERIALS. BY SIR WILLIAM H. BRAGG, K.B.E., F.R.S. Friday, 4:th November 1927. Introduction.—The examination of the structure of crystals by means of X-rays first became a possibility some fifteen years ago. Since that time the subject has been advanced by the work of a number of investigators who have been drawn to it in the hope that the new point of view might open up wide and important fields of knowledge which had hitherto been beyond the reach of inquiry. The resulting accessions to knowledge are described in several excellent and comprehensive treatises : which, however, can with difficulty keep pace with the rapid development of the subject because the method finds application in almost every branch of science. As the subject has a special bearing on the structures employed by mechanical engineers, I have thought that you might be willing to listen for a short time to an explanation of the principles on which it is based. Such statements have of course been made already, and are to be found in published form: nevertheless I hope that I may be able to put the matter before you in a way which will lead simply and directly to the explanation of some of the most recent investigations on engineering materials, particularly metals and alloys. [THE I.MECH.E.] 2 o

423

752

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

NOV. 1927.

The Fine Structure of Materials.—The properties of any material depend, as we all know, upon its fine structure. No better example of the fact could be given than the growth of metallurgical science under the influence of the microscope. The toughness, hardness, brittleness, and other qualities of, let us say, a sample of steel vary enormously with the condition to which the steel is brought by work and heat. The microscope shows that the changes are marked by striking alterations in the appearance of the steel. The grains that are brought out by suitable etching, show an infinite variety of forms which the expert has learnt to associate with a corresponding variety of qualities in the material. The fine structure thus displayed is obviously in close relation to quality. But below the structure which the microscope makes clear is a much finer structure which no eye can see and no microscope will ever help the eye to see. In this realm of small things the lens of the microscope is constitutionally incapable of acting as an optical instrument. Yet it is of the highest importance that the ultra-fine structure should be recognized and known ; it is the real foundation of the properties of the body; that which the microscope reveals is merely a consequence of the other. The minutest grain of the microscopic photograph has a structure, and effective properties which depend upon it. How shall we reach this finer view, and understand these more fundamental relations '? The use of X-rays.—We must in the first place abandon the light waves that serve us in the study of grosser details, and use ether waves more fitted for the work. The lengths of light waves have only a small range of variation, but even within that range we have some guidance to the course we must take. The finer the details to be observed, the shorter the waves that must be employed to be affected by them and to make them visible. In modern microscopy the shortest possible wavelengths, violet and ultra-violet, are employed in order to increase the power of recognizing small things. When, on the other hand, we wish to avoid the scattering of light we use long waves, as when red flares are employed to penetrate a fog, and as infra-red rays have been used to obtain photographs through a scattering atmosphere. But we wish to make a step hundreds of times as great as that between the two ends of the optical range ; and we find the requisite quality in the extremely short ether waves known as the X-rays. It is an extraordinary fact that the X-rays which are emitted by apparatus such as we can conveniently construct and use, are ether 424

Nov. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

753

waves of just the right length for our purpose. Had they been ten times longer or ten times shorter we should have been in very different case. The wavelengths are in fact of the same order as, but rather shorter than the dimensions of the fine structure we wish to examine, and this is exactly what we want. Whenever radiation —ether waves—is to be used to show the details of an object, it is essential that the waves shall not be larger than the details of the object examined, nor indeed as large. Here then is our illumination of the right quality. How do we apply it ? We are anxious to study those fine details on which the properties of substances depend, we have an ether radiation of the proper quality, what is the next step to be taken ? The Regularity of Natural Structures.—We take advantage of a certain strong tendency in nature for regularity of arrangement. The atoms of which a metal is composed are always found to be more or less in ordered array. One and the same pattern of arrangement may not, indeed, extend through the whole specimen ; there may be breaks, on each side of which are arrangements that are similar but do not dovetail into each other. But whether the boundaries of any one array include a larger or smaller mass, regularity is always there. A portion throughout which the regularity is continuous is a single crystal; usually a mass of metal is an aggregate of minute crystals orientated in different directions. This tendency to regularity is extraordinarily persistent. Whenever a substance solidifies, whether from liquid or vapour, the atoms or molecules tend to arrange themselves in some pattern proper to that substance. The longer the time given and the less disturbing the circumstances the more complete and exact is the arrangement. Even when a substance has been so rapidly hurried into the solid state that the molecules have not had time to arrange themselves correctly, the unfinished process is only hindered and not arrested entirely ; a slow crystallization continues to take place. The Diffraction Phenomenon.—Finally, and this is what makes the new analysis possible, whenever there are waves and a regularity of spacing which is of the same order of magnitude as the length of the waves, there can result a certain measurable phenomenon called diffraction. The spacings of the striations of mother of pearl are of the same order as the lengths of light waves, hence the diffraction colours that play upon its surface. The diffraction effects observed on the grating spectrometer allow us to compare the wavelengths of light with the distance between two lines of the grating. In the 425

754:

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

same way there will certainly be some kind of diffraction effect when X-rays traverse a body in which there is regularity of arrangement. The usual wavelength of an X-ray is somewhere about the hundredmillionth of a centimetre, an Angstrom Unit, so called. The dimensions of the pattern on which the atoms and molecules of the solid are arranged are generally a few such units in magnitude ; for example, the distance between the centres of two neighbouring atoms in a piece of aluminium is 2.86 A.U., and in a-iron 247 A.U., and in g-iron 2.57 A.U. In these three cases the unit of pattern is very simple and consists of one atom only. In the majority of substances the unit is more complicated, and may contain ten, fifty, a hundred atoms. The regularity consists in the even spacing between a point representing one unit, and a point similarly representative of the next. FIG. 1.

FIG. 2.

We have now got so far as to see that since there is regularity of spacing in a solid, and since the X-rays are of the proper length in comparison with that regularity, there ought to be diffraction effects of some kind; and if we can find them we ought to be able to compare the periodicity of the spacing in the solid with the wavelengths in the X-rays. If we can go on to find how that regularity depends on direction in the solid we should be able to lay out the details of the pattern. To take a very simple example, if we were told that the distance between the rows of vines in a vineyard was ten feet in a north and south direction and the same east and west, but rather more than fourteen north-east and southwest, it would be easy to find the pattern of the planting. Let us then try to see what sort of diffraction effect we might look for. In doing this we must make use of the old principle of wave interference which Young in England and Fresnel in France 426

Nov. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

755

developed in the earlier years of the nineteenth century. The application to our present problem is perhaps even simpler than in the first instance of its use, and it is worth while to grasp it, since everything else follows from it. Let us imagine a train of waves in the ether as in the Figure, advancing and passing by an object that can scatter them; if we like to take the simplest case we can think of, we may say that the object is a single electron. A set of spherical waves will spread out in all directions in space concentric with the object which is their origin. Just so a set of concentric ripples will spread out on the surface of a pond if a train of waves impinges on a post projecting above the surface (Fig. 1). If we replace the single electron by some group of electrons, atoms, and molecules, then in the immediate neighbourhood of that group there will be a jumble of interfering wavelets, but at some distance from the group the irregularities will have melted into each other and again there will be a system of spherical ether waves of which the group occupies the centre. We can observe this on the pond when the single post is replaced by a group of posts, which it should be observed need not be of the same size nor be spaced in any regular way (Fig. 2). The essential fact is that whether the scattering body be a single electron which we may look on as a point source, or a complicated group such as might be found within the compass of a single unit of pattern in the solid body, a train of secondary spherical waves will spread away into space from the point or the group. In the case of the group we must not look for a clear development of the spherical wave very close to its origin, but at a certain distance it will be well formed, and this is all that matters. The point and the complicated group give eventually similar results ; and so we will for the present talk only of point sources, or points. What now will happen if there is a set of points arranged in any regular fashion or of groups which can be represented by points because they are all exactly alike ? Here come in the principle of interference and its consequences. The application is made much easier by the fact that we have all seen it in operation under many different kinds of circumstances, and can therefore picture its extension to this case where imagination must take the place of actual vision. Suppose, for example, we have a row of posts projecting above the surface of the pond and let there be regularity to this extent that the posts are all in one line. Then each post becomes the centre of a train of spreading ripples, and what is to be noted is that these different trains combine to make linear ripples which recede as in 427

756

APPLICATION OF X-EAYS TO STUDY OP MATERIALS.

Nov. 1927.

Fig. 3, making a reflected train of ripples ; the rest of the ripples from each post spread away in the usual fashion, dying out rapidly as they spread over a wider front. The combined wave front, on the other hand, is not only strong relative to the rest but persists for a far longer time ; if the posts are many and are not too wide apart the reflected ripples are the obvious parts of the phenomenon. In fact the effect is often seen in a harbour where a row of closely planted piles reflects the oncoming waves. So also the air waves of sound are reflected by a row of palings. This is the one fundamental principle of the new analysis, extended to the case of three dimensions because we are now dealing with a number of points arranged regularly in space according to some pattern, instead of a row of FIG. 3.—Sketch of an experiment performed at a Christmas Lecture at the Royal Institution, where ripples in a shallow tank were reflected by a row of nails standing upright in water.

posts in a straight line. Let us consider this rather more complicated but exactly analogous case. Suppose that the dots in Fig. 4a are some of the representative points in the solid : each point stands for one unit of pattern. The points marked 1, 2, 3 are a few of the points lying in a plane perpendicular to the plane of the paper: we suppose ourselves to be looking edgeways at this and parallel planes, in some of which also a few points are marked, such as 4, 5, 6, 7 in one plane, 8, 9, 10, 11, 12 in the next, and so on. On account of the regularity of the pattern it must be possible to consider the points as arranged on planes in this way, just as in the two-dimensional problem of a vineyard it is possible to think of the vines as in rows, in various ways. The points are not all in the plane of the paper, because if 428

NOV. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

757

they were it would mean that certain point-bearing planes which we are regarding edgeways are at right angles to other point-bearing planes : and this would presume a certain rectangular character in the solid which it does not necessarily possess. The dots in the Figure maybe projections upon the paper of the representative points, but this does not affect what follows. The regularity with which the dots are drawn is intended to be in conformity with the general regularity of the spacing. Furthermore, only a few points are shown, but these must be taken to represent the enormous number which come into action when the X-rays sweep through the crystal.

FIG. 4a.

Suppose that I, I1 I2, I3 is a train of incident waves ; if they are X-rays they will sweep through the material as waves would sweep over a sheet of water where posts project above the surface. From each of the numbered points concentric ripples spread away into space ; they will now be spherical, not circular as in the analogous case upon the pond. But exactly the same results follow; and combination of the ripples may lead to the formation of what we may call a reflected train of waves. Suppose R, R1 R2, R3 is such a reflected train of waves ; let us take I in the incident train and R in the reflected train as reference lines, and let us suppose also that the waves are perpendicular to the plane of the Figure. The condition that combination between the scattered waves from the points 1, 2, 3 shall take place as in the analogous case of the points in the lake is that the sum of the perpendiculars from point 1 upon 429

758

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

the lines I and R is equal to the same sum in the case of points 2 and 3. When this is the case, the straight line through 1, 2, 3 obviously makes equal angles with I and R ; this is in fact a particular case of the familiar law of reflection. When this equality is satisfied, the sum mentioned is the same for all the points on the plane to which, in the figure, 1, 2, 3 belong, as already explained. FIG. 4b.

But, is there a possible combination which will take in not only points 1, 2, 3, but all the other points in the Figure ? If so, the reflection will surely be far greater. Now it is clear that the sum of the perpendiculars from No. 4 is greater than the sum from No. 1, but if that difference is a whole wavelength, the combination of No. 4 with Nos. 1, 2, 3 will take place equally well. For what difference can there be between two successive waves of a train ? FIG. 4c.

And if 4 enters into the alliance so also will 5, 6, and 7, because the line 4, 5, 6, 7 is equally inclined to I and R. Furthermore, owing to the regularity of the arrangement of the points in the Figure, 8, 9, 10, 11, and 12 will join in because their contributions are exactly two wavelengths behind those from 1, 2, 3 ; and 13, 14, 15, 16 because they are three in arrear. In fact all the points we have put in the Figure and all those we have not put, will combine ; and 430

Nov. 1927. APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

759

though the contribution of each one may be exceedingly small, the total effect is appreciable. The fine pencil of X-rays which we employ in the actual experiment sweeps over billions of such points in the material, and each makes a contribution to the reflected rays. Can we arrange that the sum of the perpendiculars on I and R from No. 4 shall be exactly one wavelength greater than the corresponding sum for No. 1 ? Surely ; for, since we can place the collection of points at a different angle to the incident train, we can make the difference in the two sums less as in Fig. 4b, or greater as in Fig. 4c, and we can therefore adjust the difference to the desired equality. When the adjustment is made there is combination of all the points in the solid, and an observable reflection results ;

FIG. 5.

when the adjustment is absent, or indeed is only slightly imperfect, there is no combination, and as we find in practice, no reflection at all. The angle between the direction of the incident rays and 1, 2, 3, and therefore that between the incident and reflected rays, must have a definite value. It will be observed that the combination effect is only possible because there is regularity in the spacing of the rows 1, 2, 3 ; 4, 5, 6, 7, etc. If then a pencil of X-rays having one definite wavelength be represented in Fig. 5 by XC, and if a crystal, say of aluminium, is placed at C so that the X-rays sweep through or over it, it will be possible to turn the crystal into various positions until suddenly a reflected ray CR flashes out; when that occurs XC makes as we have seen a certain angle with CR, an angle depending as regards size upon the magnitude of a certain spacing in the crystal. 431

760

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

NOV. 1927.

If a photographic plate is placed as in the Figure, two spots will be impressed upon it; one very strong spot P made by the primary beam of X-rays and the other by the reflected beam at R. The " Powder " Photograph.—Suppose the crystal were turned round about the axis XC; the conditions for combination would not be affected, and if the X-rays were acting all the time the spot R would be extended so as to form a circle on the plate with centre P and radius PR. To obtain this result we must first set the crystal into a reflecting position and then rotate it about an axis coinciding accurately with the line of direction of the incident rays. Now suppose that we put at C not a single crystal but a small powdered mass of the substance. The vast majority of the crystalline fragments will not be in a position to reflect at all; but there will be some that are correctly adjusted, and these will produce spots on the plate at a distance PR from the centre. And as among the reflecting crystals there will be some that take up all the possible orientations of the single crystals revolving round XC the result will be the formation of a ring on the photographic plate. This effect is easily obtained. In fact when we send a fine pencil of X-rays of definite wavelength through a specimen of powdered aluminium we find not only one ring, but several rings upon the plate. If we reconsider Fig. 4a we find a ready explanation of the multiplicity. When reflection was successful the line of points 1, 2, 3 made equal angles with I and R. But why should 1, 2, 3 be the only line so favoured ? There are many other lines of points in the diagram, e.g. 1, 4, 8, 13 ; 2, 5, 9, 14; or 1, 5, 10, 16; 2, 6, 11, 17. We might expect to find other successful positions of the crystal in which these or other sets might successively be equally inclined to I and R. In each case there would be a special value for the angle between incident and reflected rays. In a mass of powdered crystal these positions would also be represented by various individual crystals, and rings would appear on the plate, each ring corresponding to the case of one particular set of point-bearing planes. A photograph of aluminium obtained in this way is given in Fig. 6. If any other crystal in a state of powder is treated in the same way we find on the plate a set of rings which differs from that due to aluminium. Every substance has its own characteristic arrangement, because there are varying patterns according to which the points are arranged. Some of the patterns are particularly simple. The atoms of aluminium are packed together like a pile of cannon balls. As is easily discovered by trial there are two ways of packing spheres 432

NOV. 1927.

APPLICATION OP X-RAYS TO STUDY OP MATERIALS.

761

closely together, one in which there is a repeat every third layer and one every second ; aluminium follows the former. The unit of pattern is very simple in this case, merely a single atom ; so there are as many representative points as there are atoms. On this assumption we can calculate the diameters of rings which are characteristic of aluminium. I do not propose to do this now. since my purpose is to explain a method and not work out its details. The calculated system agrees perfectly with the experimental result shown in Fig. 6. Whenever this system is observed aluminium is present, and no other body gives the same result. Aluminium is not the only metal of which the structure is that of closely packed FIG. 6.—Powder Photograph of Aluminium.

spheres ; at least fifteen others are built on the same plan, for example, calcium, g-iron, nickel, copper, silver, gold, lead. In all these cases the rings are similar as regards their relative sizes, but the absolute sizes vary, because the atoms are not of the same diameter. If in Fig 4a we supposed the whole arrangement of points to be reduced in scale, and still required that the sum of the perpendiculars from a point in the second row, 4, 5, 6 should be one wavelength greater than the sum in the first row, 1, 2, 3, it is clear that we must increase the angle at which the rays are inclined to these rows, as in Fig. 4c, and so the angle between XC and CR, Fig. 5, must be increased. If the atoms are smaller and closer together the whole set of rings expands correspondingly. 433

762

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

NOV. 1927.

The Structure of Iron.—When we have a structure of different design the whole character of the system of rings changes. A good example is shown in Fig. 7 which is due to Westgren and Phragmen. These pictures are portions of the complete rings of a photograph such as that of Fig. 6. The photographs of a-, b-, and d-iron are quite different from that of g-iron, which, as I have already said, is like that of aluminium. The disposition of the lines in these three shows that the atoms are arranged in a different way. If we put down a layer of spheres in a square pattern and on them a second layer as represented by the dotted circles of Fig. 8, and on those again another layer exactly overlying the spheres in the first layer, and so on, we have the structure of ordinary iron. FIG. 7.—Powder Photograph of Iron at different Phragmen).

temperatures (Westgren and

Although iron undergoes a change in magnetic properties at 800° C. there is no change in its crystalline structure. At 900° there is a complete alteration in the design, the iron remaining non-magnetic. At 1,400° the structure reverts to its old form and the iron is magnetic again. Some other metals, notably manganese, thallium and chromium, show similar changes from one form of structure to another. Some metals like antimony, bismuth, and tin show much more complicated structures. Tin having one form of structure is brittle ; having another form, it is not. A Ring System Characteristic of Every Substance.—Not only does every different chemical element have, so to speak, its own signature, 434

Nov. 1927. • APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

763

but every compound. Iron has its own ring pattern, carbon another ; iron carbide has a pattern quite different again. A mechanical mixture of powdered nickel and powdered aluminium gives the systems of the two metals on the same plate ; but the compound Ni.Al gives an entirely different picture. Every one of the hundreds of thousands of organic compounds has its characteristic X-ray photograph, though so many of them consist only of carbon, oxygen, and hydrogen. It must also be observed that impurities do not affect the photograph ; they may add other lines, if they are present in sufficient quantity, but they do not affect the position or sharpness or intensity of the lines proper. FIG. 8.—Structure of a-iron.

The opportunities thus offered have tempted observers to follow the nature of alloys through all their various ranges of composition. The technique is difficult, and interpretation is hindered by present inexperience, but remarkable results have already been obtained. The work of Owen and Preston in England and by Jette, Westgren and Phragmen in Sweden on the copper-aluminium and the copperzinc alloys has been set out by Clark * in two very instructive tables in which the agreements and disagreements between the two sets of observers show at the same time how much has already been made possible and how much remains to be done. The X-rays have already been applied to the study of the * " Applied X-Rays," pages 207-208.

435

764

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

extremely complicated system of iron-carbon alloys ; a survey of what has been accomplished so far is given by Clark in the book to which I have already referred. There is one very remarkable feature in certain of the metal alloys, namely, the very large number of atoms in the unit of pattern. As an example I may take the case of d-bronze, which has been examined by Berrial in the Davy Faraday Laboratories of the Royal Institution. The unit cell contains more than a hundred atoms ; clearly the atoms of copper and tin have definite and unyielding characteristics ; a careful adjustment of a number of each kind is FIG. 9.—Drawn Aluminium Wire (Clark}. X-rays perpendicular to the ivire. The photograph on the right was obtained with the aid of a special filter which made the pencil of rays more homogeneous.

required to complete one unit before starting another. Even a pure metal sometimes shows a similar character ; the ordinary form of manganese has 56 atoms in the cubic cell, according to Westgren and Phragmen. The ring pictures which I have tried to explain are obtained when a powdered mass of the crystalline material is traversed by the X-rays. The completeness of the ring is ensured by the complete disorder of the fragments of crystal; they must be orientated in every possible way. If this is not the case the completeness of the ring suffers ; only parts of the circles appear on the photograph. The Effect of Treatment on the Form of the Rings.—Now if the metal experiences any mechanical treatment, the particles are put into partial order. Therefore any such treatment will alter the 436

Nov. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

765

appearance of the photograph because the rings can no longer be complete. In Figs. 9 and 10, are shown photographs of rolled sheet aluminium, and drawn aluminium wire, which are to be compared

FIG. 10.-—Rolled Aluminium Sheet: X-rays making angle of 70° with sheet. The form changes irith the angle of inclination.

with the simple complete ring system of Fig. 6. The rings are still there, but they have shrunk into spots on the circumferences. The faint radial lines are due to inhomogeneities in the X-rays which FIG. 11.—Commercial Rolled Copper Sheet (Clark"). On the left, X-rays perpendicular to the direction of rolling : on the right, parallel.

were not apparent in the simple picture because they were diffused all round the compass. Consider also the two photographs of Fig. 11 ; they are due to commercial rolled copper sheet. In one of them the rings are much more complete than in the other, because the X-rays have been 437

766

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

passed through the specimen in the direction of the rolling and therefore there is no singularity in orientation round that direction ; unless perhaps in respect to the geometry of the sheet, and this seems to have had little effect. But in the other photograph the rays have been passed through at right angles to the line of rolling ; the latter is parallel to a certain direction upon the plate. It is not possible in the time at our disposal to make any study of these effects, but it will be clear that the photograph not only shows the result of mechanical treatment, but also varies with the nature of the treatment and its direction in the specimen. The skilled observer learns to interpret the effects observed ; and much useful FIG. 12.—Rotation photograph of a Crystal of Benzoic Acid (Muller).

information has already been gained in this way. I propose to conclude this Lecture by a short statement on this branch of the subject. But before I do that, I would like to continue my explanation of the method, by adding a little about the origin of the short arcs or spots in Figs. 9. 10 and 11. The Rotation Photograph.—A powder gives complete rings. If the powder is combed out by any treatment the rings shrink into short arcs. We may understand this better by beginning at the other end. Let us take the case of a single crystal, subjected to monochromatic X-rays. We have seen that unless it is correctly held there is no reflected ray ; it must be disposed so that some plane of points in the crystal makes the proper angle with the X-rays and then the reflection flashes out. There are several such positions, in each of which one of the many planes which can be drawn to include the representative points in the crystal is prominently involved. But it is not likely that any chance holding of the crystal 438

Nov. 1927. APPLICATION OF X-RAYS TO STUDY OP MATERIALS.

767

will give a reflection. Suppose, however, that the crystal is made to revolve about an axis perpendicular to the X-rays. The inclination of the rays to the various planes will change continuously over wide ranges during the revolution, and reflection after reflection will flash out as each set of planes, and therefore each reflected ray makes its own characteristic angle with the incident beam. The resulting photograph will show a collection of spots, and we may well expect that the ordered array in the crystal will give order to the photographs. An example of such a photograph is given in Fig. 12. The axis of revolution and the line perpendicular thereto are naturally lines of symmetry, but in addition to that it will be observed that the spots are displayed on various lines. It is not FIG. 13.—Asbestos (Astbury). A bundle of fibres held at rest in a direction perpendicular to the X-rays.

difficult to explain the form of these lines, but it is not necessary to our present purpose. The " Fibre " Photograph.—Suppose now that instead of taking a single crystal and rotating it we were to take a number of minute crystals arranged so that some one common direction in them all was parallel to a certain straight line and perpendicular to the X-ray beam, but otherwise their disposition was a matter of chance. The substance might be said to possess a " fibre " structure. We should then expect the same picture as when we rotated the single crystal. In the first case one crystal occupied all possible positions successively; in the second case different crystals occupied all possible positions at the same time. A good example is given by a set of asbestos fibres, each of which is crystalline. One definite direction in each crystal is parallel to the length of the fibre and the 2 p 439

768

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

bundle of fibres is placed perpendicularly to the X-rays. Thus we get in Fig. 13 a picture which shows the same general characteristics as Fig. 12 ; it differs only in the way in which every X-ray picture FIG. Ua.—Cast Steel (Clark).

FIG. 146.—Cast Steel annealed (Clark).

FIG.

14c.—Cast Steel thoroughly annealed (Clark).

is characteristic of the substance and of the way in which it was presented to the X-rays. It will now be clear that the photographs of mechanically treated 440

Nov. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

769

metals have an appearance somewhat similar to that of asbestos. In them also, we must conclude, the separate crystals are arranged so that some one direction—in the case of the actual photographs given—is perpendicular to the X-rays. When the circles shrink into spots or short arcs it is because the substances show a certain amount of " fibre." The picture is symmetrical, as in the illustrations given, if the angle of inclination is a right angle. Now the extent of development of fibre structure, the magnitude of the crystals, and their orientation to the direction in which strain is to be expected are extremely important things. Fibre structure does not alone determine the strength and other properties of a specimen, but it is one of the most important of the influences that do so. It is one of the most obvious of the applications of X-rays to observe the changes in fibre structure due to the preparation of FIG. 15.—Lines of equal strain plotted from X-ray data (Clark).

a specimen, and its subsequent mechanical treatment and annealing. To make this point clear I will draw on the material collected by Clark in the book already quoted ; one or two of his many illustrations will give substance to what I am trying to say. Various examples of Fibre Structure.—Here are first of all three photographs of cast steel which show the removal of fibre structure by annealing ; very persistent annealing is required to remove the radial effect which is due to the fibre structure (Figs. 14a, 14b, and 14c). Here is a diagram (Fig. 15) showing the plotting by means of X-rays of lines of equal strain in a specimen which is, according to the usual tests, free from holes, cracks, inclusions, and the like deficiencies. Possibly this Figure shows that each part of the specimen is so strained as to meet most efficiently the strain it is to bear : possibly not, 441

770

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

In Fig. 16 the effects on a pulled bar are shown; the perfect FIG. 16.—Structure in a Cast-Steel Bar pulled at 1,200° F. (Clark).

FIG. 17, Satisfactory forming steel.

Unsatisfactory forming steel. (Clark).

rings in (d) are due to the fact that the direction of the X-rays employed in making the photograph is parallel to the line of pull; 442

Nov. 1927. APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

771

FIG. 18a. Transformer Steel. Hysteresis loss 0-8636 watts

/lb. (Clark).

Transformer Steel. Hysteresis loss 0.6931 watts /lb.

FIG. 18b. Transformer Steel. Hysteresis loss 0-5535 watts /lb.

in (c) the two directions are perpendicular to one another. There is a strong fibre structure. The photographs (a) and (6) relate to a part of the bar where the pull has not acted so as to produce fibre. Fig. 17 shows the difference between two steels for stamping. 2 P 2 443

772

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

Nov. 1927.

There is a fibre in the unsatisfactory steel which gives it different properties in different directions, so that it cannot respond uniformly to the stresses of the manufacturing operation. Fig. 18 shows pictures of transformer steels of varying quality. In this method heterogeneous X-rays are employed : in the methods which I have been describing in some detail the rays have been of one single wavelength, or at least of one or two wavelengths easily separable in their effects. In this latter case the crystal had to be put into special positions before reflections occurred ; in the former, Laue's, the crystal gives reflections in any position, because whatever inclination a set of planes presents to the incident X-rays there is some wavelength in the wide range of wavelengths which is at the correct angle for reflection. A single crystal gives a series of spots, a mass of crystals gives only a hazy cloud as these Figures illustrate. FIG. 19.—Duralumin (Clark).

And in Fig. 19 are two pictures of duralumin, one resistant to intercrystalline corrosion, the other not. In the latter there is greater internal strain; and a recrystallization process has set in. The examples of fibre structure are due to Clark. In Fig. 20 are shown the results of rolling magnesium. These photographs are due to Rosenhain. In this case flake crystals of magnesium are caused by the rolling to lie parallel to the surface of the sheet; their distribution with regard to the normal to the sheet is quite irregular. The photographs of Fig. 21 show the appearance and disappearance of certain crystal forms during the various stages of treatment of tungsten steel which are accompanied by changes in magnetic properties. 444

NOV. 1927.

APPLICATION OF X-RAYS TO STUDY OF MATERIALS.

773

Finally, let me instance the use that has been made of the new methods in the investigation of strain phenomena, and of planes of slip. Valuable work in this country has been done by Carpenter and Elam, Taylor, Miiller, Owen and Preston, and others. Again, I must call attention to the use of X-ray analysis in the determination of the way in which carbon enters the iron structure to form steel : in the investigation of solid solutions and indeed in a great variety of metallurgical problems. But I must perforce leave unmentioned a large number of very interesting applications to questions arising in modern industry. Summary.—I have tried to sketch the fundamental principles of a new method of examining the structures of nature. Its FIG. 20.—Rolled Magnesium (Rosenhain).

importance lies in its application to a range of dimensions where the properties of the structures are mainly formed ; dimensions somewhat larger than those of atoms and molecules, much smaller than those which are observed under the microscope. Since both method and application are so new, we might legitimately say that we have here a new science. It is a science which has links with every other ; for it is not only in the case of metals, though I have referred to metals in the main, that the X-rays show us processes of which otherwise we should know nothing. All the other substances which the engineer handles, wood and stone, bricks and cements, conductors and insulators, lubricating oils, and a host of other things, have structures on which their properties depend. The same may be said of the cottons, wools, and other fibres of the textile worker ; of the 445

FIG. 21.—5 per cent Tungsten Steel (Shearer),

Nov. 1927.

APPLICATION

OF X-RAYS TO STUDY OF MATERIALS.

775

living substances with which the physiologist deals ; of the botanist's plants, the geologist's minerals, and indeed all the solid substances that are handled by man. It would be false modesty to refrain from claiming the power which the X-rays give us of entering upon this huge field. We must pay our debt to modesty when we admit our present inexperience in handling our new power, our lack of understanding all that our observations tell us, our want of realization of what we may now attempt, and the deficiencies in our technique. However, the number of workers is already so great and grows with such increasing rapidity that not only will the foundation of our new science be rapidly strengthened and broadened, but also the superstructure will rise quickly. I am glad to say that in this country an official entry into the new field has been made by the foundation of a department at the National Physical Laboratory which is to take as its special concern the application of X-ray methods to industrial problems. The private laboratories of this country which have been interested in the subject should be mainly devoted to the fundamental principles and to applications in pure science. It would indeed be a mistake to concentrate all our energies on industrial applications, while we have still to map out the main features of the subject. It will happen here, as it has happened with every new increase of knowledge, that many of the first suggested applications will fall short of what is hoped, and that the great successes will come from directions unexpected. Moreover, the benefit which will come from the widening of outlook and the general strengthening of our knowledge of structure will be of greater importance than the special and limited applications which we think we can put in hand at once, useful and fascinating as these may be. Therefore I hope you will wish success both to the new venture at the National Physical Laboratory and to the more general investigations that are now in progress elsewhere.

Vote of Thanks. Dr. H. S. HELE-SHAW, F.R.S. (Past-President), proposed a vote of thanks to Sir William Bragg for his Lecture, which was seconded by Mr. LOUGHNAN PENDRED (Vice-President), and carried with acclamation.

447

This page intentionally left blank

Sir Arthur Stanley Eddington FRS (1882-1944) One of the most gifted scientific minds in the crowded field of 20th Century astrophysics, A S Eddington was born in Kendal, Westmorland, on 28th December 1882. Of Quaker stock, Eddington was educated privately, progressing to Owen's College, Manchester at the age of 15, and from there to Trinity College, Cambridge. His first professional post was as assistant to the Astronomer Royal at Greenwich Observatory in succession to F W Dyson. In 1913 Eddington became Plumian Professor of Astronomy at Cambridge, publishing his first book in that year - Stellar movements and the structure of the universe. The book opened a new field in astronomy, that of stellar dynamics. In 1916, Eddington's theoretical work received a boost from an unexpected quarter, when he was sent a paper on General Relativity written by the then little-known Albert Einstein. Eddington reported on the paper to the Physical Society, adding much original work on electromagnetic phenomena, and expanding his thinking into The mathematical theory of relativity (1923). Einstein himself remarked that Eddington was one of the few scientists who truly comprehended his ideas. From 1916, Eddington began research on the mechanisms of stars, and it is to this phase of his life that the paper published in IMechE's Proceedings belongs. His summary of these years, The internal combustion of the stars (1926), was immediately recognised as a standard work. In the later Hawksley lecture Eddington states explicitly that he approached the topics as 'an engineering problem viewed from an engineering standpoint'. Eddington's later life was engaged in various problems of pure mathematics including calculations of constants of nature and reconciling aspects of relativity with quantum mechanics. This fundamental work would later see him described as 'the modern Archimedes'. His inclusion here is therefore appropriate, updating the original Archimedes who appears on the Institution's coat of arms.

449

FEB.

1929.

43

THE FIFTEENTH

THOMAS HAWKSLEY

LECTURE.

ENGINEERING PRINCIPLES IN THE MACHINERY OF THE STARS. BY PROFESSOR A. S. EDDINGTON, M.A., D.Sc., LL.D., F.R.S.

Astronomy and Engineering.—The Council of The Institution of Mechanical Engineers has taken an unusual step in appointing an astronomer to deliver the Thomas Hawksley Lecture. Perhaps the closest relations between our respective sciences would be found in the problems of construction of telescopes, particularly in the monster telescopes that are now so essential to astronomical research. The greatest telescope at present existing—the 100-inch reflector at Mount Wilson—is a marvel of engineering. We have recently learnt that a further advance is planned, and it is expected that in a few years a 200-inch reflector will be at work. I need not say how the utmost skill of the engineer will be concerned in the design and delicate mounting of this huge piece of scientific apparatus with its tube, I suppose, not less than 80 feet long. On my side I can say with the utmost confidence that the new instrument will bring rapid and perhaps sensational advance of astronomical knowledge. It is true that problems are confronting us for which a 200-inch or even a 500-inch mirror will be insufficient; but there is much urgent work that lies just outside the capacity of present instruments, and the increase of power will be a most momentous accession to our means of research. For these enterprises an astronomer ought also to be a skilful [THE I.MECH.E.] 451

44

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

engineer, but I am not qualified in that way. I come as one ignorant of the subject who can only accept with gratitude these gifts of engineering to our science. I ought indeed to be rather a pariah among you, because much of my work has been devoted to that side of modern physics which is bent on destroying the mechanical basis of the universe and showing that ether and matter are not constructed as an engineer would have devised but rest rather on those strange conceptions associated with relativity and the quantum theory. We have been busy wrecking mechanical models and driving out the engineer. Our predecessors thought of gravitation as a stress carried by an ether of rigidity greater than steel; if you inquire about gravitation now we put you off with talk about curvature of space. The electron used to have at least something concrete about it; a year or two ago it was set spinning, and we conjured up visions of fly-wheels; but now we have got rid of the suspicion of mechanical contrivance and deal with it by the mathematical device of matrices. In short Dame Nature is no longer a bustling engineer ; she is a serene mathematician. But however unmechanical her methods, Nature does contrive to produce engineering work on the grand scale, and my task must be to show you something of these achievements. If any of you are concerned with the problems of power stations, I will take you to see the great power stations of the universe. Some may be concerned with the calculation of stresses in structures; I have to speak of the calculation of stresses in the stellar structure where the pressure exceeds a thousand million atmospheres. Some may be concerned with properties of material; we shall find material with strange properties transcending anything available to the terrestrial engineer. Many are concerned with throbbing machinery ; we shall find that some stars are in steady pulsation and invite conjecture as to the engine machinery which keeps them vibrating. Whilst not forgetting the title of this Lecture, I think it would be unwise to let it cramp us unduly, or to force continually the parallel with engineering problems. We ought to see something of the general setting in which engineering principles with others play their part. Moreover the principles of physics and the principles of engineering are one and the same. I remember reading in a transatlantic journal an article in praise of the benefits of science. After detailing many valuable inventions the article went on to say, " Prof. Newton of Cambridge University, Eng., has discovered three laws of motion which have been highly beneficial to engineers." In this indisputable manner I might drag in any kind of astronomical progress I desired. ,But I may be allowed to say this. In using 452

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

45

the terms " engine " and " machinery" in connexion with stars I am not merely offering a bait to engineers. So far as my personal work on the interior of a star is concerned it was an engineering problem viewed from an engineering standpoint which led me to take it up. I was attracted by the problem of the pulsating stars in which mechanical energy must constantly be supplied to maintain the pulsation against dissipative forces; and I approached it from the point of view that the interior of the star must be treated as a heat engine and therefore subject to the familiar law of heat engines, namely that in order to produce mechanical work heat must be supplied to the working substance at a high temperature and withdrawn at a lower temperature.* It was the effort to find out how the acquisition and loss of heat by the various parts of the pulsating star come to be in the right phase to conform to this law which led me to the study in more detail of the flow of heat through a star and the various other problems that unfolded themselves. If there is one principle of physics which is pre-eminently the property of engineers it is the second law of thermodynamics. That, as you know, is bound up with the theory of the heat engine ; both its history and usage stamp it as essentially a development on the engineering side. Sometimes in astronomy we use it in a direct manner as in the problem of pulsating stars to which I have just referred. Sometimes we employ its remoter consequences. But in one form or another it has been the mainstay of astrophysical investigations, and I shall allude to it from time to time. The Sun.—We may take the sun as a fairly typical star, and it will be well to glance first at modern photographs of the sun's surface. Fig. 1 shows a picture taken with the spectroheliograph. The sun has no definite surface ; we look downwards through its atmosphere towards the interior until the opacity becomes too great for us to penetrate farther. The spectroheliograph, by selecting light of particular wavelength, has the property that it shows us what is going on at one particular level in this atmosphere, whereas the ordinary photograph gives merely the blurred impression of all layers superposed. Fig. 1 shows a rather high level, and you will notice the two vortices whirling in opposite directions. The picture conveys some idea of the turmoil of the atmosphere through which the tremendous flood of heat from the interior streams out. Our task this evening is to dive beneath this layer and study the interior which lies hidden from sight. The Observatory, 1917, vol. 40, page 290.

453

46

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

The pressure at various depths in the sun can be calculated without much difficulty just as the stresses in a pillar or wall are calculated.

The pressure at the centre of the sun is between 10,000 million and 100,000 million atmospheres; that represents the accumulated weight of the long column of material stretching up to the surface 454

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

47

which is pressing down on the material at the centre. The temperature is more difficult to calculate, and indeed until recently it was thought that the problem was soluble only for the so-called " giant" stars which are much more rarefied than the sun. But we now realize that in stellar conditions matter remains a true gas even at very high densities, so that the simplifications appropriate for a perfect gas are admissible. The temperature at the centre of the sun is found to be 40,000,000° C.* It is a remarkable fact that our calculations give the same central temperature for the great majority of the stars. We divide stars into three groups which probably represent three successive phases of evolution, namely the Giant Stars, the Main Series, the White Dwarfs. The great majority of the stars are on the Main Series; they form a sequence stretching from the brightest to the feeblest stars, from the bluest to the reddest, from the most massive to the least massive ; but notwithstanding their differences in all other respects they all have practically the same central temperature. It would almost seem that there is some unknown peculiarity about a temperature of 40 million degrees; but I do not think any of the attempts to account for it can be regarded as satisfactory. E f f e c t of High Temperature.—In dealing with these high temperatures we have to recognize that a profound change has occurred in the state of matter. The atoms are broken up—or, in scientific terms, strongly ionized. We must be careful in using the phrase " broken up " because two quite distinct atomic processes may be so described. Taking, for example, an atom of iron, it is pictured as consisting of a small compact nucleus which contains nearly all the mass, together with a system of twenty-six satellite electrons more or less remote from the nucleus. The nucleus is only a minute speck at the centre, and it is the system of satellite electrons which balloons out the atom to the dimensions that we ordinarily attribute to it. This system is disarranged and torn by the terrific commotion in the interior of a star, so much so that most of the iron atoms retain only two of their electrons which are held closer to the nucleus and more tightly than the others; the remaining twenty-four are loose and wander freely about in the material as though they were little atoms on their own. We should * I shall not attempt in this brief survey to give limits of uncertainty, nor to indicate the various doubts and difficulties. I give throughout the conclusions which seem to me most plausible in the present state of knowledge, but we are constantly on the look-out for opportunities of revising and improving the results. 455

48

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

notice one point which will be important later, the big ballooning structure which gives volume to terrestrial atoms has disappeared in the star ; all that is left is the nucleus with two close electrons, a structure occupying only a millionth of the volume of a terrestrial atom of iron ; the twenty-four electrons when released from sentinel duty as outposts of the atom take up practically no room as individuals. This breaking up must not be confused with permanent transmutation of the atoms such as occurs when radium disintegrates. However shattered the electron system may be, so long as the nucleus remains intact the atom is still an iron atom. If the turmoil were stayed for an instant, i.e. if the temperature dropped, each nucleus would seize the right number of free electrons and make its system of twenty-six complete. The material would become ordinary iron vapour again. If, however, the nucleus is broken a permanent change is made in the nature of the atom and it is no longer the same element. The temperature in the stars is, so far as we know, not high enough to produce breakage of the nucleus and its effects are limited to the outer electron system. Eutherford employing other means has succeeded in artificially disintegrating the nuclei of some of the lighter elements, so that they are evidently not impregnable to attack. But what goes on in the stars is mild compared with what goes on in the Cavendish Laboratory. X-Rays in the Interior.—Besides shattered atoms and free electrons rushing about tumultuously there is something else in the star of great importance to its mechanism. Vast quantities of X-rays are present travelling in all directions. I shall presently explain how we know that they are there. (You will remember that I am speaking of regions deep in the interior which we have no direct means of penetrating.) But I will first try to indicate their importance. We can picture the scene. Atoms tear along at 50 miles a second having lost their voluminous clothing in the scrimmage. The free electrons torn from them are hastening a hundred times faster. Collisions occur in reckless profusion. After a thousand narrow shaves and side-slips an electron is caught and begins to circulate round its captor ; but barely has it settled down when an X-ray bears down on the atom and blasts the electron away again. And so it goes on, energy passing from ether to matter and back to ether. Here an ether wave is absorbed and an electron endowed with the energy darts away into freedom; there an electron is captured and its escaping energy shakes the ether into vibration. And what is the result of it all ? Only this ; the X-rays (the ether456

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

49

waves), rushing first one way and then the other, nevertheless on the average make slow progress outwards. The ethereal energy leaks slowly through the star as through a sieve. If suddenly all the atoms and electrons of the sun vanished by a stroke of the magician's wand, the X-rays in the interior would scatter through space with the velocity of light; two million years' supply of radiant energy would be squandered in an instant. The atoms dam up this store, catching and turning back the ether waves as they try to escape, and only a slight leakage dribbles out to illuminate and warm the earth and other planets. This brings us face to face with the most practical problem of stellar radiation. Regarding the star as a power station, do we find that its equipment is properly proportioned to its output ? I deal with the equipment, not with the fuel supply. Fuel supply is also an important problem; but it is much more difficult and leads us to theories of subatomic energy which are still largely conjectural. On the question of equipment progress has I think been much more substantial. We can calculate the quantity of ethereal energy in the form of X-rays imprisoned in the interior of the sun, trying to push its way out through the obstruction of the atoms and electrons. We can observe the amount which actually gets through, the energy in the form of heat and light sent out by the sun into space. The question is : is this observed rate of leakage consistent with what we know experimentally of the power of matter for obstructing the passage of X-rays, that is to say its opacity to X-rays ? I may say at once that it turns out to be at any rate of the right order of magnitude. We infer the existence of these X-rays inside the stars simply from the second law of thermodynamics and we calculate their quantity and intensity from the same law. We simply apply in the interior of a star the well-known rule that an engine working between two given temperatures cannot develop more than a certain maximum efficiency. If we are wrong, then it will be possible to make an engine worked by X-rays which has a greater efficiency than is allowed by the rule—a contingency which we are convinced is impossible. There are two stages : by a very simple application we can prove that the amount and constitution of the radiant energy within any constant-temperature enclosure depends only on the temperature and not at all on the material in the enclosure or its walls ; * by a much more elaborate mathematical application * For let A and B be two enclosures with walls of different material, both at temperature T. Our assertion is that the intensity of the ether E

457

50

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

we deduce from our knowledge of the amount and constitution of the radiation at any one temperature its amount and constitution at any other temperature (Wien's Law). Thus experiments at terrestrial temperatures are sufficient to teach us all about the nature of the radiation in material even at 40,000,000°, provided only that we admit the universal validity of the law of heat engines. The fact that at these high temperatures the radiation is in the form of X-rays, i.e. ether waves of very short wavelength, considerably simplifies our problem. If we had to discuss the leakage of ordinary light rays we should be rather nonplussed. How opaque would you expect stellar material to be to light ? You cannot say unless you have some idea of the chemical nature of the material. But for absorption of X-rays chemical composition makes much less difference. Indeed along the whole front of progress the high temperature is a great simplification, and the investigator heaves a sigh of relief when he gets above a million degrees. It is a remarkable fact that although we know next to nothing of the chemical constitution of the interior of a star this is scarcely a hindrance to the calculations that interest us. The sun might be made of any element from helium to uranium and we should still predict almost the same brightness for it; hydrogen, however, stands by itself and makes a big difference. Our results will therefore rest on the assumption that the stars do not contain an excessive proportion of hydrogen. Perhaps some years hence when we are more sure of our ground, we shall be rather sorry for this simplification ; it spoils our chance of learning from a comparison of theory and observation whether the stars are made up of light or of heavy elements. I have spoken of the X-rays leaking outwards through the star; what actually emerge are, however, not X-rays but light and heat rays. That requires a word of explanation. Perhaps it might be put this way : the stellar power station is equipped with transformers waves of any given kind, say of wavelength A, is the same in the interior of both A and B. If not, let it be greater in A. Then connect A and B by a passage barred with a screen transparent only to wavelength A. Since the intensity is greater on the A side than on the B side of the screen there will be a net passage of radiation from A to B. B will now contain more radiation than the equilibrium amount with which it had previously settled down, and the surplus must therefore pass away into the walls which will rise in temperature ; similarly the walls of A will fall in temperature. Since they now differ in temperature, B and A can be used as source and sink of a heat engine. If the engine works a valve opening and closing the passage from A to B it can run continuously, using only heat supplied at temperature T (which might be atmospheric temperature); the opening of the valve continually recreates the temperature-difference which an engine requires. 458

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

51

which step down the high-tension output of the main machines to a low-tension form for outside distribution. In the star the transformation is gradual; as they progress outwards the X-rays gradually lengthen in wavelength until they can no longer be classed as X-rays. But the last stage of the stepping down from X-rays to light rays occupies so small an outer shell of the star that we are not concerned to follow it in detail. Once we have traced the radiant flow up to the beginning of this shell our work is really ended ; it will all emerge somehow or other because there is too little material in the shell to plunder or reinforce the stream appreciably. The amount of outflow is determined by the temperature-gradient and opacity in the interior; the outflow may change in character but not in quantity in passing through the outer shell. If these outside layers tried to check it they would be blown away ; actually they adjust themselves so as to let it through. Theoretical and Observed Output of Heat.—Having perhaps sufficiently indicated the general line of the calculations, we pass on to the comparison of the theoretical output and the observed output of the stellar power station. You will recall that we have not been dealing with the question of fuel supply. The heat is supplied in the deep interior we do not particularly mind how; we deal only with the manner in which the star regulates its passage into outer space. Our problem may be compared with a calculation of water supply from the pressure and diameter of the mains, which does not involve any discussion of the problems of the pumping station. We find first that the output depends on the mass of the star. Eather surprisingly it does not depend to any important extent on the density-—whether the mass is diffused into a large volume or concentrated in a small volume—but it is assumed that the density is not too great for the material to behave as a perfect gas. Fig. 2 shows the theoretical curve giving the output plotted against the mass, or rather the logarithm of the output (absolute magnitude of the star) against the logarithm of the mass. The circles, crosses, etc. indicate observed values; they include all the observational data that is considered reasonably trustworthy and indeed some that is rather inaccurate. The circles are the most trustworthy. It will be seen that the agreement is good. Apparently the theoretical curve runs rather too low on the left; and I think we know the reason for this. There is an electrical effect prominent in stars of small mass which makes the material superperfect gas ; this gives a correction in the right direction, but since the magnitude of the correction depends greatly on the chemical composition it E 2 459

52

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

would be arbitrary to apply any particular value. It may be added that the diagram covers practically the whole range of stellar magnitudes and stellar masses, so that it is a fairly stringent test of a theory (necessarily only a first approximation) that it should be applicable through so wide a range. The sun is generally regarded as a typical star; and indeed it is just about as ordinary a star as you could find. But the stars have great variety and range widely on either side of this type. About fifteen years ago it was realized that some stars were enormously swollen in bulk. Not containing much more material FIG. 2.—The Mass-Luminosity Relation.

than the sun they nevertheless fill a thousand or a million times the volume. The well-known star Betelgeuse is so great that the whole orbit of the earth could be put inside it, and its density is so low that it corresponds to that of a fairly good vacuum. Capella, which is one of the stars represented in Fig. 2 has a mean density about equal to air; and many other examples of these bloated " giant stars " are known. Stars of density equal to or below that of air may evidently be treated as perfect gas, and should agree with the curve calculated for a perfect gas; but the surprising thing is that these are not the only stars that fit the curve. The sun has a mean 460

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

53

density greater than water but it is very near the curve. Other stars have the density of iron or lead : they also fit, or if they deviate at all it is on the superperfect side. As soon as the diagram called attention to this behaviour the explanation suggested itself. A gas ceases to be perfect when it has been compressed so much that its atoms begin to jam in contact, or when the volume actually occupied by the atoms is a considerable fraction of the whole volume of the gas. But we have already seen that the structure which swells out the volume of a terrestrial atom is broken up in the star. The tiny fragments which survive will not jam at any ordinary density. The breaking-up of atoms at temperatures of the order of ten million degrees has the effect that stellar material of the density of water or even of platinum is still a true gas. Matter of High Density.—In saying that material as dense as platinum is (under stellar conditions) a perfect gas, I mean that it still has the characteristic compressibility of a gas, namely that if you double the pressure you halve the volume. There is nothing to prevent advantage being taken of this compressibility so that the density need not stop at that of platinum. Under sufficient pressure a density may be reached far transcending anything known to us on the earth. There is no limit to the compressibility until the tiny fragments of atoms begin to jam and that will not happen until we reach densities something like 100,000 times greater than those of terrestrial liquids and solids. There are two or three stars known to us which are believed actually to possess enormous densities which a few years ago seemed incredible. They are called " white dwarfs." They are intrinsically faint, giving out much less light than the sun ; and yet the white colour of their light and the character of their spectra show that their surfaces must be at high temperature glowing very brilliantly. The surface is brilliant though the total light is small; this combination can only be accounted for by supposing that they are small objects. The mass or total amount of matter is found to be much the same as that of the sun ; and since this has to be packed into a much smaller globe the resulting density is enormous. The best known of these stars is the Companion of Sirius—a faint star which first betrayed itself because it was noticed that Sirius was moving irregularly as if under the attraction of an unseen mass. Like Neptune its influence was perceived long before it was actually seen. However, it is now visible in large telescopes and can be fully examined. The density is found to be 60,000 times that of water; a ton of its material would go inside a match-box. 461

54

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

As soon as it was realized that the calculated density of the Companion of Sirius was after all not incredible but quite consistent with the modern view of the nature of stellar material, efforts were made to apply some more crucial test. Einstein's theory predicts a small displacement of all spectral lines to the red end of the spectrum when light comes to us from a place where there is a strong gravitational field. This effect is very small on the sun and it is even yet not entirely certain that it has been definitely observed, though there is a general consensus of opinion that it is confirmed. But since the effect is inversely proportional to the radius of the star it becomes greatly magnified in these very dense stars, and it should be conspicuous in the Companion of Sirius if the small volume attributed to it is correct. The test was applied by W. S. Adams in 1925 and the large Einstein shift was duly found. Since then it has also been confirmed by J. H. Moore at the Lick Observatory. It seems therefore to be established that the Companion of Sirius is an example of the enormous density which can be attained by matter under stellar conditions. Dense material such as I have been describing must appear to us at first almost as an incredible fairy-tale. But I am sure we are looking at it from the wrong attitude. You terrestrial engineers are concerned with matter in a very abnormal state. Nine-tenths of the matter of the universe (at any rate of the matter that we are directly aware of) is tucked away in the interior of stars under temperatures of more than a million degrees. That is the normal condition of matter and we must not take our experience of matter on this chilly little planet as at all typical. We have at last realized the extreme emptiness of matter ; a terrestrial atom is like a miniature solar system. If we eliminate that emptiness, if we pack the little electric charges (which are all that could possibly be considered substantial) close together, a man's body is reduced to a speck just visible with a magnifying glass. What is there that stops such close packing ? If the temperature is high enough there is nothing to prevent it; but, of course, it will be necessary to exert a very large pressure to keep the material so compressed. At low temperature, as on the earth, there is a bar which the utmost pressure is unable to overcome. The atoms are protected by a guard of sentinel electrons which ward off nearer approach. We have some idea of the way in which this atomic phenomenon operates. It is called the Exclusion Principle. In general terms it means that every electron insists on being in some way a little bit different from its neighbours. So when pressure tries to insist on electron A packing a little closer to electron B, A replies " No. We are 462

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

55

already so nearly in the same position that people can only just manage to tell us apart." But it is open to persuasion by an offer of some other distinction as a substitute for difference of position. If A differs sufficiently from B in energy or in momentum that will do just as well. So at high temperatures when there is plenty of energy to go round the electrons can distinguish themselves by seizing different quantities of it and then they will not mind losing their distinction by position. Poor things ! they are all turned out exactly to pattern by Nature's lathe, so they treasure these ways of insisting on their individuality—not to be just like one's neighbour. And so it comes about that at low temperatures the exclusion principle devotes its efforts to separating the electric charges in position and gives a large effective volume to the atom, whereas at stellar temperatures it is more concerned with distinguishing their momenta and energies and is lax about keeping them apart in position. Crystalline Gas.—Those who are curious as to the possible states of matter may be interested to know that the problem of the stellar interior leads us to consider a rather remarkable state, namely, a crystalline gas. I do not think the gas inside a star is crystalline, but it is not so far removed from that condition that we can leave the possibility out of all consideration; and at any rate material which is at the same time a crystal and a true gas seems to be possible under the right conditions. In a star the iron atoms are broken up so as to provide twenty-four electrons each carrying a unit negative charge; the residual atom by way of compensation carries a positive charge of twenty-four units. Thus in any volume there will be many small negative charges and a few big positive charges. It can be proved that the small negative charges will on the average spread fairly uniformly over the volume, crowding a little in the neighbourhood of the positive charges but not to any great extent. The large positive charges tend to take up positions as far from one another as possible, the arrangement of minimum energy being that of a crystal lattice. Whether they will actually keep to this crystalline arrangement depends on the temperature ; the energy of agitation of the material may keep them stirring and " melt " the crystal. But it seems that for heavy elements conditions are possible in which the negative charges are freely mobile like a gas whilst the positive charges remain immobilized at the points of a crystal lattice. Such matter would be simultaneously a gas and a crystal, its gaseous character being manifested chiefly in its mechanical properties of expansion and compressibility, whilst its crystalline structure would appear chiefly in its optical properties. 463

56

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

Pulsating Stars.—I come now to pulsating stars. The variable star d Cephei is a typical example. It is a rather large star probably about ten times as massive as the sun, and it is in a diffuse condition with an average density less than that of air. This globe swells and contracts periodically so that the volume at maximum is about 20 per cent greater than the volume at minimum. The variation has gone on regularly with a period of 51/3days ever since the star was first observed 150 years ago. During the contraction the inner material is being compressed and consequently rises in temperature. The adiabatic law will apply very approximately because the whole process is on so vast a scale that heat leakages are relatively small. During expansion there will similarly be an adiabatic cooling. Owing to these changes of internal temperature the outward flow of heat varies periodically and the brightness of the star waxes and wanes in the period of 51/3days. It is this varying brightness which calls our attention to stars of this class. (Not all variable stars are pulsating stars ; in many of them the change of brightness is caused by a faint companion occulting the principal star.) Among the ordinary stars we know of nearly 200 which behave like d Cephei, and in addition many hundreds have been observed in the distant starclusters. Besides the indication of varying brightness, we can also by measuring the Doppler displacement of the spectral lines observe the motion of approach and recession of the surface as it heaves up and down with the pulsation. This phenomenon immediately provokes the question: how is the pulsation maintained ? If there were no energy to renew it the vibration would gradually dissipate itself. The decay would not be very rapid and I do not think it would be noticeable in the 150 years during which we have been watching the star, but there would certainly be a considerable falling off in the course of 10,000 years. Ten thousand years is a mere nothing in the long life of a star ; and the fact that we find a large number of stars in this condition shows that it is not merely a transitory incident. We must look for some mechanism by which, as a star reaches a certain phase of evolution, a pulsation is started up and maintained for a long while. It will die down again when the star has evolved further and reached a condition no longer favourable for pulsation. There is no immediate difficulty about finding the required energy. We are bound to suppose that from some source or other heat is being liberated in the interior of the star to keep up the stream of radiation emitted into space. If only a small part of this can be diverted for use in starting or maintaining the mechanical energy of pulsation it will be quite adequate. But we must here apply 464

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

57

the principle of the heat engine. Mechanical work is only obtained from heat if there is a " working substance " taking up heat at a high temperature and giving out heat at a low temperature. Accordingly some part of the star must behave as a thermodynamic engine converting heat into work according to the above condition: Generally speaking it is in the nature of things to acquire heat from their surroundings more readily when they are at a low temperature than at a high temperature—the wrong phase for a thermodynamic engine ; that is to say, the natural tendency is to dissipate mechanical work not to produce it. The engineer overcomes this by the use of valves which artificially control the flow of heat. In the star the " valve " which controls flow of heat from one part of a star to another is the opacity—the obstruction to X-rays which constitute the more mobile part of the heat. The opacity depends on temperature and density and therefore alters in the course of the pulsation; when it decreases the effect is like opening a valve. We are therefore faced with the problem of determining which way the valve mechanism in the interior of a star is operating. Is the phase such as to admit heat into any region at the right moment and let it out at the right moment ? Using the general laws of opacity so far as they are known we have worked out this problem, and the conclusion is adverse. The valves open and close at the wrong time and increase the dissipation instead of setting the engine going. But when you come to think of it, that result is not so bad after all. Cepheid variables are the exception, not the rule; and it would have been disastrous if we had proved that the normal conditions in a star are such that the slightest provocation will start up its engine and set it pulsating. Clearly the Cepheid pulsation sets in at some rather critical temperature or density when the opacity or some other physical characteristic is put out of its normal stride. The most likely suggestion is that it is associated with the ionization or breaking up of the electron systems of the atoms ; this does not increase steadily with temperature but increases by fits and starts as successive rings of electrons are attacked. Nothing very definite has been made out; but it appears that, although normally the conditions are not such as to cause pulsation, the margin of safety is not very great for the more massive stars; and a modification of physical constants, corresponding to the stage in which the preponderant elements are undergoing a rapid change of ionization, may cause such stars to overstep the danger point. There is another possibility which seems much simpler, but I am afraid there are grave objections to it. We believe that the 465

58

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

heat of a star is replenished from some subatomic store of energy. Instead of worrying about the flow of heat from one part of a star to another we can study the " valve " which regulates the supply of heat from this store. All we have to arrange is that the tap of subatomic energy shall be turned on fully when the material of the star is hottest—at the time of greatest compression—and closed when it is coldest—at the time of greatest extension. Then the star will behave as a heat engine working between these two temperatures and converting heat into work with something like maximum efficiency. Although we have no experimental knowledge of the conditions of release of subatomic energy it seems almost obvious that the turning on and off of the tap would occur in the way suggested. If temperature has any effect at all it will presumably stimulate liberation of subatomic energy ; and if more than one atom or electron is concerned in the liberation, the increased density at the time of greatest compression will accelerate the interaction. We thus arrive at the following view. During the state of compression the liberation of subatomic energy is stimulated by the higher temperature and density, and the star is gaining more heat than it loses by radiation; this heat gives greater vigour to the rebound from compression. During the next half period of rarefaction the star correspondingly suffers a net loss of heat, and this aids the ensuing return to compression. The pulsation is thus encouraged at each stage so that a small pulsation will grow up to the greatest possible limits. But once again we have to remind ourselves that the problem is not so much to explain why stars are able to pulsate, but rather to explain why stars, with comparatively few exceptions, do not pulsate. The problem of the Cepheids sinks to relative unimportance ; we see that there is a more fundamental problem of stability of ordinary stars. The law of release of subatomic energy must be such that in general it does not upset the stability of stars. I carniot here enter into the intricacies of the problem; I will only mention one circumstance which makes it somewhat different from analogous problems in our terrestrial experience, namely, that a star has a negative specific heat. Negative Specific Heat.—It is our usual experience that when a body loses heat it gets colder. Nearly sixty years ago Homer Lane showed that a star behaves paradoxically, and when it loses heat it gets hotter. It is true the deception practised on us is rather transparent. It happens in this way. When a star loses heat (by squandering it in radiation, for example) the pressure in the interior 466

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

59

is relaxed a little so that the whole star contracts. But this contraction or falling in of material generates heat, converting gravitational energy into heat; and the generation of heat more than replaces the loss. The result is that internal temperature rises. Lane's law was only asserted for gaseous stars ; for liquid stars the contraction would no longer be great enough to replenish the heat radiated; but we now believe that all ordinary stars are perfectly gaseous, so that the law is general. If then we ask how much heat must be added to a star in order that its temperature may rise by 1°, the answer is that no heat must be added; heat must be taken away. As Professor H. N. Kussell has put it—the specific heat is negative. You will understand that there may be serious difficulties about the fuel supply of a body which has this strange property; and actually we find considerable difficulty in arranging a supply which will neither die out prematurely nor explode into pulsation. Only a narrow range of possible laws of release steer between the Scylla of instability on the one hand and the Charybdis of pulsation on the other hand. It is only because there is some dissipative resistance damping the pulsations that any channel exists at all. So far as I can see the conclusion towards which we are led is that the influence of temperature and density on the rate of liberation of subatomic energy must be an indirect one. The energy is released from certain active substances formed inside the star; the rate of formation of these substances increases with temperature and density, but they break up and liberate the energy at a rate unaffected by temperature and density.* Whether this is the correct solution or not, we can at least show that many of the simpler theories of release of subatomic energy are untenable, since they would necessarily lead to pulsations of all stars. One further point may be mentioned before we leave the Cepheids. The stars differ from one another in two respects, namely in mass and density. It appears that there is just one stage of evolution at which Cepheid pulsation is set up, so that for any given mass there is just one corresponding density which permits pulsation. Now the period of pulsation depends mainly on the density, being longest in the most diffuse stars ; there is in fact excellent agreement between theory and observation as regards the periods. It follows that just * Thus when the star is in a state of compression there is no immediate increase in the rate of liberation of energy and therefore no assistance to pulsation. But if the compression is long continued the rate of liberation gradually increases, because of the increasing amount of active material formed. 467

60

THOMAS HAWKSLEY LECTURE.

FEB. 1929.

one kind of star can have Cepheid pulsations of given period; and if we meet with a Cepheid variable of period 5J days anywhere in the universe, it ought to be in all respects a replica of d Cephei. This has been stringently tested by observation; there is a definite periodFIG. 3.—Star-Cluster w Centaur i (Cape Observatory).

luminosity law, so that as soon as we know the period of a Cepheid variable we can say at once what its absolute brightness must be. Fortunately the Cepheid variables are among the most luminous of the stars and they can be observed in the remotest parts of the universe ; wherever we observe them we have at once a gauge of distance. Results of the greatest interest have been found in this 468

FEB. 1929.

THOMAS HAWKSLEY LECTURE.

61

way by Hertzsprung, Shapley and Hubble, and our present idea of the scale of the universe is based almost wholly on this evidence. FIG. 4.—The Great Andromeda Nebula (Yerkes Observatory).

Fig. 3 shows a famous star-cluster w Centauri. Among the minute stars shown in the picture seventy-six Cepheid variables have been 469

62

THOMAS HAWKSLEY LECTURE.

FEB. 1929,

discovered. For each of them we learn the absolute brightness from the period, and can then judge what must be the distance of the cluster to reduce this brightness to the apparent brightness observed. They all agree that the distance is 20,000 light years. Fig. 4 shows the great Andromeda nebula. In the outer part of this Cepheid variables have also been discovered. Period for period they are much fainter than those in the cluster w Centauri indicating much greater distance. By this comparison of the apparent brightness of the Cepheid variables, we learn that the light we are now receiving from the Andromeda nebula left it 900,000 years ago. Fuel Supply of a Star.—The question of the fuel supply of a star is one which the astronomer finds it difficult to keep away from, although there is not very much progress- to record. It seems clear that the stellar furnace must be stoked from inside, not from outside. The problem is not just to maintain the radiation from the surface, but to maintain the tremendous temperature in the interior which keeps the star distended to its actual size. Thus it is no use supposing that the sun's heat is supplied by a bombardment of its surface with meteors ; you cannot maintain a temperature gradient by supplying heat at the bottom end, and the bombardment would not prevent the heat flowing away from the interior. The source of stellar energy must supply heat in the deep interior. The principal basis for discussion was supplied by Einstein, who showed the equivalence of mass and energy. A mass of 1 gram represents an energy of 9.1020 ergs, and by the same proportion we at once find the total amount of energy contained in the sun or any other star of known mass. Accepting our conclusion that the heat cannot be supplied from outside, this calculation gives the total store of energy which the star can draw on to provide heat for the rest of its life. The store is ample for all reasonable demands, and it would suffice to maintain the sun's present radiation for 20 billion years. But the question arises whether any considerable proportion of it is actually utilizable. There is the same enormous store of subatomic energy in the material of this earth, but it is a very remote dream that we shall ever be able to utilize it. Most of this energy is energy of constitution of the electrons and protons of which matter is built up. If it is released the matter must disappear. It is as though something might slip in the structure of the world, and where we had previously a proton and electron we have now only an ether-wave carrying off the released energy. Is such a slipping back possible ? We can only speculate. But it is not at all an idle speculation, because our whole theory of stellar 470

FEB. 1929.

THOMAS HAWKSLEY LECTUKE.

63

evolution turns on the answer. If this energy of constitution which forms 99 per cent of the whole store is not being released, the corresponding 99 per cent of the star's mass is permanent. It,seems impossible that a star can lose mass by actual escape of material anything like so fast as it loses it by radiation. Hence the star will remain of practically constant mass until the fuel supply gives out and the star becomes dead. But if the energy of constitution is releasable, the star can gradually radiate away any proportion of its mass, and big stars will evolve into little stars. According to the results illustrated in Fig. 2 brightness and mass go together ; so the second hypothesis provides for an evolution of faint stars from bright stars whereas the first hypothesis requires that the star shall remain practically of the same brightness all its life (except at the extreme beginning and end). There is much observational evidence which favours the view that a star grows fainter and less massive as it grows old; and this is our main ground for accepting provisionally the hypothesis that a star's heat is provided by destruction of matter inside it. But on the other hand there are some observational results which are hard to reconcile with this. I have already .referred to the difficulty of assigning laws controlling the rate of this (or any other) subatomic process, and the fact that the simpler theories must be rejected as making the star unstable. At present we seem to be back at the beginning of the problem of stellar evolution, scarcely able to make up our minds whether faint stars are evolved from bright stars or whether they were born bright and faint, respectively. A biologist cannot have got far towards a satisfactory theory of evolution if he is still in the position of wondering whether puppies grow into dogs or whether dogs and puppies were born different. That is more or less our state to-day. Nevertheless our feeling is not one of despair at the indefiniteness of present conclusions, but rather of interest and anticipation with regard to a great problem unfolding itself before us, which it seems possible to advance by many and varied lines of research. And if you should at some future date again select an astronomer to deliver this Lecture, it is not unlikely that he will have much to tell of progress in the field of stellar evolution coupled with a clearer understanding of those stores of energy held so fast and so tantalizingly within the atom.

471

This page intentionally left blank

Herbert Nigel Gresley (1876-1941) Gresley began his engineering training at the Crewe Works of the London and North Western Railway. Moving North, he joined the Lancashire and Yorkshire Railway and by 1905 he was at Doncaster, eventually becoming locomotive engineer to the Great Northern Railway. During the First World War, Gresle/s contributions to improved wartime production practices were noteworthy. Gresley introduced the prototype of the "Pacific" class of locomotive (4-6-2) for long-haul passenger traffic during his time at Doncaster, work which he extended as Chief Mechanical Engineer of the London and North Eastern Railway. Here he was responsible for the construction and repair logistics of a truly large company. His design team increased the power and efficiency of locomotive stock. From 1928, this included close collaboration with the French engineer Andre Chapelop. Much of Gresley's efforts involved the rebuilding and improvement of existing locomotive designs. He was very interested in the possibilities of promoting express speed by streamlining locomotives, a topic which took up much of his Presidential address to the Institution of Mechanical Engineers in 1936. Noting developments in France, Germany and the United States, Gresley applied them to a British design. The Silver Jubilee which entered service in 1935 was the first British engine of this type. A later "Pacific" locomotive, Mallard attained a world record speed of 126 mph. Although best-known for his steam locomotives, Gresley worked with electric and diesel machines. He also made great contributions to passenger comfort and safety by redesigning coaches, restaurant cars, and sleepers, and by pioneering all-electric facilities.

473

101

JAN. 1931.

HIGH-PRESSUBE LOCOMOTIVES.

BY H. N. GKESLEY, C.B.E., M.I.Mech.E. (Member of Council).

At no time during the history of the steam-locomotive have such radical changes been introduced as during the past ten years. These changes have been introduced more or less simultaneously in countries supplying the locomotives of the world, and have no doubt been stimulated by the competition of electricity and oil-engines which offered more efficient and in some cases cheaper motive power. Twenty years ago it was thought that the steam-locomotive had attained practically its maximum development, as it had nearly reached the limits imposed by the load-gauge. The radical changes necessitated by the adoption of extra high pressures seem to have opened up a new era. Not only do these changes render possible an increase in tractive power, but should at the same time increase the overall efficiency. In Great Britain alone there are 23,000 locomotives, and a sum of approximately £45,000,000 per annum is spent on their maintenance, renewal, and running. Of this sum, nearly £12,000,000, or 25 per cent, is the cost of fuel burnt by these locomotives, and another £12,000,000 has to be spent on their maintenance and renewal. The purpose of the novel forms of locomotive which have been introduced during the last five years has been to effect economy principally in fuel consumption, but the Author wishes to draw attention to the importance of the cost of maintenance, and the necessity for designers of new and improved forms of high-pressure locomotives to remember that the expense incurred in maintaining locomotives is equal to the cost of the great quantity of coal which [THE I.MECH.E.]

475

102

HIGH-PRESSURE LOCOMOTIVES.

JAN. 1931.

they consume. It will be seen what a large field there is for economy if both cost of fuel and maintenance can be reduced, and what an influence such economy can have on the cost of transportation on railways. From the early days of the locomotive it has been recognized that increase of boiler pressure results in decrease of coal consumption, but with the conventional type of locomotive boiler this has generally resulted in increasing the cost of firebox repairs, and reducing the life of boilers and fireboxes. When the "Rocket" was built by George Stephenson a boiler pressure of 50 Ib. per sq. in. was employed, and pressures have gradually increased until now there are many engines running in this country with a boiler pressure of 250 Ib. per sq. in. Pressures up to 325 Ib. per sq. in. have been experimentally tried for locomotives in Germany and America, but 250 Ib. per sq. in. can be looked upon as approximately the maximum pressure which can be carried in a boiler of the Stephenson type, having regard to the cost of boiler maintenance. It has taken 100 years to increase the pressure of locomotives from 50 Ib. to 250 Ib. per sq. in., but during the last five years pressures have leapt up to 450, 900, and now to 1,700 Ib. per sq. in. The use of pressures above 250 Ib. per sq. in. has necessitated the design of a completely novel form of boiler built up of tubes and circular drums in which, generally, all flat surfaces have been eliminated. Ingenious means have been adopted for transmitting the heat from the fire to the water. The use of very high pressures, and the consequent high temperature, result in conditions which are much more exacting to the materials used in the construction of the boilers. These conditions are aggravated unless the heat transference is very rapid. In striving for economy by the use of higher pressure, designers of locomotives are only following the lead which has been set by the designers of large stationary plants and marine engines. Their problem, however, has been made more difficult by loading-gauge and weight restrictions, and for these reasons they are unable to take advantage of condensers and have had to extend the pressure gradient upwards to a greater extent. It does not necessarily follow that the use of high pressures in boilers is more dangerous than that of low pressures. An explosion of a boiler at a pressure of 50 Ib. per sq. in. can have disastrous results, and there is no reason, if proper precautions are taken both in design and maintenance, to assume that there is any greater liability to explosion or failure as a consequence of increase in pressure. Whilst high steam pressure gives greater economy in fuel 476

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

103

consumption, it demands complication in design, and care must be taken that the economies in fuel are not absorbed in the increased cost of maintenance of the boiler and of the machine as a whole. Simplicity of design is an important factor, because simplicity generally results in accessibility. Time alone can prove which of the designs, if any, that have been recently produced, will result in such overhead economies as will justify their general adoption. The results which may reasonably be expected from the use of high steam-pressures on locomotives are so attractive that encouragement should be given to the production and development of designs by which these results can be attained. Keciprocating pistons have been adopted in all the latest highpressure locomotives, as this form of conversion of energy appears to be the most advantageous for meeting all conditions which a locomotive is required to fulfil. It is interesting to note that in the high-pressure locomotives which have been produced during the last five years, both two-, three-, and four-cylinder compounds have been adopted. The Author proposes to describe in some detail the notable high-pressure locomotives which have been produced in America, Germany, Switzerland, and England since 1924. Delaware and Hudson Two-Cylinder Compound Locomotives.— The first of these was built in 1924 by the American Locomotive Company to the designs of Mr. John E. Muhlfeld, for the Delaware and Hudson Eailway, and marks the first distinct advance in boiler pressure on a main-line locomotive, the pressure being 350 Ib. per sq. in. The engine was called " Horatio Allen." It is a two-cylinder compound with 2-8-0 wheel arrangement, the tender being fitted with a booster. The boiler barrel follows usual practice, but over it, on each side, two cylindrical drums are fixed. These extend from the rear end of the firebox nearly to the front of the boiler barrel. Two shorter drums form the lower sides of the firebox, the front wall being formed by a flat water space with ordinary stays. The upper drums are connected to this flat water space, and pass through it. The boiler barrel is secured to the front wall of the water space. The rear plate of the water space is cut away for the firebox tube-plate which is fixed in it, and the front ends of the lower drums are also attached to it. The back end of the firebox is formed by a similar flat water space into which the rear ends of the four drums are secured. The top of the firebox between the top drums is formed by eight horizontal water-tubes connecting the front and back water spaces, 477

FIG. 1.—Elevations and Cross-Sections of High-Pressure Boiler with Water-Tube Firebox. Delaware and Hudson 2-8-0 Compound Locomotive "Horatio Allen."

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

105

and each of the sides by six rows of curved water-tubes joining the upper and lower drums. The front ends of the upper drums are connected to the boiler barrel by headers, and four pipe connexions are also provided; sections of this boiler are illustrated in Fig. 1. With this type of boiler about one-third of the evaporative heating surface is in the firebox. In ordinary locomotive boilers the firebox heating surface is only about one-tenth of the total. There is nothing of exceptional novelty about the rest of the locomotive. It must have given satisfaction, because the Delaware and Hudson Railway Company obtained a very similar engine from the American Locomotive Company early in 1927 which was called " John B. Jervis." The boiler is of very similar construction, but the pressure has been raised to 400 Ib. per sq. in., and a greater superheating surface given. The firebox is slightly larger, and it is covered by sheets of heat-resisting steel lagged on the outside, with iand-hole openings to allow ashes to be cleared from the fire-tubes and tops of the drums. The brick arch extends the entire length of the firebox, and causes the gases to flow outwards, and up through the side water-tubes into the combustion space above the arch, and to the fire-tubes. TABLE 1.—Delaware and Hudson High-Pressure Locomotives. " Horatio Allen " Date built Grate Area . . sq. ft. Heating Surface : — Firebox . ,, Total Evaporative ,, > Superheater . ,, Boiler Pressure, Ib. per sq. in. Cylinders :— Diameter H.P. . in. : ,, L.P. . ,, Stroke . . ,, Driving- Wheel Diameter „ Adhesive Weight . tons Total Weight of Engine „ Tractive Force : —Simple . . lb. Compound . . ,,

1924 71.4

" John B. Jervis "

"James Archbald"

1927

1930

82

82

1,187 3,200

1,217 3,121

579 350

700 400

1,114 3,439 1,037

22 1/4 38

23 1/2 41

30 57

i

30 57

133 156

132 150

84,300 70,300

85,000 70,800

500 20 1/2 35 1/2 32 63 134 159

85,800 71,600

479

FIG. 2.—Delaware and Hudson Railway Locomotive " James Archbald."

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

107

The results obtained in advancing by progressive stages having proved satisfactory, the Delaware and Hudson Railway Company in 1930 produced a third engine in which the pressure was increased to 500 Ib. per sq. in. which was named " James Archbald." This engine has a boiler of the same type as the previous engines, but its external appearance is noticeably different, the boiler being enclosed by an outside casting which effectively insulates and stream-lines the upper part. The superheating surface has been considerably increased, giving a steam temperature of 700° to 750° F. A photograph of this engine is shown in Fig. 2. The main dimensions of these three locomotives are given in Table 1. The Schmidt-Henschel Three-Cylinder Compound Locomotive.—In 1926 one of the standard three-cylinder 4-6-0 engines of the German State Railways was converted by Messrs. Henschel and Sons of Cassel into a high-pressure locomotive. This engine was fitted with a multi-pressure type of boiler built to the design of the Schmidt Superheater Company, in which steam was produced at a pressure of 850 Ib. per sq. in. by means of a closed circuit in which the heat transfer medium was distilled water. The design was the result of the invention and many years' investigation of the late Dr. Schmidt. A full description of this engine was given by Mr. R. P. Wagner, M.I.Mech.E., Chief Engineer of the German State Railways, before the Institution,* in 1927, and only a brief account of the principle of its operation, therefore, is given in this Paper. In this system the tubes forming the walls of the firebox are filled with distilled water. They are subject to direct radiant heat, and owing to the purity of the water there is no corrosion or scaling to cause overheating of the tubes. The principle is illustrated diagrammatically in Fig. 3. The steam is generated in the firebox tubes at a pressure of 1,200 to 1,600 Ib. per sq. in., the pressure depending on the rate of firing. The steam passes through separating drums to the heat-transfer elements situated in the high-pressure drum. The difference in temperature between the saturated steam of 1,200 to 1,600 Ib. pressure in the heat-transfer elements, and the water in the high-pressure drum at 850 Ib. per sq. in. pressure causes the latent heat in the steam to evaporate the water. The steam in the closed circuit is condensed by the removal of its latent heat and returned as water by way of the down-pipes to the firebox bottom ring. * Proc. I.Mech.E., 1927, vol. ii, page 961. 481

FIG. 4.—Schmidt-Henschel High-Pressure Locomotive.

FTG. 3.—Diagrammatic Sketch of the Schmidt HighPressure Locomotive Firebox.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

109

The boiler consists of two sections ; one section generates the high-pressure steam at 850 Ib. per sq. in. in a high-pressure drum by means of the closed circuit as described above, and the other section, which is similar to the barrel of an ordinary locomotive boiler, produces steam at about 200 Ib. per sq. in. pressure. Fig. 4 shows sections of this locomotive, and from these the design of the boiler can be seen. The high-pressure drum is a solid forging of about 3 feet internal diameter and is protected from the direct heat of the firebox gases by the cross-over tubes and by steel protection plates covered with asbestos sheets. The heating elements are inserted into the drum through a manhole at the rear end and consist of pipes 11/2inches diameter arranged in groups of coils. The inlet on each element is connected to the separator drums and the outlet to the firebox ring. The low-pressure boiler has flues 31/2inches outside diameter and nearly all these contain superheater elements. The flues at the top contain the low-pressure elements, and those at the bottom, elements for high-pressure steam. The water from the tender is fed into the low-pressure boiler by ordinary injectors or feed-pumps. The feed for the high-pressure drum is delivered by a pump from the low-pressure boiler, the water being at a temperature of 390° F., which is the saturated steam temperature of the low-pressure boiler ; consequently most of the solid matter in the water has been deposited in the low-pressure boiler. In operation, superheated steam from the high-pressure boiler enters the centre cylinder and exhausts into a receiver where it mixes with superheated steam from the low-pressure boiler, the combined steam passing to the low-pressure cylinders and so to the blast-pipe. The operation of the locomotive is similar to one of ordinary type as the two regulators are coupled together and the reversing gear is controlled by one lever. The locomotive was delivered in 1926, and after several minor alterations was put into ordinary service and has given satisfactory results. Before being put in regular service, it was subjected to a series of comparative trials with locomotives of the standard threecylinder simple type as used for express trains on the German State Railways. The coal consumption per drawbar horse-power hour of the Schmidt engine was about 2.5 Ib., and the steam consumption about 18 Ib. Since then certain modifications have been made resulting in further reduction of fuel and steam consumption to 2.15 Ib. and 15 Ib. respectively, giving an economy of 35 per cent in fuel and 23 per cent in steam as compared with engines of the standard type. The 483

1. Steam-Drum. 2, 3. Bottom Drums. 4, 5, 6. Water-Walls. 7. Water-Tubes. 8. Stay Tubes. 9. Firebox. 10, 12. Combustion Chamber.

FIG. 5.—Winterthur High-Pressure Locomotive. 11. Grate. 13. Safety-Valve. 14. Kegulator. 15, 28. Superheaters. 16, 27, 31, 32. Stop-Valves. 17, 18. Main Feed-Heater and Delivery. 19, 24. Smokebox and Chimney.

20, 21, 22. Air-Heater. 23, 36. Exhaust Pipe. 25, 26. Ash Outlets. 29. Feed-Pump. 30. Feed-Preheater. 33. Tank. 34. Water-Gauge. 35. Engine. 36. Drain.

JAN. 1931.

HIGH-PBESSUEE LOCOMOTIVES.

111

calorific value of the fuel used was approximately 12,600 B.Th.U. per Ib. The proportion of the total power supplied by the high-pressure boiler at low output is 65 per cent and at higher output falls to 55 per cent. As the remainder of the steam is formed at ordinary boiler pressure, it is now appreciated that greater savings will be possible if a greater proportion can be generated at the higher pressure. In the latest designs the ratio has consequently been altered, the function of the low-pressure boiler becoming mainly that of a preheater and supplying only a very small proportion of the low-pressure steam. TABLE 2.—Comparative Data of Schmidt Multi-Pressure Locomotives. German L.M.S. State Ry.

Railway . Type

.

.

.

.

Fuel. Grate Area

sq. ft.

High- Pressure Boiler — Heating Surface : — Firebox . sq. ft. Heat-transfer elements ,, Superheater ,, Boiler Pressure Ib. per sq. in. Low-Pressure Boiler — Heating Surface : — Flues . sq. ft. Superheater „ Boiler Pressure Ib. per sq. in.

'

4-6-0

4-6-0

N.Y.C.

C.P.R.

P.L.M.

4-8-4

2-10-4

4-8-2

Coal

Coal

Coal

Oil

Coal

26.6

28

70

77

41

217.6

—

422

515

323

426 430

368 274

660

750

407 507

850

900

850

850

850

1,335 355

3,229

3,746

1,680 525

250

250

250

200

1,265 426 205

High-Pressure Cylinder — Diameter . inches Stroke . „

11.4 24.8

11.5 26

13 30

15.5 28

9.5 25.6

Low-Pressure Cylinders — Diameter . inches Stroke . . „

19.7 24.8

18 26

23 30

24 30

22 27.5

Driving Wheel Diameter . inches

78

81

69

63

70.8

Weight of Engine, tons

91

89.3

181

206

115

Tractive Effort.

—

83,300

—

Ib.

33,200 | 66,000

485

1. 2, 4, 7. 8. 9. lo!

Fia. 6.—Winterthur High-Pressure Locomotive. Steam-Drum. 11. Grate. 3. Bottom Drums. 12. Refractory Floor. 5, 6. Water-Walls. 13. Safety-Valve. Water Tubes. 14. Regulator. Stay Tubes. 15. Superheater. Firebox. 16, 18. Non-return Valves, Combustion Chamber. 17. Main Feed-Water Heater.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

113

During last year a locomotive having a boiler of this type was built by the North British Locomotive Company of Glasgow for the London, Midland and Scottish Railway Company, and later on in the year another of these engines was built by Messrs. Henschel and Sons for the Paris, Lyons and Mediterranean Railway of France. Engines having boilers of this type are at present under construction for the Canadian Pacific Railway and the New York Central Railway. Table 2 gives the leading dimensions, so far as at present available, of locomotives equipped with Schmidt high-pressure boilers. Winterthur High-Pressure Locomotive.—The Swiss Locomotive and Machine Company, Winterthur, completed a 2-6-2 tank locomotive about the end of 1927, to the designs of Mr. Buchli, their Chief Engineer, in which the steam is generated in a boiler of the water-tube type, carrying a working pressure of 850 Ib. per sq. in. The general principle is shown illustrated in Fig. 5. The boiler is fed by a compound feed-pump, its steam supply passing through a small superheater in the firebox. The feed-water is preheated by exhaust steam to a temperature of about 180° F., and is then pumped through the feed-heater situated in the front of the boiler, where its temperature is raised to about 450° F. The air to the grate passes through a heater situated at the front of the smokebox. The engine is totally enclosed and has three simple cylinders, the cranks being set at 120°. It is fitted with cam-operated poppet-valves, and is of the " Uniflow " type, having exhaust passages at the centres of the cylinders. The single-seated poppet-valves allow of the use of a higher degree of superheat than would be possible with piston-valves, and the high revolution speed and " Uniflow " principle also assist in preventing excessive temperature drop during expansion. The cam-shaft is operated by means of bevel gearing from the crank-shaft, and different rates of cut-off and reversal are obtained by moving the cam-shaft transversely to bring other cams into engagement. When the locomotive is coasting, the valves can be lifted off their seats to prevent compression. The crank-shaft is geared to the jackshaft by means of a 21/2to 1 flexible reduction gear. The rods transmitting power to the coupled wheels are attached to the centres of the leading coupling-rods, thus limiting the effect of angularity. The boiler is illustrated in Fig. 6. It consists of a top drum which acts as a steam-reservoir, and two small bottom drums. The drums are connected together by the water-space walls. These walls are pierced by a number of short tubes which allow the furnace 8 487

114

HIGH-PRESSURE LOCOMOTIVES.

JAN. 1931.

gases to pass through. The water-tube elements consist of two vertical tubes secured at their lower ends to the two lower drums. Their upper ends are joined by a cross-tube which in turn communicates with the upper drum by means of a T-piece. The space between the rear and centre water-walls is fitted with a grate and forms the firebox. The superheater is situated at the rear of the front section and communicates with the header which is on the engine side of the regulator. The main feed-water heater is situated in front of the superheater, and is fitted with end caps to allow of the removal of scale. Before assembly, the various components of the boiler were subjected to exhaustive experiments in order that adequate factors of safety should be obtained. For example, the water-walls were FIG. 7.—" Winterthur " High-Pressure Locomotive.

subjected to a steam pressure of 2,800 Ib. per sq. in. The tube elements were also subjected to rigorous tests, and the complete boiler was used for some time as a stationary boiler so that the various accessories could be experimented with. A photograph of this locomotive is shown in Fig. 7, and its main particulars are given below :— Boiler pressure : 850 Ib. per sq. in. Grate area : 14.4 sq. ft. Heating surface : 1,040 sq. ft. Heating surface of superheater : 215 sq. ft. Water capacity of boiler : 594 gallons. Number of cylinders : 3. Diameter of cylinders : 81/2inches. Stroke of piston : 13f inches. Ratio of gear-wheel transmission : 1/2.5. Diameter of driving wheels : 60 inches. 488

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

115

Maximum speed : 50 miles per hour. Water tank capacity : 1,320 gallons. Coal tank capacity : 2.7 tons. Weight of locomotive, empty : 68 tons. Weight of locomotive, in working order : 78 tons. Extensive trials have been made with this locomotive in Switzerland, Austria, and France, and during some recent trials a coal consumption of about 21/2Ib. per d.h.p.-hr. has been recorded when delivering 800 h.p. at the drawbar, the corresponding water consumption being slightly under 15 Ib. The calorific value of the coal was 13,500 B.Th.U. per Ib. During the trials it was found that most of the solid matter contained in the feed-water was deposited in the feed-water heater, but some trouble has been experienced by the formation of scale in the boiler tubes. The makers are taking steps which they expect will prevent this in future. It will be seen from the list of dimensions and from Fig. 7 that the locomotive is only of small size : with a larger locomotive it is expected that considerably better results could be obtained. SchwartzJcopff-Lqffler Three-Cylinder Locomotive.—In April 1930 an engine was completed to the order of the German State Railways by the Berlin Machine Works, formerly known as Messrs. L. Schwartzkopff and Company, constructed to the designs of Dr. Lofiler, in which a pressure of 1,700 Ib. per sq. in. is used. This engine is a three-cylinder engine having two high-pressure and one low-pressure cylinder. As will be seen from Fig. 8, the principle on which the boiler is designed is entirely different from that of any locomotive which has previously been built, and whereas in the Schmidt engine distilled water is used as a medium for transmitting the heat from the fire to the water from which steam is generated, in this case the medium of transfer is steam highly superheated in tubes forming the walls of the firebox. Steam in a saturated condition is drawn from the high-pressure drum by means of a circulating pump and is forced through the superheater tubes referred to above, in which it becomes superheated to a temperature of about 900° F., and the pressure raised to approximately 1,700 Ib. per sq. in. About one-quarter of this steam is used for supplying the high-pressure cylinders, and the remaining three-quarters is returned to the high-pressure evaporating drum. This raises the temperature of the water in the drum, and the steam so generated passes again to the circulating pump for superheating as described above. The exhaust steam 489

FIG. 8.—Schwartzkopff-Lojjler Three-cylinder Locomotire.. 1 High-Pressure Drum. 17. Safety-Valve. 36, 39. Non-return Valves, Water. 2! 3, 51. External Steam Supply. 18. Oil-Separator. 37. H.P. Feed-Water Heater. 4 5 Delivery and Outlet Pipes. 19. L.P. Feed-Water Preheater. 40, 41, 42. Supplementary H.P. Feed6,, 6,. Pipes to Circulating Pumps. 20, 21, 22, 23. Low-Pressure Boiler. Water System. 7 1 ,7 2 . Circulating and Feed Pumps. 25. H.P. Feed-Water Tank. 43, 45. Injector and Feed-Pump to L.P. 8.Delivery to H.P. Superheater. 26, 29. L.P. Steam-Pipe. Boiler. 91, 92. H.P. Superheater. 27. L.P. Superheater. 44, 46, 47. Feed-Delivery Pipes, L.P. 10, 12. H.P. Steam-Pipe. 28. L.P. Regulator. Boiler. 11 Non-return Valve, Steam. 30. L.P. Cylinder. 48,49. Air-PreheaterandDuctto Ashpan. 13 H. P Regulator. 31, 32. Valves and Steam-Pipes for driving 50. L.P. Exhaust Pipe. 14. H P Cylinders. Pumps. 52, 53. Supplementary H.P. Steam to 15. Return Pipe to H.P. Drum. 33, 34, 35, 38. Feed-Water Pipes for H.P. L.P. Boiler. 16 H P. Exhaust Pipe. Drum. 54, 55. Emergency Valve and Pipe.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

117

from the high-pressure cylinders passes first through an oil-separator, then through the feed-water heater from which the low-pressure boiler is supplied, and thence through tubes in the low-pressure boiler, in which pressure up to about 225 Ib. per sq. in. is generated. The condensate from the high-pressure exhaust passes to a feedwater tank whence it is pumped back to the high-pressure drum through the high-pressure feed-water heater, any leakage in steam in the high-pressure system being made up from the low-pressure boiler. Steam generated in the low-pressure boiler passes through another superheater in which it is raised to a temperature of about 640° F., from whence it passes to the central low-pressure cylinder, and is finally exhausted up the blast-pipe. In addition to supplying the low-pressure cylinder, the lowpressure boiler supplies steam for operating all the pumps. The pumps are in duplicate, a set being situated on each side of the engine. Each set consists of three pumps :—(1) the,circulating pump for circulating steam from the high-pressure evaporative drum through the superheating tubes forming the firebox ; (2) the pump feeding the high-pressure evaporative drum with the condensate from the high-pressure exhaust ; and (3) the pump feeding the evaporative drum with the make-up water from the low-pressure boiler. Each of the pumps is capable of delivering about 75 per cent of the maximum requirements of the boiler, and absorbs about 4 per cent of the power developed. In order to improve the efficiency of the engine, the air supply to the firebox is preheated to a temperature of 300° F. From an examination of Fig. 8 it will be seen that the heat generated in the firebox is utilized first of all to superheat the steam for the high-pressure cylinder and the evaporative drum. The gases then pass through a nest of tubes raising the temperature of the steam supplied to the low-pressure cylinder, thence through the high-pressure feed-water heater, after that the air-preheater, finally passing up the chimney. The two outside high-pressure cylinders are machined from steel forgings carried in cast-steel brackets bolted to the frame. They are free to expand in this bracket in a longitudinal direction to avoid stresses. Owing to the small size of the cylinders, tail-rods have had to be provided to equalize the power on the forward and backward strokes. Steam is distributed to the cylinders by piston-valves and in all other respects the construction of the motion, frames, etc., is similar to that employed by the German State Railways. The Author has described only the main principles of the LofEer locomotive, but whilst no doubt every effort has been made to utilize 491

118

HIGH-PRESSURE LOCOMOTIVES.

JAN. 1931.

as far as possible the heat generated in the firebox, it must be admitted that this has only been effected by a very intricate and complicated system of tubes and drums. It will be appreciated also that the design involves the use of a very large number of joints and connexions, a failure of any one of which might have serious results. Another serious drawback to this locomotive is that steam can only be raised if a steam supply from another source is available for the initial heating up of the water in the high- and low-pressure steam generators. The pressure in the latter must rise to about 70 Ib. per sq. in. before the pumps can be started, and until then the fire cannot be lit in the grate. It is usually convenient to obtain this steam from FIG. 9.—Schwartzkopff-Loffler

Three-cylinder Locomotive.

another locomotive. The most important feature, however, in connexion with the working of the engine, is the necessity for absolute reliability in the circulating and feed-pumps, and this no doubt is the reason why the designer has thought fit to duplicate this feature in the engine. In the Loftier boiler there is no necessity to use distilled water, ordinary feed-water being quite satisfactory because the concentration and deposit of scale is only possible in the high-pressure evaporative drum which is not heated by either fire or high-temperature circulating tubes as in the case of the Schmidt boiler. Boiler scale will therefore have no injurious effect. Special precautions have been taken to ensure safety as far as possible. Non-return valves are situated in the high-pressure steam 492

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

119

and feed-water pipes so that in the event of a superheater pipe bursting, the high-pressure drum is isolated, and only the steam in the superheater can escape. Safety devices have also been incorporated for rapidly reducing the firebox temperature in case of purnp failure. It is claimed that with this locomotive a fuel economy of 45 per cent will be attained and its overall thermal efficiency will be raised to nearly 18 per cent. It is difficult to reconcile these claims, as an efficiency of 18 per cent is more than double that of an ordinary locomotive. Trial runs have been made with this locomotive but the Author has not been supplied with the actual results obtained. A photograph of this locomotive is given in Fig. 9, and its main dimensions are given below :— Grate area : 26 sq. ft. Heating Surfaces :— High-pressure superheater : 970 sq. ft Low-pressure superheater : 344 sq. ft. High-pressure preheater : 764 sq. ft. Low-pressure boiler : 882 sq. ft. Boiler Pressures :— High-pressure boiler : 1,700 Ib. per sq. in. Low-pressure boiler : 215 Ib. per sq. in. Motion :— Diameter of high-pressure cylinders :85/8inches. Stroke of high-pressure cylinders : 26 inches. Diameter of low-pressure cylinder : 235/8inches. Stroke of low-pressure cylinder : 26 inches. Diameter of driving wheels : 78| inches. Diameter of bogie wheels :331/2inches. Diameter of trailing wheels : 49-1/4 inches. Weight empty (calculated) : 110 tons. Weight in working order (calculated) : 113 tons. Adhesive wieght: 59 tons. Four-Cylinder High-Pressure Compound Locomotive, London and North Eastern Railway.—The first British high-pressure locomotive to be built is the four-cylinder compound engine of the London and North Eastern Railway, which was completed at the end of 1929. Unlike his Continental colleagues, the Author, in designing this engine, thought it advisable to be content with what may be regarded as only a moderate increase in boiler pressure to 450 Ib. per sq. in. In coming to this decision he was largely influenced by consideration

493

120

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

FIG. 10.—Four-Cylinder High-Pressure Compound Locomotive, London and North Eastern Railway. (See Folding Plate opposite).

LEADING DIMENSIONS AND RATIOS. Grate— Length on slope 7 ft. 6 in. Width . . . 4 ft. 8 in. 34.95 sq. ft. Grate area . Boiler— Steam-drum :— Length . . 27 ft.115/8in. Diam. inside 3 ft. 0 in. Water drums, forward :— Length . . 13 ft. 53/4in. Diam. inside 1 ft. 7 in. Water drums, back :—• Length . . 11 ft. O5/8 in. Diam. inside 1 ft. 6 in. Smokebox length 16 ft. 1 in. Working pressure, 450 Ib.per sq.in. Tubes— Small :— Number . . 444 Diam. outside 2 in. Combustion Chamber :— Number . 74 Diam. outside 21/2in. Firebox :— Number .

.

Diam. outside Heating Surface— Firebox . . . Combustion Chamber . . Small Tubes . Total evaporative Superheater— Number of elements . Diam. inside . Heating Surface Total Heating Surface Axles— Journals :— Bogie Coupled wheels Carrying „ Bissel ,, Crank-pins— Outside . Inside Coupling Pins— Leading Driving . Trailing

494

250

21/2in. 919 sq. ft. 195 ,, 872 „ 1,986 ,, 12 1-18 in. 140 sq. ft.

2,126 sq. ft. 61/2in. Xll in. 91/2in. Xll in. 6 in. X ll in. X11in. X ll in. 51/2in. X 6 in. <6 in.x5£in. 4 in. X 43/8in. 6 in. X 41/4in. 4 in. X 41/4in.

Springs— Bogie wheels :— laminated. Type . . Length . 2 ft. 6 in. centres. Number of plates . . 8 Size of plates,41/2in.wide X 7/16-in. Coupled wheels laminated. Type. . . Length . 3 ft. 6 in. centres. Number of plates 11 Size of plates 5 in. wide X5/8-in Carrying wheels laminated Type . . Length . 4 ft. 6 in. centres. Number of plates . 10 Size of plates 5 in. wide X 5/8-in. Bissel:— Type . . helical 51/2in. Outside diam. in. Free length 10 3/16 Cylinders— High-pressure Low-pressure Motion—• Type .

.

(2) 10 in. X 26 in. (2) 20 in. X 26 in.

.

Gresley-Walschaert.

Valves :— Type . . piston. 6 in. Diam. H.P. „ L.P. 8 in. Max. travel, H.P. . . 89/16in. Max. travel, L.P. . . 611/16in. Steam lap, H.P. . . 1* in. Steam lap, L.P. . . 15/8 in. nil. Exhaust lap Cut-off in full gear, H.P. 90 per cent. Cut off in full gear, L.P. 75 per cent. Tractive Effort . 32,000 Ib. Total Adhesive 140,000 Ib. Weight Adhesive Weight % Tractive 4.37 Effort . . steam brake and Brakes vacuum ejector.

This page intentionally left blank

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES,

121

of maintenance costs and the desirability of advancing by stages. Past experience of revolutionary designs has been that the spectacular advancements have not always been justified by results, and consequently the Author deemed it wiser to seek progress on a less ambitious scale. He also recognized that as the pressure increases the economies to be expected in fuel consumption are in a diminishing ratio. He decided to adopt a boiler of the water-tube type, in view of the successful application of such boilers to high pressures in marine practice and in large power stations. In September 1924. he accordingly approached Mr. Harold Yarrow, M.I.Mech.E., of Glasgow, whose firm are so well known as designers and builders of water-tube boilers, and suggested to him a design of boiler of the water-tube type which might be applied to locomotives. A water-tube boiler suitable for a locomotive involves a radical departure from the usual design of such boilers for marine and land purposes, and upwards of three years of work on the part of Mr. Yarrow and the Author resulted in the completion towards the end of 1927 of the final design, which was patented in their joint names. Early in 1928 an order was placed with Messrs. Yarrow to proceed with the construction of the boiler, which was completed and tested in October 1929. The engine was built at the Darlington Works of the L. & N.E.R., and ran its trial trip on 12th December 1929. The drums of the boiler are of sufficient diameter to allow a man to get inside them for the purpose of expanding the tubes. To suit the conditions peculiar to a locomotive, it was felt that tubes of a large diameter only should be used, the tubes in the firebox end of the boiler being21/2inches external diameter and the tubes in the forward part 2 inches. Fig. 10 gives a sectional elevation and cross-section of the boiler and shows its general form of construction. The considerations which govern the design of marine and land boilers are so entirely different from those required in a locomotive boiler, that there is very little similarity between the boiler used on this engine and the ordinary type of water-tube boiler, as will be seen from the photograph, Fig. 11. In the ordinary water-tube boiler resting on foundations, the boiler can expand freely in any direction, and the tubes, not being subjected to vibration and racking stresses, are not liable to leak. In a locomotive, the boiler must be so secured to the frame that in addition to standing the shocks and vibrations consequent upon the engine running at high speed upon a railway, it must be capable of withstanding the shocks which occur when a locomotive is shunting, or comes in contact with buffer stops, or possibly becomes derailed. 497

FIG. 11.—Boiler.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

123

It is necessary also to have due regard to the fact that the boiler provides an important structural element in the construction of a locomotive and contributes to its rigidity. It will be seen from the diagram that in this boiler the large steam-drum forms the backbone of the boiler, from which the tubes and the small drums depend. At the forward end this drum is carried in a cast-steel cradle into which it is firmly secured by large strap bolts, and any fore-and-aft movement is entirely prevented by stops which are machined on the lower side of the drum. The cradle in turn is secured to the engine frame by 1-inch steel plates extending downwards inside the main engine frames, to which they are securely riveted. The steel drum must be free to expand longitudinally, and accordingly the back end is secured to the top of a triangular-shaped transverse plate which in turn is secured at its lower extremity to the engine frame. The drum therefore, whilst being free to expand longitudinally, is constrained from side or vertical motion. The four smaller drums are not supported, but hang from the water-tubes. Rectangular lugs are riveted to the lower side of each of these drums at their forward and back ends. These rectangular lugs are free to slide longitudinally in grooved castings which are secured to the engine frame, but as they do not reach the bottom of the grooves, they are also free to move vertically. These lugs and grooves are provided to restrain the drums from any side movement. It was considered that this was the best way of preventing the transmission of vibration and shocks to the tubes in such a way as might cause leakage. It is interesting to note that this method of construction has been completely successful, in that there has been no case of the slightest leakage occurring at any of the 1,536 points at which the tubes have been expanded in the drums. In adopting a water-tube boiler, the Author was not unmindful of the troubles which might reasonably be expected to result from scale formation in the tubes. Unlike marine and power-house boilers, in which the boiler feed is derived from the condensate and only a very small make-up of fresh water is required, the locomotive boiler requires 100 per cent of make-up water. In order to prevent the formation of scale as far as possible, the Author decided to introduce the feed-water at the highest possible temperature. It will be seen from the diagram that the feed-water is introduced into a chamber at the forward end of the top drum in front of the water-tubes and separated from the evaporative portion of the boiler by a weir at a height of about half its diameter. The water is supplied from the tender by means of two ordinary injectors, and is delivered into this forward space of the top drum, 499

124

HIGH-PRESSURE LOCOMOTIVES.

JAN. 1931.

after passing through a form of injector heater. The latter has two sets of cones in which the injector action is repeated by steam from the steam space in the boiler. Heat is absorbed to such an extent by the feed-water that its temperature when delivered into the water chamber is over 400° F. and is therefore only about 50° less than that of the saturated steam in the boiler. Much of the scale and mud is consequently thrown down in the forward portion of the top drum. But for this arrangement for dealing with the feed-water, considerable trouble due to the formation of scale would have to be expected; as it is, after running some 15,000 miles, a slight deposit of hard scale was found on the inner rows of tubes in the firebox. This was at first not easy to detect, and in consequence about half-adozen tubes on each side of the firebox adjacent to the brick arch showed signs of overheating and had to be changed. The experience gained has resulted in the development of an apparatus by which the hardest scale can be readily cut out without damage to the tubes, and this will probably be necessary when the engine has run some 10,000 to 15,000 miles, according to the class of feed-water used. On the other hand, experience has also shown that this boiler can be worked for a much longer period than the ordinary type of locomotive boiler before requiring to be washed out. When stationed at Gateshead the locomotive worked express trains from York to Edinburgh and back, involving a daily run of about 420 miles. Whilst other engines of the " Pacific " type in the same " link " require washing out after running 1,000 to 1,500 miles, this engine ran 5,000 miles without washing out, and when opened up it was found that the boiler was exceptionally clean and the tubes were in good condition. It was, however, found that the accumulation of mud and scale, which usually occurs above the foundation ring in an ordinary boiler, had fallen into the drums below the grate on each side of the firebox. This material remained in the form of mud in the drums, from which it was easily removed, whereas if it had- been in the sides of an ordinary locomotive firebox, much of it would have formed into scale. Other means, which will be described later, have been recently embodied, by which the temperature of the water when delivered in the feed-water space in the top drum is further increased before it passes to the evaporative portion of the boiler. In a test which was carried out at the maker's works, an evaporation of 20,000 Ib. per hour at 450 Ib. per sq. in. was maintained for a period of four hours by the introduction of a steam jet up the chimney. This high rate of evaporation is possible owing to such a large proportion of the heating surface available being subject to 500

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

125

FIG. 12.—Engine Cab.

501

FIG. 13.—Four-Cylinder High-Pressure Locomotive, L. & N.E.R.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

127

direct radiant heat. In the ordinary form of locomotive boiler only the firebox is subject to radiant heat, and the evaporation per square foot of heating surface of the tubes is only about one-fifth of that of the firebox sides. The superheater elements in this boiler are located in the forward portion of the central flue and are also subject to radiant heat. In order to prevent the flame impinging directly on the ends of the elements, a brick column is provided in the centre of the main flue immediately in front of the brick arch. Notwithstanding this precaution, and owing to the fact that there were no data available as to the effect of radiant heat on superheater elements, the temperature to which the steam was superheated during the preliminary trials was excessive, temperatures of 900° F. being obtained; consequently the lengths and area of the superheater elements have been reduced so that a temperature of approximately 700° F. can now be obtained, and this is regarded as sufficient. The superheater elements are situated between the boiler and the regulator and are, therefore, always subject to full boiler pressure. In order to prevent overheating when the regulator is closed, the steam supplied for auxiliary services is taken from the superheater and passes through a coil of ribbed pipes laid in the feed-water chamber, thus raising further the temperature of the boiler feed and at the same time de-superheating the steam. The de-superheated steam is led to the reducing valve where its pressure is reduced to 200 Ib. per sq. in. Steam from the reducing valve supplies a manifold pipe on the footplate across the front of the boiler above the firehole door. From the manifold pipe steam at 200 Ib. per sq. in. pressure is taken for supplying all the auxiliary services, such as the injector, vacuum and steam brake, reversing gear, steam-sanding, steam-heating, whistle, and turbo-generator. It has been possible to retain the standard steam fittings for all these purposes. Fig. 12 shows the arrangement of these fittings. With a pressure of even 450 Ib. per sq. in. special designs of boiler fittings and valves have to be used owing to the cutting action of high-pressure steam, but in this engine only the safety-valves, regulators, and water-gauges have had to be made suitable for this pressure. After the hot gases have passed between the water-tubes they enter the flues located at each side of the boiler. Naturally the walls of these flues are very hot, and in order to reduce their temperature and to make effective use of this heat, which otherwise would be wasted, the boiler is surrounded by an air space lying between the casing of the flues and the insulated outer clothing. All the air 503

128

HIGH-PRESSURE LOCOMOTIVES.

FIG. 14.—Front view of Engine.

504

JAN. 1931.

JAN. 1931.

HIGH-PRESSURE LOCOMOTIVES.

129

required for combustion traverses this air space from the intake at the front of the engine to the ashpan, and in so doing its temperature is raised to about 250° F. It has been found that even when working hard it has not been necessary to supplement this supply of hot air by opening the ashpan door. Preheating the whole of the air supply is bound to augment the thermal efficiency of the boiler. The only other feature of the boiler which calls for comment is the construction of the front end and chimney. Incidentally, the apparent absence of the chimney has caused more public comment than any other feature of the locomotive. (Side and front views are shown in Figs. 13 and 14.) In order to provide sufficient length for the water-tubes, it was necessary to have the top steam-drum as high as the limits imposed by the load gauge permit, consequently there was no room for a chimney of the conventional type. Engines having large high-pitched boilers can only have very short chimneys, and trouble has been experienced in such engines owing to smoke and steam beating down on the front windows of the cab and interfering with the driver's view of signals. The Author enlisted the assistance of Professor W. B. Dalby and constructed a wooden model of an engine of such a type. This model was placed in an air flume, and powdered chalk was blown up the chimney at the same time as a current of air was drawn through the flume at 50 m.p.h. Observations through a glass window showed the course pursued by the powdered chalk, and as a result of various modifications, the design finally adopted was arrived at. In this design the whole of the powdered chalk was lifted sufficiently to clear the cab windows, and it is satisfactory to record that in actual service the smoke and steam, whether running at slow or at high speeds, is deflected upwards sufficiently to clear the cab, and in no way obstructs the driver's vision. As originally built the engine had two high-pressure cylinders 12 inches in diameter, and two low-pressure cylinders each 20 inches in diameter, all cylinders having 26 inches stroke. It has been found that by reducing the diameter of the high-pressure cylinders to 10 inches a more equal distribution of work between the high-pressure and low-pressure cylinders results. The Author regarded it as necessary, whilst having only two sets of valve-gear, to be able to vary the cut-off of the high-pressure cylinder independently of that of the low-pressure cylinder. He felt that only by trial at varying cut-offs could the best results be realized. He therefore arranged that in the rocking link by which the highpressure valve is actuated, provision should be made by means of a slot and die-block to vary the travel of the valve, at the same time 9 505

FIG. 15.—Motion Arrangement.

132

HIGH-PEESSURE LOCOMOTIVES.

FIG. 16.—High-Pressure Cylinders. 10 inches diameter ; 26 inches stroke. Section through E.F.

508

JAN. 1931.

JAN. 1931.

HIGH-FEESSURE LOCOMOTIVES.

133

retaining the combination lever to keep the lead constant. The motion arrangement for both high- and low-pressure cylinders is shown in Fig. 15. The reversal of the low-pressure valve-gear, and consequently of the whole engine, is actuated by the ordinary form of steam-reverser, and a similar equipment is provided to vary the high-pressure cut-oil. Both these equipments are attached to the shafts they actuate, and being so remote from the footplate, their delicate control was not easy. This has been successfully effected by the use of telemotors ; another telemotor is also provided for operating cylinder cocks. The high-pressure cylinders, steam-ports, passages, and low-pressure steam-receiver are formed by a single steel casting as illustrated in Fig. 16, the cylinders and steam-chests having special close-grained cast-iron liners. Various minor alterations and adjustments have had to be made since the engine was built, but with the exception of reducing the diameter of the high-pressure cylinders and reducing the heating surface of the superheater, it has not been necessary to alter any of its main features. The locomotive has worked trains of over 500 tons' weight for long distances at express speeds with consistent success and reliability, and although it has not been possible so far to carry out any extensive trials, there is every indication that it will prove more economical in fuel consumption than express engines of the latest normal types. Any economy effected in maintenance cost will only become fully apparent after the engine has run a few years. It has been ascertained that the cost of a water-tube boiler similar to that fitted on this engine will not be appreciably greater than that of the ordinary wide firebox type as fitted on " Pacific " engines. The most expensive components of the water-tube boiler are the solid-forged steam- and water-drums. These are not subjected to the action of the fire, and consequently may be expected to have a long life. On the wide firebox type of ordinary boiler the copper firebox is the most costly section, and it is well known that its life is short and its renewal is an expensive item. Again, in the ordinary type of locomotive boiler tubes and firebox stays are sources of trouble involving costly maintenance and occasional failures. In the design of boiler under consideration there are no stays ; the tubes are more effectively secured, and are not subjected to variation in temperature and stress at the points where they enter the drums. In conclusion, the Author submits that, with the moderately high pressure and simple design which he has adopted, economy both in fuel and maintenance costs will be secured and at the same time the reliability so characteristic of British locomotives will be fully maintained. 509

FIG. 17.—Arrangement of Injector Heater.

This page intentionally left blank

Ernest Rutherford FRS Rutherford of Nelson

(1871-1937),

Baron

A New Zealander of Scottish extraction, Rutherford's father was an engineer and flax-miller. Shortly after graduation from the University of Canterbury, Christchurch, he began experimental researches in magnetism. Significant as this work was, Rutherford's scientific career started in earnest from 1895 after winning a scholarship to Trinity College, Cambridge. Like Bragg, Rutherford was fascinated by the discovery of X-rays. His first work at Cambridge was on electromagnetic wave transmission, but he shifted to examining the nature of radiation. From his experimental results he concluded that X-rays were different from uranium emissions, and that the latter were not uniform, having distinct components which he called alpha and beta radiation. Continuing his thinking as professor of physics at McGill University in Canada, Rutherford reached the startling conclusion that these phenomena accompanied the transformation of matter. His disintegration theory was incorporated into the book Radio-activity (1904). Rutherford travelled to Manchester in 1907, commencing a great "middleperiod" of scientific discovery. This centred on investigations into the alpha particle and included the determination of Avogadro's number (the number of molecules per unit volume of gas). The main outcome of his experimental work was the gradual theorizing of the nuclear structure of the atom and, consequently, the properties of elements. Interrupted by war, Rutherford managed to make one further breakthrough in 1919. In bombarding nitrogen atoms with alpha particles, the physicist recorded the liberation of hydrogen nuclei - evidence of transmutation of matter. Rutherford had "broken" an atom. Rutherford returned to Cambridge as Cavendish Professor of Physics in succession to his old teacher J J Thomson. He had already been honoured with the Nobel Prize for Chemistry (1908) and became President of the Royal Society in 1925. However, whilst his own personal researches continued, he also collaborated with other leading scientists, such as Frederick Soddy and Niels Bohr. His students amounted to a Who's Who of contemporary science. With Rutherford's encouragement, these younger men led the way into new areas of subatomic physics. Thus, Rutherford's prediction of the existence of the neutron was confirmed by James Chadwick. J D Cockcroft was another research student, extending Rutherford's influence into the later, commercial exploitation of the atom.

513

183

Nov. 1932.

THE NINETEENTH

THOMAS HAWKSLBY LECTURE.

ATOMIC PROJECTILES AND THEIR APPLICATIONS. BY PROFESSOR LORD RUTHERFORD, O.M., M.A., LL.D., D.Sc., F.R.S.

For my Lecture to-day I have selected a topic for discussion which I hope may prove of some general interest to the members of the Institution, even if it has only a remote bearing on their professional activities. I propose to deal this evening in a general way with the methods adopted to obtain high-speed atomic projectiles and their application to the solution of important physical problems. This age will be for ever memorable for the development of new and swift methods of transport over land, sea, and in the air. The spectacular advances in the speed of the motor-car, and motor-boat, and still more of the flying machine, are familiar to you all and are triumphs of which the engineer may well be proud. It is of interest to consider the maximum speeds that can be communicated to matter in bulk and to compare them with the speed of the atomic projectiles used every day by the physicist in his experiments. The maximum velocity that can be communicated to the rim of a suitably shaped rotating steel disk is controlled by the tensile strength of the material and does not exceed at the best 2 miles per second. The peripheral speeds used in practice are much lower. For example, if it is to be efficient, the tip of an aeroplane screw should not travel faster than the velocity of sound, or about one-fifth of a mile per second. We shall next consider the velocities that can be communicated to matter by the use of explosives. The maximum speed of the rifle bullet under the best conditions is about half a mile per second, while the shell from " Big Bertha " which bombarded Paris had a

[THE I.Mech.E.] 515

184

ATOMIC PROJECTILES.

Nov. 1932.

maximum range of about 80 miles and a muzzle velocity of about 1 mile per second. No doubt the speed which can be communicated to projectiles by such methods will be exceeded in the future, and much higher speeds may be ultimately obtained by the development of a rocket type of apparatus. We may take it, however, that at present the maximum velocity that can be communicated to matter in bulk is not more than 2 miles per second. It is of interest to note that this is of the same order of magnitude as the average speed of the molecules of gases under ordinary conditions. The average velocity of agitation of the molecules of oxygen gas at normal temperature is about 1/4-mile per second and of hydrogen molecules about 1 mile per second. The molecular speed varies as the square root of the absolute temperature, so that even at a temperature of 7,000° C. the average speed for an undissociated gas only increases by a factor of 5 times. No doubt if we were able to pass to the interior of a hot star where the temperature may be of the order of 100 million degrees, the average velocity of agitation of the hydrogen atoms reaches high values, of the order of 1,000 miles a second, and thus becomes comparable with the velocities acquired by charged atoms of hydrogen in an ordinary discharge tube. While in our normal experience we rarely obtain or observe velocities of matter greater than a mile per second, yet occasionally we have evidence of the destructive effects of large masses of matter moving at much higher speeds. For example, it is estimated that the great Siberian meteorite which fell in 1908 had a mass of about 100 tons and reached the earth with a speed of 40 to 50 miles a second. The conversion of its kinetic energy into heat caused a pressure wave which destroyed the forest for a distance of about 40 miles from its centre, and was of such a magnitude that it was clearly recorded in this country.* Such a spectacular but fortunately rare event brings home to us the enormous energy of motion resident in large masses of matter moving at high speeds. Even if large sources of energy are available, our experience has shown that it will be difficult to set matter in motion with a speed of more than a few miles per second. Yet if we pass from matter in bulk to the individual atoms of matter, the situation is very different, for we shall see that simple methods have been developed in the laboratory to produce atomic projectiles moving with enormous speed. This has depended on producing charged atoms or molecules of matter by means of electrical discharges or other appropriate agencies and accelerating them by strong electric fields in a good vacuum. * See F. J. W, Whipple, Roy. Met. Soc., July 1930, 516

Nov. 1932.

ATOMIC PROJECTILES.

185

As you know, the passage of an electric discharge from an induction coil through a gas at low pressure—of the order of 1/100 mm. of mercury—leads to the liberation of electrons and the production of charged atoms and sometimes charged complex molecules which acquire energy of motion in their passage through the electric field. In this way we can obtain a copious supply of fast electrons and swiftly moving atoms of matter. The maximum speed that can be acquired by the charged particles can be calculated in a very simple way. Suppose the particle of mass m and charge e moves freely in a vacuum through a difference of potential V. The energy acquired by the particle in moving through the electric field is Ve and this is a measure of the kinetic energy acquired by the particle. For speeds small compared with the velocity of light so that the mass m remains nearly constant, the kinetic energy1/2mu2==Ve or u =

If the charge and mass of

the particle are known, the velocity acquired by the particle can at once be calculated. It will be seen that the velocity acquired for a given charge varies as the square root of the voltage applied, and inversely as the square root of the mass of the particle. In this way it can be shown that an electron moving freely from rest through a difference of potential of 1 volt acquires the great velocity of 590 km. per sec. Since the charge carried by the atom of hydrogen is numerically equal to that of the electron but its mass is 1,840 times greater, the velocity acquired by the hydrogen atom under the same conditions is 13-8 km. per sec. For a singly charged atom of mercury of mass 200 times that of hydrogen the speed is 1 km. per sec. If the particles fall through a difference of potential of 100 volts instead of 1, the speeds are all increased by a factor of 10. When the speed of the particle becomes comparable with that of light, we have to take into account the change of mass of the particle with speed. This mass for any speed is given by the relativity formula

where m0 is the mass of the particle for slow speeds and b = u/c, the ratio of the velocity of the particle to the velocity of light. Under such conditions, the energy of motion E acquired by the particle is given by the relativity formula

517

Energy in electron-volts.

m m0

103

109

106

Electron

1.980 x 10-3

6.261 X 10-2

9.411 x 10-1

0.99999987

H+

4.618 X 10-5

1.461 X 10-3

4.618 X 10-2

0.875

Hg+

3.273 x 10-6

1.034 X 10-4

3.273 X 10-3

0.103

Electron

1 + 1.960 X 10 -6

1 + 1.960 x 10-3

1 + 1.960

1 + 1.959 X 108

H+

1 + 1.066 X 10-9

1 + 1.066 x 10-6

1 + 1.066 x 10-3

1 + 1.066

Hg+

1 + 5.354 x 10-12

1 + 5.354 X 10-9

1 + 5.354 x 10-6

1 + 5.354 x 10-3

3.371

1.068 X 102

4.740 x 103

3.336 X 106

H+

1.445 X 102

4.569 X 103

1.447 x 105

5.656 X 106

Hg+

2.042 x 103

6.456 X 104

2.042 x 106

6.478 X 107

Electron Hp

1

Nov. 1932,

ATOMIC PROJECTILES'-.

187

For the study of many physical problems, streams of charged particles of appropriate speed, generally produced by the acceleration of the particles in an electric field, are often required. For brevity it is very convenient, for example, to speak of a thousand-volt electron or proton or other charged particle, meaning thereby that the particle has the speed and energy equal to that gained in passing freely between two points differing in potential by a thousand volts. Similarly, it is convenient to express the energy of the electron or other particle in terms of electron-volts. Methods have been devised in the laboratory to produce charged particles varying in energy between a fraction of a volt and one million volts. By the use of the swift a- and b-particles spontaneously expelled from radioactive substances, this range of energy is increased to 10 million electronvolts, while we have direct evidence that ultra-penetrating radiation in our atmosphere gives rise to particles of energy reaching as high as 1,000 million electron-volts. In order to illustrate the velocities acquired by a particle for different voltages and the changes of mass with velocity, a Table is given on page 186 for three typical projectiles, namely, an electron, a proton (H + ), and a mercury atom (Hg+), each of which carries the unit charge (4.77 X 10-10 electrostatic units). The velocity acquired by a particle is conveniently expressed by b — u/c, the ratio of the velocity of the particle to the velocity of light c (3 X 1010 cm. per sec.). The ratio m/m0 expresses the change of mass of the particle with speed. In our investigations we have to deal with particles over a very wide range of speed, and for electrons of high energy like those observed in the ultra-penetrating radiation in our atmosphere, the velocity acquired approaches exceedingly closely to the velocity of light—the limiting velocity of any particle on existing views. Even for a million-volt electron, the value of b is already close to unity. The mass of the electron shows great variations over the range of velocities considered. The increase of mass is quite marked for a 106 volt electron, and becomes very large for a 109 volt electron, exceeding even the mass of the proton at slow speeds. The relative changes of mass over the same range of energy are much less marked for the proton and are hardly appreciable for the massive Hg+ atom. A charged particle can be deflected from its path both by a magnetic and electric field. If the particle is acted on by a uniform magnetic field H perpendicular to its direction of motion, it describes a circle of radius p where Up = mu/e. As the deflexion by a magnetic field is a convenient method for directly measuring the velocity of a known particle, the values of Hp for particles of different energies 519

188

ATOMIC PROJECTILES.

Nov. 1932.

are included in the Table. In a uniform magnetic field of 10,000 gauss, a 1,000 volt electron is bent into a circle of a tenth of a millimetre radius while a 106 volt electron describes a circle

p = 047 cm. For a 109 volt electron p = 334 cm. so that the curvature of the path of the particle in the magnetic field would be slight and not easy to measure under practical conditions. The difference in the curvature of the path of the electron and proton for equal energies is very marked for slow speeds, but tends to 520

Nov. 1932.

ATOMIC PROJECTILES.

189

equality for very high speeds. For example, the values of p for a 109 volt electron and proton only differ by a factor of 2. Production of Fast Particles.—While the same general method of acceleration in an electric field is employed to produce swift particles of all kinds, the practical arrangements for the purpose are naturally dependent to some extent on the method of production of the charged particles and the magnitude of the accelerating voltage. A copious stream of electrons of appropriate speed is most simply obtained by heating a tungsten wire to incandescence in a highly exhausted vacuum tube and applying a suitable difference of potential between the heated filament and a neighbouring electrode. To obtain a source of high-speed protons or other charged atoms of matter approximately homogeneous in velocity, a somewhat different arrangement is necessary. This is diagrammatically illustrated in Fig. 1. The apparatus consists of two tubes, the discharge tube A separated by an opening from the accelerating tube B. If a supply of fast protons is required, a stream of hydrogen at low pressure is continuously passed into A and escapes through a tubular opening into the tube B where it is rapidly removed by fast pumps of the diffusion type. The size of the opening is so adjusted that a convenient pressure of hydrogen is maintained in A while the tube B is kept at such a good vacuum by the pumps that a high accelerating voltage V can be safely applied without an electric discharge passing. A supply of protons is obtained by the electric discharge through A produced by a suitable difference of potential V1. Some of these protons pass through the opening where they are further accelerated in the tube B. Since the hydrogen atoms may acquire their charge at any point of the discharge tube, the protons escaping have not all the same speed, so that the energy of the individual protons arriving at the lower end of the tube B may vary between Ve and (V + V1)e. If V1 is small compared with V, the stream of accelerated protons is, however, nearly homogeneous in velocity. It should be mentioned that not only protons (H + ) but also positively charged hydrogen molecules (H 2 + ) are produced by the discharge and pass into the tube B. The H2+ particles will have the same maximum energy as the H+ particles, but only 1/ 2 times their velocity. By passing appropriate gases or vapours into the discharge tube, it is obviously possible to produce streams of charged atoms of various kinds which can then be accelerated to the degree required. In general, a difference of potential of 50,000 volts applied to A gives a convenient supply of protons for experimental purposes. In the experiments of Cockcroft and Walton on the artificial transmutation 521

190

ATOMIC PROJECTILES.

Nov. 1932.

of matter, to be referred to later, a steady difference of potential up to 600,000 volts could be maintained in the tube B. By this method, it was not difficult to obtain a stream of swift protons corresponding to a current of 20 microamperes. The corresponding number of protons produced per second is very large and exceeds the number of a-particles emitted per second from more than 1,000 grams of radium —a quantity greater than the world's supply. No doubt by special devices, far more intense streams of protons could be obtained for special experiments. Numerous experiments are in progress throughout the world to develop methods of production of high-speed particles for experimental purposes. Coolidge was one of the first workers in this field and succeeded in obtaining a stream of swift electrons of energy approaching one million volts. Subsequent experimenters have mainly concentrated on methods of obtaining swift protons or other atoms to be used in investigations on the transmutation of matter by atomic bombardment. Various methods have been used to obtain the requisite high potentials including induction coils, transformers, Tesla coils, and special types of electrostatic machines. There is no insuperable difficulty in devising tubes to withstand an accelerating voltage of several million volts, but the size of the apparatus and the size of the laboratory to contain it, not to mention the cost, increase rapidly with the voltage. It would be a great advantage for many investigations to obtain streams of atoms of energy as high as 10 or 20 million electron-volts. For this reason, experiments have been made to devise methods of accelerating atoms which do not involve the almost brutal method of generating very high potentials to be applied directly to a vacuum tube. A very interesting method has been tried by Lawrence and Livingston of the University of California. This depends on the multiple acceleration of charged atoms so that high-speed particles can be obtained with the aid of comparatively low voltages. By using a high-frequency radio-generator, a rapidly oscillating electric field is applied within a highly exhausted vessel placed in a uniform strong magnetic field produced by a large electromagnet. The protons or other ions to be accelerated move perpendicularly to the magDetic field and describe nearly circular orbits. The time for traversal of a complete circular path by the charged ion is independent of the velocity of the particle and of the radius of its path. The strength of the magnetic field or the frequency of the oscillating field is so adjusted that the particle is regularly accelerated during each rotation. As the speed of the particle increases, the radius of its circular path increases. In a fairly good vacuum, it should be 522

Nov. 1932.

ATOMIC PROJECTILES.

191

possible for the particles to make several hundred complete rotations without serious retardation or scattering due to collision with the residual gas molecules in their path. By this ingenious method, Lawrence and Livingston state that they have been able to obtain a stream of protons of energy as high as 1,200,000 volts by the use of a voltage as low as 4,000 volts. They consider that it should be possible by a development of this method to obtain a reasonable stream of protons of energy as high as 10 or 20 million volts. It seems likely that devices of this or similar kind will have to be utilized in the laboratory if we wish to obtain projectiles for experimental purposes much more energetic than can be produced by a million volts applied directly to a vacuum tube. The development of these new methods is thus of great interest and importance for the future. Applications.—-The study of the effects of collision of atomic projectiles with the atoms or molecules in their path has proved of great importance to the advance of physics in many directions. Time does not allow me to refer more than briefly to the more important results obtained by these methods. If a stream of electrons of a few volts energy passes into a gas at low pressure, there are occasional collisions of the electrons with the atoms or molecules which are elastic in character. When the energy of the electrons exceeds a certain value, some of the collisions are inelastic and the colliding electron may lose part or the whole of its energy either in exciting the atom or in removing one of its outer electrons and thus ionizing it. By the use of this general method, we have been able to determine the energy required to ionize the atoms or molecules, and to determine the energy required to excite the atom in order to emit some of its characteristic radiations. Experiments of this kind have proved of great importance in giving us information as to the structure of the outer atom and particularly of the values of the energy levels which are so closely connected with the explanation of the complicated light and X-ray spectra of the elements. Since charged particles are deflected both by a magnetic and an electric field, it has been found possible to devise arrangements to bring to an approximate focus all particles of the same speed in a diverging beam of atomic projectiles. The well-known focusing method, in which the particles are bent into a semicircle in a uniform magnetic field, has proved very useful in many fields of work including the study of isotopes and the detailed examination of the homogeneous groups of a- or b-particles which have proved of such importance not only in fixing the frequencies of the radiations from the nucleus in the 523

192

ATOMIC PROJECTILES.

Nov. 1932.

form of g-rays, but also in throwing light on the mode of origin of these penetrating radiations. When an electric discharge is passed through a complex gas or vapour at low pressure, in general, as we have seen, a variety of kinds of charged atoms and molecules are formed. In the hands of Sir J. J, Thomson, a method of chemical analysis of these flying particles has been perfected, depending on the variation of the deflexion of the particles of different charge and mass when acted on by a combined electric and magnetic field. This general method of analysis has been carried further by Aston, who devised a " mass spectrograph " in which, by a suitable combination of an electric and magnetic field, particles of the same mass but different velocities can be brought to a sharp focus on a photographic plate. In this way, he has been able to bring out in a striking way the complexity of the chemical elements and the relative intensities of the isotopes of which they are composed. In addition, he has been able to compare the relative masses of the individual isotopes with great precision. This new method of atomic analysis has proved invaluable, not only in throwing new light on the constitution of the elements, but in fixing the chemical atomic weights with an accuracy difficult to realize by the older chemical methods. Space does not allow me to enter into the very interesting question of the scattering of electrons by collision with the atoms of matter and the laws that govern the absorption of swift atomic projectiles in their passage through matter. This has opened up a wide and important field of work involving the study of projectiles of different masses and charges over a very wide range of speed. The electron or proton of a few volts energy is very easily stopped by the thinnest sheet of matter while the fast electron or proton connected with the ultra-penetrating rays may readily pass through several feet of lead. In the latter part of my Lecture, I shall confine myself to a consideration of the way in which swift a-particles have been used to throw light on the dimensions of the atomic nucleus and on the laws of force which hold in its neighbourhood. This will be followed by an account of experiments on the transmutation of matter, which has been effected by the bombardment of matter by swift atomic projectiles of different kinds. Atomic Collisions.—On modern views, all the atoms of the elements have a similar type of electrical structure consisting of a positively charged nucleus of minute dimensions surrounded at a distance by a distribution of electrons in rapid motion. Practically all the mass of the atom resides in this minute nucleus and the ordinary physical 524

Nov. 1932.

ATOMIC PBOJEOTILES.

193

and chemical properties of the atom apart from its mass are controlled by the magnitude of the nuclear charge. The swift a-particles which are expelled from radioactive substances are known to be helium nuclei of mass 4, and may escape with an energy of motion as high as 10 million electron-volts. Such swift a-particles are able to pass freely through the outer electronic structure of the atom and in general the tracks of these particles through a gas are nearly straight. Occasionally, however, one of these particles approaches so closely to the nucleus of an atom in its path that it comes into violent collision with it and may be deflected through a large angle. On account of the minute dimensions of atomic nuclei such encounters are rare, and it may be that only one a-particle in 1,000 or 10,000 may suffer a large deflexion in its path through air. The detailed study of these rare encounters has given us information of fundamental importance. In the majority of cases, the collisions are perfectly elastic and the laws of conservation of momentum, and energy hold. The collisions between the charged nuclei in fact resemble the collision between two minute perfectly elastic billiard balls. Using the well-known expansion method devised by C. T. K. Wilson, Blackett has obtained a number of photographs of the trails of the particles involved in these rare collisions. Examples are shown in Fig. 2. The rectilinear tracks show the trails of the a-particles and the forks show the trajectories of the a-particle and the nucleus with which it collides. Photographs are taken from two positions at right-angles so that the trajectories of the particles in space can be determined, and the angles between them accurately measured. The first shows the collision of an a-particle with the hydrogen nucleus of mass 1. The a-particle is only slightly deflected but the hydrogen nucleus flies off at an angle with high speed. The second shows the collision with the helium nucleus of the same mass as the a-particle. It is impossible to distinguish between the tracks of these two identical particles. Ordinary collision theory shows that the two tracks should be always at right-angles to each other and this is found to be the case by measurement. The third photograph shows the collision with an oxygen nucleus of mass 16. The short fork shows the trail of the recoiling oxygen nucleus. It must be borne in mind that these collisions must not be regarded in the ordinary mechanical sense. The actual nuclei never come in contact in the collision, but the deflexions are due to the intense electrostatic forces between the colliding nuclei. Considering the minute masses of the nuclei involved, these collisions are of enormous intensity. For example, in a head-on collision of a swift a-particle with a heavy nucleus in which the a-particle is deflected backwards, 13 525

FIG. 2.

526

Nov. 1932.

ATOMIC PROJECTILES.

195

it can be calculated that the maximum force of repulsion between the nuclei, notwithstanding their minute masses, corresponds to several pounds in weight. In general, the law of the inverse square holds for the electric forces between colliding nuclei, so that the a-particle describes a hyperbolic path in its collision with a heavy nucleus. The classical experiments of Geiger and Marsden on the large angle deflexion of a-particles show that the number of a-particles scattered at different angles is in complete accordance with that to be expected for a Coulomb law of force. By measuring the fraction of the a-particles scattered through a definite angle by thin foils of metal, Chadwick has shown that the nuclear charge on an element is numerically equal to its atomic number—a result deduced by Moseley from his famous experiments on the X-ray spectra of the elements. We can readily calculate the closest distance of approach between the colliding nuclei for a Coulomb law of force. If the distribution of the scattered a-particles is consistent with this law of force, it is clear that the colliding nuclei behave as charged points even in the closest collisions. In this way we have been able to fix an upper limit to the dimension of the nucleus of the elements, for it is to be expected that the Coulomb law of repulsion would break down when the nuclei approach very closely to each other, and still more when the a-particle actually enters the structure of the nucleus. This question has been carefully examined in a series of experiments by Dr. Chadwick and myself. It is to be anticipated that the breakdown in the law of force would be first observed in collisions of a-particles with light atoms, since the repulsive forces are much smaller than in the case of heavier nuclei. We have found by experiment that the forces become abnormal for all the light elements examined. An illustration of the results obtained by a study of the scattering of a-particles by nuclei of aluminium is shown in Fig. 3. If the Coulomb law of force holds, the number n of a-particles scattered should vary as 1/m4 where u is the velocity of the colliding a-particle. The product n.u4 should thus be constant for all velocities and is shown by the horizontal line in the Figure. It will be seen from the experimental curve that this is far from being the case. For slow a-particles, the numbers of scattered a-particles agree approximately with the law of Coulomb, but with increasing speed the number falls to a minimum and then rises again rapidly. Similar results were obtained for other light elements, but, for all elements heavier than copper, the scattering results were entirely normal. The abnormal laws of force observed for aluminium and other light elements show clearly that a swift a-particle in a direct collision 527

196

ATOMIC PROJECTILES.

Nov. 1932.

must approach very nearly to the nucleus. Actually we shall see later that there is definite evidence that, in a certain fraction of these collisions, the a-particle undoubtedly enters into the nucleus and is captured by it. This at once leads to a consideration of the experimental methods that have proved successful in transmuting the atoms of matter. The Transmutation of Matter.—The development of chemistry during the last century showed that the atoms of the elements appeared to be indestructible entities, for they were not affected by the most drastic physical and chemical agencies. A definite

FIG. 3.

attack on the problem of transmutation had to await a clearer understanding of the structure of the atom and the development of methods for detecting individual atoms in swift movement. The discovery of radioactivity in 1896 provided the first clue to the elucidation of the problem of transmutation. The proof by Rutherford and Soddy that the heavy radioactive elements were undergoing spontaneous transformation was a great step in advance and revived the dormant interest in this age-old problem. The study of these uncontrollable transformations, accompanied by corpuscular radiations of great individual intensity, gave us for the 528

Nov. 1932.

ATOMIC PROJECTILES.

197

first time some idea of the powerful forces that must exist within the atom. The a-particles emitted from radioactive bodies are the most energetic projectiles known to science, if we except the very swift particles accompanying the ultra-penetrating rays in our atmosphere. We have seen that the nuclear structure of all atoms was first demonstrated by studying the scattering of a-particles. It is important to emphasize at the outset the clear-cut distinction between the nucleus and the external planetary electrons. By the action of light and X-rays or by the action of swift-moving particles, one or more of the outer electrons of an atom may be readily removed. This, as we say, leads to an ionization of the atom and profoundly alters its main properties. This change, however, is only temporary, for at the earliest opportunity the nucleus captures new electrons from its surroundings and the atom regains its original state. If we wish to produce a veritable and permanent transformation of an atom, it is necessary to alter the charge on the nucleus, or what is equivalent, to add or remove one or more of the charged units of which the nucleus is built up. In 1919, I made some experiments to test whether the nuclei of light elements could be altered by bombardment with the energetic s-particle. As we have seen, it seemed probable that in a direct collision, the a-particle must approach very closely to the nucleus of a light atom and might even enter its structure. When a powerful stream of a-particles was passed through nitrogen gas, a number of penetrating particles was observed by the scintillation method. These were found to be hydrogen nuclei—protons—travelling at high speed. Under the experimental conditions the appearance of these swift protons could only be accounted for on the assumption that they were ejected from the nitrogen nuclei as a result of close collision with the a-particles. In this way, we obtained for the first time definite evidence of the transmutation of an element by artificial methods. On account of the minute dimensions of the nuclei, these disintegrating collisions are very rare, for only one proton is observed for about 100,000 incident a-particles. Later, Blackett and Harkins were able to obtain direct photographic evidence by the expansion method of a few of these disintegrating collisions, and showed that the bombarding a-particle was actually captured by the nucleus, resulting in the expulsion of a swift proton. The nature of this transformation is now clear. The nitrogen nucleus of charge 7 and mass 14, captures the a-particle of mass 4 and charge 2, while a proton of mass 1 and charge 1 is expelled. The resulting nucleus thus has a charge 8 and a mass 17, or in other words is an isotope of oxygen of mass 17. Long after these experiments, ordinary oxygen was found to contain 529

198

ATOMIC PROJECTILES.

Nov. 1932.

a small quantity of an isotope of mass 17, showing that the product of transmutation of nitrogen is a stable element. In conjunction with Dr. Chadwick, a number of other light elements were found to be transformed by a-ray bombardment. In all cases protons were ejected, but the number and speed of the protons varied for the different elements. In the case of aluminium, some of the protons were ejected with an exceptionally high velocity and had an energy even greater than that of the bombarding a-particle. It appears probable that in all these cases of transmutation, the a-particle is captured by the nucleus and a proton ejected, so that the process leads to the formation of an element of mass 3 units greater and charge 1 unit higher than the bombarded nucleus. In the earlier experiments the protons were counted by the scintillation method, but in the last few years, powerful electric methods have been developed to obtain a photographic record of each proton entering the detecting chamber. By these developments the counting of protons has been made much easier and more reliable. A number of experimenters including Bothe, Pose, Chadwick and Constable and others have shown that the expelled protons may be divided into a number of groups each of definite speed. The interpretation of these results has already thrown much light on the mechanism of transformation and given us valuable data on the energy levels present in the lighter nuclei. In particular, the difference of energy between some of the proton groups can be correlated with the emission of energy in the form of g-rays which has been observed for some of the elements. Discovery of the Neutron.—Another unexpected type of transformation has been disclosed by these experiments on artificial disintegration by a-particles. When the light element beryllium is bombarded by a-particles, no protons are emitted, but Bothe observed some years ago that a very penetrating type of radiation appeared. This was initially supposed to be of the g-ray type, but a more detailed investigation by M. and Mme. Curie-Joliot disclosed a very surprising property of this radiation. When this radiation from beryllium passed through material containing hydrogen, it led to the appearance of swift protons. Chadwick carried out further experiments by counting methods and found that the radiation was able to set the nuclei of all light elements in rapid motion. He concluded that the radiation was not of the g-ray type at all, but consisted of a flight of material particles or " neutrons " as he termed them. The neutron is supposed to consist of a very close combination of a proton and an electron which 530

Nov. 1932.

ATOMIC PEOJECTILES.

199

FIG. 4.—Tracks of Protons produced by collisions with Neutrons.

b-particles which are probably due to g-radiation.

The thin tracks are due to fast

behaves as a unit, of charge zero and mass nearly equal to 1. The swift neutron occasionally collided with the nuclei in its path and set them in rapid motion. In Fig. 4 is reproduced a photograph, obtained by Mr. Dee, which shows some tracks of protons set in 531

200

ATOMIC PROJECTILES.

NoV. 1932.

motion by the collision of neutrons. The protons have been ejected from a layer of gelatine on the roof of the expansion chamber. All the experimental data fitted in with this " neutron " hypothesis, and there seems to be no doubt that a new and strange type of particle has been disclosed for the first time. The idea of the possible existence of such neutrons had been suggested by me in 1921, and various experiments had been made to detect them in electric discharges but without success. In the light of these results, it seems clear that the reaction which leads to the expulsion of a neutron from a beryllium nucleus may be explained in the following way. The a-particle is captured by the beryllium nucleus while a neutron is liberated, expressed by the relation Be9 + He4 = C12 + " neutron," or in other words, the carbon isotope of mass 12 is produced as a result of the transformation which gives rise to the neutron. On this and other data, Chadwick showed that the mass of the neutron is 1.0065, somewhat less than the mass of the hydrogen atom 1-0077. This indicates, as we should expect, that the electron in the neutron is more tightly bound than in the ordinary hydrogen atom. Similarly, neutrons have been found to be liberated when boron is bombarded by a-particles. Time does not allow me to do more than mention the main properties of fast neutrons. On account of their very weak external field they pass freely through the electronic structure of the atom, and, as Dee has shown, they produce little if any ionization in their path. Their presence can only be detected when they collide with an atomic nucleus and set it in rapid motion. The recoiling nucleus is brought to rest in traversing matter, producing an intense ionization along its path. While many of the collisions of a neutron with a nucleus are elastic, occasionally the neutron may enter a nucleus and in turn cause its transformation. This has been clearly brought out by the experiments of Feather, who has obtained definite evidence that both nitrogen and oxygen suffer transformation under a bombardment of neutrons. The transformation too is of a novel character, for in many cases an a-particle is liberated both from nitrogen and oxygen. Feather concludes that in the case of nitrogen the neutron is in some cases captured by the nucleus according to the reaction N14 + neutron1 = B11 + He4, an a-particle and the isotope of boron 11 being a product of the transformation. There is also some evidence that other types of transformation can occur in nitrogen, but the data are not yet sufficiently precise for a definite decision on these points. 532

Nov. 1932.

ATOMIC PROJECTILES.

201

The transformation of oxygen seems to be analogous to that of nitrogen. The neutron is captured by the oxygen nucleus according to the reaction 016 + neutron1 = C13 + He4 an a-particle and the isotope of carbon 13 being a product of the interaction. This is the first time that evidence has been obtained of the transformation of oxygen for it appears to be unaffected by bombardment with a-particles or protons. The use of neutrons as bombarding particles thus promises to throw much new light on the possible modes of transformation of the light elements. In these various types of transformation both by a-particles and neutrons, it is believed that the conservation of energy and momentum are valid. In testing this conclusion, it is necessary to take into account not only the kinetic energies of the incident a-particles and of the products of the transformation, but also the masses of the nuclei before and after transformation. It is known that mass and energy are closely related and a change of mass dm of a system is equivalent to a change of energy c2dm where c is the velocity of light. Since the masses of many of the lighter nuclei are known from the work of Aston with considerable accuracy, the energy balance before and after disintegration can be estimated when the kinetic energy of the bombarding particle and recoiling nuclei are known. In this way, it has been possible to show that the transformations proposed are consistent with the conservation of energy within the limits of accuracy of the experimental data. Experiments with Fast Protons.—I have already described the general methods that have been adopted to obtain streams of high velocity atoms for experimental purposes. During the course of the present year, the first definite evidence has been obtained that certain atoms can be transformed by a stream of fast protons produced artificially in a discharge tube. In the experimental arrangement adopted by Cockcroft and Walton, a steady stream of protons of energy up to 600,000 electron-volts was utilized to bombard matter. Preliminary experiments were made by them to determine the penetrating power of protons of different speed, and it was found that a sheet of mica of stopping power equivalent to 1 cm. of air was sufficient to absorb completely the fast protons used in the experiment. The general arrangement is diagrammatically shown in Fig. 1. The stream of protons of known energy fell on the target and observations were first made by the scintillation method to detect the presence of any penetrating particles produced by the bombardment. These passed through an opening in the side of the 533

202

ATOMIC PROJECTILES.

Nov. 1932.

tube covered with sufficient thickness of mica to stop completely any fast protons scattered from the target. When a sheet of lithium metal was used as a target, a large number of brilliant scintillations was observed for an accelerating voltage of 300,000 volts, using a proton current of one microampere. The scintillations disappeared when the proton current was cut off. The particles coming from the lithium target had a definite range in air of about 8 cm. and from the brightness of the scintillations it seemed likely that they were of the same mass as the a-particles (He4) spontaneously ejected from radioactive substances. These particles from lithium were also counted by electric methods and the ionization produced in air at different points of their path was in close agreement with that to be expected for a-particles of the same range. It thus appears that some of the lithium atoms under the influence of the proton bombardment have been transformed with the emission of a-particles. The general nature of the transformation seems clear. The lithium nucleus of charge 3 and mass 7 captures a proton and then breaks up into two a-particles of mass 4 according to the scheme Li7 + H1 = He 4 + He4. If this view be correct, we should expect from the principles of momentum that the transformation of the lithium atom should be accompanied by the expulsion of the a-particles in nearly opposite directions. Cockcroft and Walton have given strong experimental evidence of the correctness of this important deduction. We have no definite knowledge of the structure of the lithium nucleus except that it consists of 7 protons and 4 electrons, but whether these units are free or combined is unknown. We may suppose in a general way that the capture of the proton by the nucleus leads to an instability of the nuclear system which at once forms two a-particles and breaks up. No doubt we have to consider the whole process of capture of the proton and the break-up of the resulting nucleus into a-particles as one process which we cannot at the moment analyse further. The number of a-particles observed with a lithium target increases rapidly with the voltage. A few particles were seen at a potential as low as 70,000 volts and a very large number for 500,000 volts. The variation of number with voltage found by Cockcroft and Walton is shown in Fig. 5. The energy of the a-particle from lithium is about 8-6 million electron-volts and is thus nearly as great as that of the swift a-particlo expelled from thorium C. If two a-particles of equal energy are 534

Nov. 1932.

ATOMIC PROJECTILES,

203

liberated in the transformation, the total energy released is about 17 million electron-volts. Since a transformation is occasionally effected by a 100,000 volt proton it is seen that the energy liberated is about 170 times the energy of the proton which enters the nucleus and causes its disintegration. There is thus a large gain of energy

FIG. 5.

in the single process, but it must be borne in mind that only a minute fraction of the incident protons are effective in producing transformations. For example, even for 500,000 volt protons, only one disintegration is observed for 100 million protons incident on the target. The energy liberated in the lithium disintegration is in approximate accord with the conservation of energy, when we take into account the masses and kinetic energies of the nuclei concerned. 535

204

ATOMIC PROJECTILES.

Nov. 1932.

The bombardment by swift protons seems to be remarkably effective in promoting the transformation of atoms. As far as observation has gone, a-particles are liberated from many elements, but there is also some indication that in some cases other types of swift particles may be liberated. Boron and fluorine give a large number of a-particles but of smaller range and energy than those coming from lithium. It is natural to suppose that in these cases also the proton is captured and that the resulting nucleus breaks up with the emission of an a-particle. On this view, it may be that the fluorine nucleus of mass 19 is converted by the capture of a proton into oxygen and helium according to the reaction F19 + H1 = O16 + He4.

A smaller number of particles have been observed for beryllium, carbon, aluminium, potassium, and many other elements. Even in the case of heavier elements like cobalt, nickel, copper, silver, definite evidence has been obtained of the emission of a small number of particles resembling a-particles. It is of interest to note that pure iron gives a very small emission compared with the neighbouring elements cobalt and nickel. Still more surprising, a number of swift particles were observed from the heavy elements lead and uranium. Gold gives very little effect. The experimental results so far obtained by Cockcroft and Walton are of a preliminary nature, and much detailed experiment will be required before we are in a position to understand the exact nature of these transformations. It is obviously necessary to pay close attention to the purity of the elements exposed to the protonic bombardment to be sure that the particles originate in the element under examination and not from some easily transformed element existing as an impurity. It should be pointed out that on the wavemechanical theories, as Gamow first showed, there is a finite probability that a swift proton falling on light elements should be occasionally captured by the lighter nuclei. The number of nuclei disintegrated by the bombardment of light elements like lithium, boron, and fluorine are roughly of the order of magnitude to be expected on the general theory. On the other hand, it is difficult on the present views to account for the magnitude of the effects observed in heavier elements like copper and silver, and still less for elements like lead and uranium, which have a high nuclear charge. The chance of a proton entering a heavy nucleus like uranium should be exceedingly small unless we suppose that there is some type of resonance between the bombarding proton and the heavy nucleus which is at present difficult to understand. 536

Nov. 1932.

ATOMIC PROJECTILES.

205

The proof that protons artificially generated are able to produce transformations in a number of elements opens up a wide and very interesting new field for investigation. It will be of great interest to carry the experiments to higher voltages and to examine the effects produced by swift bombarding particles of different kinds. It seems clear that this new method of attack, so successfully begun by Cockcroft and Walton, will in the course of the next decade give us much new information on the structure of nuclei and the problem of the transmutation of the elements. We have seen that the bombardment by a-particles leads to the building up of nuclei of heavier mass, while the bombardment by the neutron and proton in general seems to result in disintegration which lowers the mass of the nucleus. A close comparison of the transformation effects produced by such different types of bombarding particles cannot fail to give us new information to interpret the mechanism of the interactions which lead to transmutation. It is obvious also that the results to be obtained by these methods will help us to understand the processes of production and destruction of nuclei which must occur in the interior of our sun or other hot stars under the influence of the swiftly moving nuclei of different kinds which arise from thermal agitation. It will be of much interest also to study the transmutation that may be produced in matter by the extraordinarily energetic particles observed in our atmosphere, and to examine also the effect on nuclei of radiations of the g-ray type of very high quantum energy. We can be sure that a vigorous and concentrated attack on this fascinating problem cannot fail to add much to our knowledge of that minute world of its own—the atomic nucleus.

Vote of Thanks. Mr. LOUGHNAN ST. L. PENDRED (Past-President) proposed a vote of thanks to Lord Eutherford for his Lecture, which was seconded by Lt.-Col. E. KITSON CLARK, T.D., M.A. (Past-President) and carried with acclamation.

537

This page intentionally left blank

Reid Anthony Railton (1895-1977) Railton was educated at Rugby School (1909-1912) and Manchester University (1912-1915). He gained his pilot's licence in 1915, and in the same year he joined Leyland Motors as a works apprentice. His service with the RNVRs Motor Boat Section in 1917-1918 completed the early experience with land, sea, and air which would shape his later career. After further work at Leyland Motors, Railton designed and built cars for Arab Motors of Letchworth, before becoming Technical Director of Thomson and Taylor Ltd. This gave him charge of the firm's racing car development work. Ample evidence of this is in his written account, reprinted here. Railton supplemented this by acting as a consultant in automobile speed records from which his fame as an engineer is derived. Railton collaborated with both Malcolm Campbell and John Cobb, who were each attempting the world land-speed record. Railton worked with Campbell, on Bluebird, summarising vehicle design developments in the paper "Bluebird 1930-1933" for the Institution of Automobile Engineers. He modified Campbell's car to take a 1,400 bhp Napier aeroengine, and then a Rolls-Royce power-plant providing in excess of 2,400 bhp. These efforts involved major reshaping of the automobile, but extended the record to over 300 mph. John Cobb approached Railton for an entirely new car in 1935. The Napier-Railton Special was a futuristic fantasy. Teardrop-shaped and aluminium-hulled, it combined lightweight bodywork with unorthodox design elements. Having no radiator, the vehicle was cooled by an inboard ice-box. Power was by angled aeroengines driving both front and back wheels. Showing an astonishing turn of speed, the car took the world record from 350 mph to 394.2 mph by 1947. Railton also worked on speedboats for his two main clients - Bluebird II and Crusader. However, it was the Cobb car which marked the high-point of his career, as the Railton Special held the land-speed record for 25 years.

539

279

RACING MOTOR CAR DESIGN By R. A. Railton, B.Sc.* The development of the racing car to-day is taking place almost entirely in Germany and Italy. It is true that during the last few years British cars have figured prominently in the lists of world speed records, but the cars used for these attempts are not racing cars at all in the true sense of the term. The most important single factor in the suspension of a racing car is to keep the unsprung weight down to an absolute minimum, and this is one of the chief aims of the designer. Independent suspension at all four wheels has been found to have advantages quite out of proportion to the mere saving in weight. There is now a fairly general agreement that if the centre of gravity of a racing car is too low, the tendency to skid is increased. Racing cars built on the frontwheel drive principle have had a considerable success in America> but they suffer from the drawback of instability at the critical speed when rounding a corner. Brakes and transmission have both been developed to a high degree of perfection, but resemble in principle those now used on touring cars. Supercharging is general, largely as a result of special circumstances which the author discusses. The use of heat-resisting alloy steels has also become an important factor in the design of the engines. The author considers also the influence of fuel, tyres, and streamlined shape on the maximum speed attainable on land. Introduction. In preparing this paper the author has been in some doubt as to the best way of approaching the subject. In the first place, from the ambitious nature of the title it might be assumed that the author regarded himself as an expert on racing car design. This is not so. There are, unfortunately, no such experts in this country. The development of the racing car to-day is taking place almost entirely in Germany and Italy. These two countries take the view that success in international road races is well worth the cost, simply as a matter of national prestige. In both countries certain factories receive Government "encouragement", which in various ways amounts to a very considerable subsidy, so long as they perform with reasonable success in the big European races. There is, of course, nothing of this sort in England. Under these conditions the only possible source of supply for the * Chief Engineer, Messrs. Thomson and Taylor (Brooklands), Ltd. [I.Mech.E.] 541

280

RACING MOTOR CAR DESIGN

large sum of money required is the British motor trade itself. Not being directly interested in national prestige, the leaders of the industry have so far shown no inclination to finance a British racing car. It is true that, during the last few years, British cars have figured prominently in the lists of world speed records, but the cars used for these attempts are not racing cars at all in the true sense of the term. They are generally vehicles built for one particular purpose, and are useless for competitive racing on road circuits. By comparison with the development of a first-class racing car their cost is small, and their value as a means of "improving the breed" is in about the same proportion. There is, however, one thing to be said in favour of these machines, and that is that their exploits have a certain amount of "prestige value". The author has been told by those well qualified to judge that the retention by England of the world land speed record is well worth while so far as overseas sales are concerned. The same sentiment was felt in the aeroplane industry at the time when this country held the air speed record as well. Having uttered these excuses, the author had better make it clear that, so far as modern racing cars are concerned, he can only claim to be an expert observer. Of world record cars he can speak with more confidence, having been engaged in their design and manufacture for the last six years. Another question which has not been easy to decide is what aspect of the subject would be of most interest to the Institution. The subject of the paper represents a specialized branch of a highly specialized industry, whose methods are so hemmed about by considerations of cost and expediency that it is sometimes difficult to reconcile them with accepted principles of engineering. This being so, the author proposes to avoid as far as possible the discussion of why certain methods are used in preference to others. He proposes to confine himself to an account of some of the special problems involved in making a motor car travel fast and to the methods adopted to meet those problems. The author's imagination has frequently been fired by the speeds achieved by modern ships, aeroplanes, railway trains, etc., and his curiosity has been aroused over the nature of the problems and obstacles which their designers have had to face. This paper is an attempt to supply just that information for any engineer who may feel the same curiosity about motor cars. Springing. 542

It is no use making a car go fast if it will not stay on

RACING MOTOR CAR DESIGN

281

the road, so it seems logical to start with some of the considerations which affect this supremely important point. In a racing car the suspension system, that is to say the flexible link connecting the frame with the road wheels and the pneumatic tyres, has as its chief object the maintenance of contact between the tyre and the road with as nearly constant a pressure as possible. The comfort of the driver is a secondary consideration. If the road were a true plane surface, no flexible spring or pneumatic tyre would be required. Unfortunately, this condition never occurs in practice. Whenever the road surface is such as to impart a vertical movement to the road wheels, it becomes necessary to devise means to allow the wheel to follow the surface of the road as closely as possible. Obviously, if the tyre, wheel, and axle (or unsprung parts) had no mass, there would be no difficulty about this. In practice, however, the inertia of the unsprung parts, acting vertically, tends to cause excessive pressure at the start of a bump, followed by reduced pressure at the top of the bump; and it is this reduction of pressure which, by reducing the adhesion, has a bad effect on the road-holding qualities of the car. Actually, it is difficult enough to prevent this pressure dropping frequently to zero under racing conditions, in other words, to stop the wheels leaving the ground altogether, as in Fig. 1, Plate 1. Clearly then the most important single factor is to keep the unsprung weight to an absolute minimum, and this is in fact one of the chief aims of the designer. Next, let us consider the nature of the link between the axle and the frame of the car. It is nearly useless to attempt an analysis of the action of the spring and damping mechanism, which, in one form or another, invariably form this link on a racing car. The conditions are so variable, and the requirements are so conflicting, that there is no rule or formula that can be of much use to the designer here. The interaction of the kinetic energies of the sprung and unsprung weights are directly dependent on such variables as the size of the tyres, the air pressure, the wheel centres, the stiffness of the springs, the spring centres, the disposition of the chassis weight, the torsional stiffness of the frame, its stiffness as a beam, the nature of the road surface, etc. However, one or two guiding principles do emerge: (1) the greater the ratio between the sprung and the unsprung weights the better, for obvious reasons; (2) the greater this ratio, the more flexible the springs that will give the optimum condition of safety; (3) the greater this optimum flexibility, the worse the surface that can be traversed with safety; (4) the worse the surface, and the greater the amplitude of vertical movement, the more absorption of energy must be allowed for in the damping mechanism. 543

Fig. 3. German Auto Union Car

RACING MOTOR CAR DESIGN

283

From this it is clear that the nature of the road upon which the car is expected to run should have a big influence on the design of the suspension. This is in fact true, but it is only too often overlooked. About the worst surface that a racing car ever has to face is the track at Brooklands. Owing to the gradual sinking of the foundations, its surface is rich in wave-like formations, which at high speeds have the effect of terrific bumps. A car whose suspension is designed and adjusted for a good road circuit is often uncontrollable here, owing to the inadequate axle movement usually allowed and the excessive stiffness of the springs. It was mentioned above that the comfort of the driver is a secondary consideration, but one reservation must be made. No one who has not tried it can have the slightest conception of the physical punishment inflicted on the driver of a racing car travelling fast on a bad surface. It can be bad enough on some road circuits and tracks, but under extreme conditions it approaches the point where the jolting of the driver's spine begins to interfere with his efficiency, This is a second reason why adequate axle movement must be allowed on very uneven surfaces, even if the stability of the car does not demand it. This movement can only be safely allowed if the unsprung weight is sufficiently low. "Independent" Suspension. The whole problem then hinges on the question of unsprung weight, and it is impossible to overestimate its importance. This being so, it might reasonably be asked why more effort is not made to spring the wheels independently, and so to save the weight of the axle beam and driving mechanism. Independent suspension is in fact adopted on nearly every modern racing car, but its chief advantages are not directly consequent upon its effect on the unsprung weight. In any type of axle the wheels, tyres, and brake gear make up considerably more than half of the total weight of the unit, and for an independent system it is safe to say that the unsprung parts of whatever arrangement is used to support the wheels will increase this proportion still further. Nevertheless the complete change-over from normal to independent suspension has been such an outstanding feature of racing car design during the last ten years, that it deserves our attention. There is no doubt that it was first introduced (in Germany) in an attempt to reduce the unsprung weight to the lowest possible figure. It was soon found, however, that a car so equipped possessed certain advantages quite out of proportion to the mere saving of unsprung weight. It was found that such a car held the road on corners in a quite uncanny manner. It was also found possible to transmit more 545

Fig. 4. Italian Alfa-Romeo Car

RACING MOTOR CAR DESIGN

285

torque through the driving wheels without wheel slip. In a word, the adhesion was greatly improved, and it was also noticed that the worse the road surface the more pronounced was the difference. Further, it was found that these advantages were present whatever type, within reason, of independent suspension was used. In other words, what made the difference was the absence of a rigid member connecting the wheels. The nature of the mechanism employed to take its place was relatively unimportant. Three modern examples are shown in Fig. 2, Plate 1, and in Figs. 3 and 4. Just why the absence of the solid connecting axle beam should make such a difference to the adhesion was not at first very clear. Nor is it

Fig. 5. Showing the Effect, with a Solid Axle, of the Vertical Movement of one Wheel A, upon the Adhesion of the other Wheel B now, for this is a point upon which no two experts will agree. The basic principle is, however, probably as follows. Consider a pair of wheels A and B running along the road (see Fig. 5). If A meets an obstruction and suffers temporarily impaired adhesion, B has to provide extra adhesion if there is to be no skid. With a solid axle, the movement of A over the obstruction is transmitted to B in such a way as to cause a sudden lateral shift of B's point of contact with the ground. This may well precipitate a skid if B's adhesion is already taxed to the utmost. With independent suspension of each wheel this reaction does not occur. Chassis Frame. One other factor, which has a very far-reaching effect upon the behaviour of the car on the road, is the stiffness of the 547

286

RACING MOTOR CAR DESIGN

main frame. It is obviously useless to lavish care and expense in attempting to arrive at a nice adjustment of the suspension system, if the chassis is sufficiently flexible to act itself as a spring, and an undamped one at that. The provision of the necessary strength usually ensures that the frame is sufficiently stiff, considered as a beam; but it is equally important that there should be a great resistance to torsional strains. With normal methods of construction, this torsional stiffness is not so easy to obtain without increasing the weight, but the resulting increase in stability is well worth it, As with touring cars, the frames of racing cars are often constructed of two channel-section side members, connected by cross members at several convenient points. Unless these cross members are of tubular section, the whole structure offers very little resistance to torsion, with the result that the desired action of the suspension may be completely upset. There has lately been a tendency to fill in the open side of the channel section, thus forming a box member and greatly increasing the torsional rigidity, but the resulting structure is inclined to be heavy. From a consideration of the stress distribution in a motor car frame, there is no doubt that the ideal construction would be a single tubular trunk extending the whole length of the vehicle. In spite of the constructional difficulties involved, developments are taking place in this direction which may well become standard practice in the course of the next few years. The rear end of a chassis constructed on these lines is shown in Fig. 6, Plate 2. Tyres. The requirements of the tyre manufacturer have an important bearing on the suspension. For long-distance, fast work the durability required calls for a large heavy tyre, working at a high pressure. This increases the suspension difficulty in two ways. The extra unsprung weight brings the usual troubles in its train, and the high pressure means that road shocks are transmitted without much diminution. This is not a grave matter on a good surface, but on an uneven road it is a serious handicap. On the other hand, a consideration which encourages the use of a tyre of large section is its increased lateral stability. Any tyre, when the car is rounding a corner, has a tendency to roll, or to "go over on its ankles", and under these conditions the path of the wheel is not at right-angles to its axis (see Fig. 7). This phenomenon is termed "creep" and, though not actually a skid, it has a bad effect on roadholding. Tyre equipment, then, is a compromise between these and other conflicting effects, and must be decided for each individual car, 548

RACING MOTOR CAR DESIGN

287

according to the road or track conditions that it is called upon to face. Weight Distribution. There are few features of racing car design that receive less direct attention than the effect of the weight distribu-

Fig. 7. Showing the Tendency of a Tyre to "Creep" under the Influence of Lateral Forces tion upon the behaviour of the car on the road. The guiding principle, until recently, seems to have been a determination to keep the car as short as possible and as low as possible, and to leave it at that. As a general rule, this gives fairly good results, but even a quite elementary analysis of the problem shows up the limitations of such a rule. To 549

288

RACING MOTOR CAR DESIGN

start with, there are several traps to avoid when trying to keep a car very low. There are two obvious reasons for keeping the height of a racing car down. First, the lower it is, the less air it disturbs and, other things being equal, the less wind resistance it offers. Second, the lower the centre of gravity, the less likely is the car to turn over if it skids sideways into an obstacle. In other words, the lower the car the safer it is for the occupants in case of accident. This is not the same thing as saying "The lower the car the less it is likely to skid." This is one of the "traps". There is a quite general idea that the lower the centre of gravity of a car, the less likely it is to skid; in other words, the faster it will go round a corner without skidding. The belief is probably a survival of the days when cars were built so high that the limiting factor in cornering was the tendency to turn over. Naturally, then, the lower the car, the faster it would go round a corner without turning over. To-day, even the ordinary saloon touring car is built so low that it will skid outwards long before the turn can be made sufficiently abrupt for it to capsize. There is now a fairly general agreement that if the centre of gravity is too low, the tendency to skid is definitely increased. One thing is certain, and that is, that a car having a very low centre of gravity behaves in a peculiar way when taken round a corner at or near the critical speed. It will maintain a perfectly true course until that speed is exceeded, when it will slide right out in an uncontrollable skid. It seems probable that this phenomenon is related to the observations of locomotive engineers in connexion with the effect of the height of the centre of gravity on rail wear and on the tendency to derailment. There is another "trap" in going to extremes to secure a low centre of gravity. Certain races are organized for cars in "touring trim", that is to say, they have to carry lamps, battery, spare wheel, etc. If the centre of gravity is to be kept very low, the individual items must obviously also be as low as possible; that is to say, they cannot be put one on top of the other, but must be more or less strung out in line. Thus heavy units like the fuel tank and the batteries may have to be placed behind the rear axle, and heavy accessories may find themselves in front of the radiator. The objection to this may not be apparent at once. The relative weights on the front and rear axles may be correct, and the chassis may lend itself to an admirably streamlined body. Nevertheless, the car will almost certainly have the grave drawback of being slow to answer the steering, the large mass behind in conjunction with the mass of the engine forward having considerably increased the polar moment of inertia of the car, or its "flywheel effect", about a vertical axis. 550

RACING MOTOR CAR DESIGN

289

Consider the two cases of a car moving (1) at a constant speed in a straight line; and (2) round a curve of constant radius. In the first case, the angular velocity of the car about a vertical axis is zero, and, in the second, its angular velocity is constant. In neither case does its polar moment of inertia affect the situation. When, however, the car is turned from a straight course into a curved one, or when the radius of the curved path varies, its mass is subject to an angular acceleration. Now the amount of the couple producing this angular acceleration is limited by the friction of the tyres on the road, and has a definite maximum for any given vehicle. If, therefore, the polar moment of inertia is high, the maximum angular acceleration will be low, and the car will take an appreciable time to follow the movements of the steering wheel. This effect can be very noticeable in extieme cases, and is particularly objectionable in a race over a twisty road. A very low centre of gravity therefore is not necessarily desirable in a road racing car. In the case of a track racing car, where the first consideration is the ultimate speed, one of the most important requirements is the reduction of the frontal projected area of the car in order to reduce its resistance to the air. This frequently does involve a very low centre of gravity, which in this instance is not objectionable. In addition, the added safety from capsizing, in case of accident, is a real advantage. There is one special type of racing car where the height of the centre of gravity has a very important bearing on the performance, namely, the car specially built for hill climbs and short sprint races. These events are really tests of acceleration, for which the most important consideration is the reduction of weight of the vehicle. At the same time, there must be sufficient weight on the rear axle to supply the necessary adhesion for the tyres. With a fairly high centre of gravity and a very short wheelbase, quite an appreciable transference of weight from the front to the rear wheels occurs when accelerating, and the effect of an upward road-gradient still further increases this transference. This, in turn, increases the available adhesion, and permits a greater torque to be applied to the driving wheels without slipping. Incidentally, a car of these proportions is very suitable for negotiating the sharp bends usually included in hill climbs. The question is often raised as to what is the ideal proportion of track to wheelbase for racing purposes. Here again the requirements are conflicting. Generally speaking, a long wheelbase is steadier on the straight, and on curves at speeds below the critical or skidding speed. On the other hand, it is fairly generally accepted that a short wheelbase will negotiate a sharp corner faster without skidding. The long wheelbase is more comfortable and therefore less tiring to the 22 551

290

RACING MOTOR CAR DESIGN

driver. This is an important point in a long race, but it has the disadvantages of extra weight, and, for a given construction, of greater flexibility of the frame. For these reasons, the use of a long wheelbase (that is to say a wheelbase longer than is actually required to find room for the driver and various impedimenta) is confined to cars intended for straightaway world record attempts, and occasionally for long-distance work on banked or circular tracks. The actual wheelbase is nearly always dictated by the exigencies of space; and it is usually kept as short as is consistent with this requirement. There is remarkably little variation in width of track between designs otherwise very dissimilar. This dimension is connected with the question of the height of the centre of gravity, and with the considerations already mentioned. There seems to be remarkable unanimity in favour of making the wheelbase approximately twice the track width. This proportion has been arrived at as the result of a great amount of experience, and it probably represents as good a general compromise as is possible. Front Wheel Drive. Any discussion of the question of road-holding would be incomplete without some reference to the possibilities of front wheel drive. Racing cars built on this principle have had a considerable vogue in America, where special conditions have favoured the experiment. The majority of races there are held on saucer tracks, where a uniform high speed is maintained, and road races of the European type are unknown. The advantages of front wheel drive on a motor vehicle are chiefly of an indirect nature, and consist almost entirely in the better chassis layout obtainable by its use. So far as racing cars are concerned, such an improvement may result in a lighter and more compact construction, with consequently improved performance and reduced wind resistance. When first introduced, great claims were made on account of the fact that, on a corner, the propelling force in a front-driven car has a slight component radially inwards toward the instantaneous centre, whereas with rear drive there is an outward component. This effect is illustrated diagrammatically in Fig. 8. It was therefore claimed that the front-driven car could be taken round a corner faster than the other. There is a certain amount of truth in this, but there is also one very serious drawback. Imagine a front-drive car negotiating a corner at about the critical speed. In order to take advantage of this effect, the driver must have the throttle open, so that this centripetal tractive effort may be exerted. If he overdoes it a fraction and the car begins to skid there are only two things that he can do. Either he can follow his instinct and take his 552

RACING MOTOR CAR DESIGN

291

foot off the throttle, in which case the sudden reversal of this inward component of the tractive effort may cause the car to skid uncontrollably ; or, brave man, he can open the throttle still further, in the hope of further increasing the inward component so as to restore equilibrium. This action also increases the velocity of the car, so that

Fig. 8. Showing Inward and Outward Components of the Propelling Force with Front and Rear Wheel Drive Respectively his security is only temporary. In a word, the condition is unstable. For this reason, it is seldom safe in practice for him to attempt to avail himself of the increased speed which is theoretically possible. It may be that the reverse effect in a rear-drive car is in practice a definite advantage, in that the driver gets a warning, in the shape of an incipient skid, which he can immediately check by taking his foot off the throttle. 553

292

RACING MOTOR CAR DESIGN

Steering. The actual steering mechanism used on racing cars differs very little from that employed on the ordinary touring vehicle. The most important requirement is the maximum of rigidity in the linkage and other mechanism between the steering wheel and the road wheels. To be effective, this involves great lateral rigidity of the means of location of the front axle itself, and the whole of the front end of the frame. These are points which are frequently overlooked. It is sometimes found, in the case of steering arms, frames, etc., which are quite adequate for touring purposes, that their slight deflexions under load combine to produce a definite lag in the steering which is objectionable for racing purposes. This point is important chiefly because of its psychological effect on the driver. Nothing engenders confidence more quickly than a quick-acting solid-feeling steering gear. In these days road-race driving has developed into a fine art. The few first-class men available may handle many different makes of car in the course of a season, and the opportunity for real practice before the race often amounts to a total of less than an hour. If the handling of the car presents any peculiarities, or if the steering is anything but very good, the driver may only begin to settle down about half-way through the race. The best engine in the world will be of little use under these circumstances. Brakes. Another factor that can make or mar the driver's confidence is the brakes. The development of brakes in the last few years has been remarkable. To anyone who has not experienced it, the stopping powers of the modern road-racing car are almost incredible. To see drivers approach a corner at 140 rn.p.h., and cram on the brakes so late that disaster seems inevitable, is probably the most spectacular part of motor racing. Some years ago it was fashionable to use a servo-motor driven off the gearbox to apply the brakes, and to use a high pressure on comparatively small drums. This was quite effective, but it was found that it was difficult to obtain that delicacy of control upon which the driver's feeling of confidence is largely dependent. A sensitive brake, responding instantly to every variation of pedal pressure, is only second in importance to perfect steering in this respect. Shortly afterwards, great improvements in materials led to the production of brake linings having a higher coefficient of friction and greater hardness. The use of these materials, in conjunction with increased drum diameters, made it possible to dispense with the servo-motor ; and, except for the very largest cars, nearly every maker has reverted to the use of direct pedal operation. For many years the brakes were operated by a system of rods and 554

RACING MOTOR CAR DESIGN

293

levers connecting the foot pedal with the ordinary double cam which expanded the shoe against the drum. With the advent of independent suspension, such a mechanical linkage became very complicated, and the recent perfection of the hydraulic system on touring cars has resulted in its sudden and almost universal adoption for racing. Special care has to be taken to insulate the working fluid from the hot drum, and special liquids of high boiling-point are used as an extra precaution. Fig. 9 shows the layout of a racing brake. It will be seen

Fig. 9. Hydraulic Brake Layout

Fig. 10. Typical Brake Drum Construction

that surrounding the operating cylinder there is a shroud, insulating it from the hot drum and supplied with cool air through an external scoop. The provision of adequate braking for a car weighing nearly a ton and capable of 180 m.p.h. is no easy problem. "Adequate braking" in the driver's view calls for a negative acceleration of at least 1/2g from top speed and about3/4gfrom 100 m.p.h. down. Bearing in mind that, for a given negative acceleration, the rate of heat flow to the drum varies as the square of the speed, a moment's thought will show how 555

294

RACING MOTOR CAR DESIGN

much care has to be given to the design of the drums. Add to this the fact that the drums are placed in the wheels just where the weight must be kept to a minimum, and the nightmare is complete. The solution up to now has been a thick and heavily ribbed light alloy drum of large diameter, with a thin liner of hard steel shrunk inside it. Such a composite drum is shown in Fig. 10. Transmission. The elements which transmit the power from the engine to the driving wheels differ very little from the corresponding parts in the ordinary motor car. The familiar combination of clutch, sliding gearbox, and bevel-driven axle is to-day almost universal in both types of vehicle. The relative position of the units varies, an unusual example being shown in Fig. 4, p. 284, but their working principle remains the same. So far as the clutch is concerned, there is a preference for a multiple disk for racing on account of its lighter weight. It is seldom used to-day on touring cars, owing to its tendency to "drag" when disengaged, and to the resulting difficulty of engaging the gear silently— an unimportant point in racing. The gearbox is generally of completely orthodox design, usually with four speeds and sometimes with only three. The modern engine delivers a high torque over such a big range of speeds that the need for several closely spaced ratios in the box is rapidly disappearing. The modern refinements of the sliding gearbox, such as "synchromesh" and "freewheel" devices, are not used, as they add to the weight of the gearbox and offer no advantages for racing. Several racing cars in this country have been fitted with planetary gearboxes of the "Wilson" type. These dispense with the clutch altogether and provide a nearly instantaneous change of gear which is as nearly as possible foolproof. Considering the intricate nature of their mechanism, they have been brought to a wonderful pitch of reliability and efficiency, and their future in this connexion will be watched with interest. Universal joints, where they occur in the transmission, are usually of the Hooke type. The latest practice is to furnish the trunnion bearings with needle rollers packed with grease and sealed for life (see Fig. 11, Plate 2). The combination of load, speed, and angle which these comparatively tiny joints will withstand is a tribute to the skill of the few specialists in whose hands their manufacture rests. The front-wheel drive car presents a special problem in universal joints. The front wheels must swivel about 25 deg. for steering, and, as they are also driven, they must accommodate a universal joint of small bulk and weight capable of transmitting the drive through this 556

RACING MOTOR CAR DESIGN

295

angle with uniform angular velocity. To meet this requirement, a very ingenious joint has been developed in America, where, in spite of the special machines required to make it and of the comparatively small

Fig. 12. Constant Velocity Universal Joint (Rzeppa) demand, it is already in regular production. In view of the admirable way in which it fulfils its difficult function, it deserves to be better known in the field of general engineering. Fig. 12 is a section drawing of this joint. 557

296

RACING MOTOR CAR DESIGN

The final drive to the axle shafts remains what it has always been— the straight-tooth bevel. On touring cars this has been entirely superseded by the Gleason or spiral bevel, chiefly on the score of its greater silence. For racing purposes silence is unimportant, and the straighttooth bevel has the advantage that it is easier and cheaper to produce in small quantities. Further, the axial thrust on the pinion is in the same direction on both drive and overrun, a consideration which simplifies and lightens the construction generally. Engine. The author has purposely reversed what would seem to be the natural sequence of his subject and has dealt with the chassis before the engine, on the assumption that the racing car would be of more general interest considered as a vehicle than merely as an application of the internal combustion engine. He has, in a word, put the cart before the horse. Having dealt with the cart we may now pass to a brief examination of the horse. It is sometimes thought that an engine designed for racing is a thing apart from the ordinary commercial petrol engine, a thing constructed regardless of cost, durability, or ease of handling. For this reason, it is sometimes asked what possible use racing can be to a manufacturer, considering that his racing engine bears no resemblance to the one by which he earns his living. In actual fact, the lines of development of racing and touring engines are steadily converging. The knowledge gained in the evolution of the production engine has had its effect on the racing engine, and vice versa. The cast iron cylinder and detachable head, developed for touring purposes, have been so perfected that they are now frequently used on racing engines. In the same way, the racing engine, after years of allegiance to ball or roller bearings for the crankshaft journals, has now reverted to the plain whitemetal bearing. Years of experience with bearing metals, methods of tinning, and the design of the supporting parts, have produced a wonderful degree of reliability in the plain bearing, even under racing conditions. On the other hand, the touring engine is now almost invariably equipped with aluminium pistons, narrow piston rings, heat-resisting alloy valves, adequately stiff crankshafts, and high-pressure lubrication, all direct legacies from the racing engine. It may be asked why it should be necessary to go to the expense of building racing cars, and entering and running them in races, when the same knowledge could be obtained on the test bench and in private trials on a track. The answer seems to be that it would not be necessary if the manufacturer were a machine and a 100 per cent efficient one at that. But he is not. He is simply a human being. It is 558

RACING MOTOR CAR DESIGN

297

primarily the human appeal of the sporting element in racing that provides the incentive for this development work. The harassed executive, worn out with the endless worries of his ordinary routine, turns with relief and genuine keenness to the consideration of his racing programme. He just cannot raise the same enthusiasm over the activities of his experimental department. The general arrangement of the racing engine is quite orthodox. It is usually a multicylinder affair with one or two rows of six or eight cylinders in line. The cylinder head has usually a nearly spherical surface, with radially disposed valves. There are generally two camshafts, one for the inlet and one for the exhaust valves. The ignition plug is disposed vertically between the valves and is nearly always fired by a magneto. Two views of a typical combustion head (eight cylinders) are shown in Figs. 13 and 14. The working mixture is either aspirated naturally from one or more carburettors, or, as is now usual, pumped in by a rotary blower. At various stages of its development there have been many factors which, from time to time, have limited the power output of the racing engine. As each difficulty has been overcome, the output of the engine has been increased until a limit has been set by some quite different factor. It may be of interest to trace the course of some of these difficulties and their solutions. For some time, the durability of the exhaust valve set a very definite limit to the flame temperature, and to the speed at which the valve could be operated. Over these limits there was serious burning of the valve, even if actual fracture did not occur. The solution of this problem has been the adoption of special steels, notably highchromium, cobalt, and tungsten steels, the latter having properties similar to ordinary high-speed tool steel. This improvement resulted at once in greatly increased flame temperatures and compression ratios, as well as an increase in the rate of revolution. The next limit to be reached was that imposed by the heat-resisting qualities and strength of the aluminium alloy pistons. This problem was sometimes evaded by the use of a large number of cylinders of small bore, the superior cooling of the small pistons making the use of higher compression ratios possible. Here again, the use of special heat-resisting alloys, coupled with improved foundry methods, has provided the remedy. For the sizes in general use, the provision of reliable pistons does not present much difficulty to-day. This solution of the piston difficulty sent engine speeds up with a rush, and brought to an acute state the troubles already being experienced with bearings, particularly the big end of the connecting rod. Designers were beginning to give up plain bearings in despair, and 559

298

RACING MOTOR CAR DESIGN

Fig. 13. Typical Valve Gear (Alfa-Romeo)

Fig. 14. Typical Cylinder Head (under side)

560

RACING MOTOR CAR DESIGN

299

were turning to roller bearings, in spite of the enormously increased cost. This was largely on the strength of the success experienced with this form of bearing in motor cycle engines. Owing to the form of construction used, there is very little stress in the crankpin of a motor cycle engine, and therefore the pin and consequently the diameter of the bearing can be kept down to reasonable limits. On the other hand, the crankpins in a car engine of six or more cylinders have to be of sufficient diameter to provide the stiffness necessary to avoid dangerous periodic vibrations. This means that the roller bearing surrounding this pin must be of such a diameter that the roller speed is very high indeed. In addition, the outer race often has to be split to facilitate assembly. Added to this again is the destructive effect of the rapid changes in angular velocity in a big-end bearing. These changes are constantly accelerating and retarding the rollers, and the consequent skidding between the rollers and their races affects their life very adversely. Fig. 15 is a cut-away view of one of the most successful engines of this period. It gives a good idea of the complication and expense involved in the use of roller bearings for the crankshaft. At the same time plain bearings were being developed intensively for aero-engines, where the weight of the roller bearing was prohibitive, so it was not long before the pendulum swung back, and the much improved plain bearing came into favour again. To-day it is quite common to run plain big-end bearings at rubbing speeds of upwards of 3,000 ft. per min. with unit pressures of 1,000 Ib. per sq. in., and failure is comparatively rare. So far as the question of bearing lubrication is concerned, the only point of unanimity is the method of introducing the oil, that is, under pressure from the hollow shaft. There is a wide divergence of practice in successful engines in the oilways provided on the bearings. Some have none whatever, whilst others are engraved with what appear to be ornamental designs. Their presence or absence seems to make very little difference one way or the other. Another factor, which in some engines has set a limit to the performance, has been the persistent breakage of valve springs. On this subject it is enough to say that the designer's technique has sometimes fallen short of an otherwise high standard. In other words, if sufficient care is taken in the original design, this trouble can always be avoided. At this stage, progress in design had arrived at a point where it was comparatively free from mechanical limitations. The engine could be relied upon to deal efficiently with all the fuel that could be got into it, provided the conditions (particularly the compression ratio) were suitable for that fuel. If the compression ratio was increased beyond 561

Fig. 15. Delage Roller Bearing Engine

RACING MOTOR CAR DESIGN

301

that limit, either the mixture would pre-ignite or the character of the combustion would be found to change suddenly, and the charge said to "detonate". It is in this field that most of the changes and improvements of the last few years have taken place. Up to the present, the chief developments have been twofold. First, external pumps, or superchargers, have been used to introduce more working fluid into the cylinders. Second, special fuels have been evolved, which can be used with very high compression ratios without producing the phenomena of pre-ignition and detonation. Of these two developments the latter represents by far the more important advance. Supercharging. To get a proper perspective of this subject, so far as it concerns racing cars, it must be borne in mind that in nearly every road race since the War and up to a few years ago, entries were confined to cars below a certain limiting size. This was done partly in the interests of safety, and partly with a view to limiting the cost of racing. The nature of this limit was not, as might seem reasonable, the weight of the car or the amount of fuel allowed. The limit was always applied to the capacity of the engine, that is to say the swept volume of its cylinders. The amount of this maximum volume varied, and was at various times 3,000, 2,000, and 1,500 cu. cm. Naturally it was not long before it occurred to someone that, although the regulations were explicit about the maximum volume of working fluid allowed per cycle, nothing was said about the pressure at which it might be admitted. Here was an obvious way to get ahead of one's rivals. If the inlet pipe pressure were raised from about 14 1b. per sq. in. abs. to say 24 1b. per sq. in. abs., and if at the same time the compression ratio were lowered to a point where pistons, valves, etc., could cope with the extra heat flow, clearly extra power would then be gained as the result of the flattened expansion curve. First principles showed that, with a reasonable blower efficiency, a substantial net increase in power would still be obtained, even after allowing for the energy required to drive the blower. An ordinary Roots blower of suitable dimensions was therefore geared to the crankshaft and coupled to the intake side of the carburettor. Shortly afterwards it was found that it worked better if the carburettor were placed on the suction side of the blower, as the latent heat of evaporation of the fuel helped to keep the blower cool, while the solution of a small amount of lubricating oil in the fuel happened to solve the problem of lubricating the blower. That, with minor modifications and different types of blower, is the supercharged engine of to-day. 563

302

RACING MOTOR CAR DESIGN

It is obvious that an engine of this type is likely to have a lower thermal efficiency, or a higher fuel consumption, than one developing the same power on atmospheric induction ; but this, under such regulations, is of little importance. If the general race regulations at the time had been based on car weight, or fuel consumption, instead of on the bore and stroke of the engine, designers would almost certainly have persevered with the development of the normally aspirated engine, an activity which would have been of infinitely more value to the manufacturer in his development of the touring car. There is, however, another side to the picture. In the last year or so the big Continental races have been thrown open to engines of any size. This, of course, has removed the main advantage of the supercharger, and has forced it to stand on its own merits; it must be said that it has stood the test very well. New cars built for these races have, without exception, been supercharged. This is no doubt partly due to the fact that designers have so far mastered the technique of forced induction, that they naturally hesitate to take the risk of experimenting in other directions. The main reason, however, is that, with present methods of engine construction, and for equal power outputs, a small engine plus supercharger weighs less than the larger unsupercharged one. So long as this remains true, engines designed to meet the present race regulations will probably continue to be supercharged. Three types of blower are in general use: the centrifugal type, the Roots blower, and the rotating vane type. The Roots blower is used almost entirely on the Continent, while the centrifugal type is popular in America. Great efforts have been made in this country to perfect the rotating vane blower and it has been used extensively. There seems to be no special reason for these preferences, except that in America most races are run on special tracks, at a more or less uniform high speed. This condition lends itself to the use of the centrifugal blower, which is obviously only efficient at or near its designed speed. Given this condition, it can be made lighter than the other types, for a given output. It is usually of the single-stage type, having a hightensile steel rotor about 7 inches in diameter, revolving at speeds up to between 30,000 and 40,000 r.p.m., the delivery pressure being from 8 to 12 Ib. per sq. in. An example is shown in Fig. 16, Plate 2. By far the most largely used type is the Roots. A great deal of development work has been carried out with it, chiefly on the Continent, with results that are amazing to those familiar only with its disreputable ancestor of the blacksmith's shop. The blowers are now run at speeds up to 10,000 r.p.m., delivering air at up to 25 1b. per sq. in. It is difficult to obtain any reliable figures as to the efficiency under these conditions, but the satisfactory results obtained from the 564

RACING MOTOR CAR DESIGN

303

power unit as a whole seem to indicate a figure of the order of 60 per cent. The rotors are usually of steel, either two- or three-toothed, and the casings of aluminium alloy, heavily ribbed for both stiffness and cooling. The clearances, of course, have to be considerable, and this seriously affects their performance at low speeds, though this is not a matter of grave importance for racing. Such a blower and the method of driving it are shown in Fig. 15. The third type, the rotating vane machine, is also a development of an industrial blower. As shown in Fig. 17, it has a rotating drum revolving in an eccentric casing, and carrying sliding radial vanes

Fig. 17. Rotating Vane Blower which follow the contour of the casing. Its adaptation for high-speed work has presented many difficult problems, chiefly in connexion with the high rubbing speeds involved, and with the durability of the surfaces concerned. In spite of these difficulties, superchargers of this type have been produced which give remarkable results with a very fair degree of reliability. They appear to show rather better efficiencies than the Roots type, particularly at low speeds. They are, however, essentially more complicated and delicate mechanisms, and in their present form at any rate they are very sensitive to lubrication. This in itself is not a fault, but it is obviously difficult to supply adequate lubrication to the working parts of a blower without blowing an excessive amount of oil through the inlet pipe and into the engine. 565

304

RACING MOTOR CAR DESIGN

Up to the present it has not been found possible to run superchargers of this type reliably and efficiently over speeds of about 5,000 r.p.m., and, for this reason, they are considerably heavier and more bulky than a Roots blower for the same duty. There is one other point in connexion with supercharging that is worth notice, and that is the mechanical and thermal effects upon the mixture of fuel and air as it passes through the blower. Modern multicylinder engines are called upon to function efficiently over a wide range of speeds, and the size of the gas passages has to be made sufficiently large to provide the least possible pipe loss at the highest speed in the range. This means that the gas velocity at the lower speeds has to be kept very low. Now the working mixture, as it emerges from the carburettor, contains very little fuel vapour, but is chiefly a moving column of air containing droplets of fuel, and not very small droplets at that. One of the biggest problems in the design of a multicylinder engine is to prevent these droplets from depositing on the walls of the pipe, and to distribute them evenly between the cylinders. Thus, when it becomes necessary for the inlet pipe to branch off to the various inlet ports, the inertia of these comparatively heavy particles of fuel may cause them to overshoot certain branches, thus starving certain cylinders and choking others. On the other hand, when this raw and imperfect mixture is led straight from the carburettor to a rapidly rotating blower, the violent mechanical agitation, coupled with the rise in temperature due to the sudden compression, certainly tends to make the charge more nearly homogeneous, and probably actually vaporizes a considerable portion of the fuel. In this condition, the charge behaves more nearly as a true gas, and even distribution among the cylinders is less difficult. It is sometimes found that, even with a blower so small that the delivery pressure is scarcely above atmospheric, the performance of the engine at low speeds, and consequently the acceleration of the car, are greatly improved, simply due to the pulverizing action of the blower on the fuel. The question is often asked, why little or no effort is made to cool the mixture, after compression, and before admitting it to the working cylinder. This is, of course, desirable theoretically, as it would increase the density of the charge, and therefore the amount of work obtainable from it. Attempts at intercooling have been made from time to time without, however, meeting with very marked success. It is obviously difficult to cool a mixture of this kind without producing deposition of the fuel spray and condensation of the vapour upon the cooling surfaces, and the disadvantages of such effects probably more than counterbalance the theoretical gain. 566

RACING MOTOR CAR DESIGN

305

Fuels. Finally, a few words on the most recent and by far the most important development of all—special fuels. As is well known, the racing motorist dearly loves a little mystery. In the old days, the mystery used to surround cams and valves and sparking plugs and so on. Now it is fuel—and the less he knows about it, the more mysterious he becomes ! It was mentioned above that one of the factors setting a limit to the output of an engine is the tendency of the charge to detonate when the compression ratio is increased beyond a certain point. This limiting compression ratio varies with the size of the cylinder and the design of the combustion chamber. Another limiting factor is the tendency to actual pre-ignition as the compression ratio is raised. Now both these factors depend, in turn, upon the nature of the fuel used, and it has long been known that certain grades of petrol, and also benzol, have a marked effect in delaying detonation. Of these, only benzol was commercially obtainable, and for a long time, therefore, racing engines were usually run on a fuel containing a proportion of benzol sufficient to eliminate detonation. The compression ratio was nevertheless always limited by the tendency to pre-ignition, the liability to which was not greatly affected by any fuels then available. The next step was the discovery of the effect of fuels having alcohol as their base. The pressure at which pre-ignition occurs is largely influenced by the temperature of the hottest part of the combustion chamber—usually either the exhaust valve or the sparking plug. The temperature of these in turn varies with the maximum temperature of the working charge—in other words, the flame temperature. If, therefore, the flame temperature can be kept down, the compression ratio can be correspondingly increased, and hence the power, although of course not in the same proportion. Although alcohol has a poor calorific value, it has two advantages in this connexion. First, it has a high latent heat of evaporation. Second, it is peculiar in that its vapour will burn when mixed with a good deal less air than is needed for complete combustion. Here it differs fundamentally from petrol, whose vapour will only burn satisfactorily when mixed with approximately the correct proportion of air. If, therefore, an engine is run on a very rich mixture of alcohol and air, so much heat is absorbed in the vaporization of the excess alcohol and in further heating the vapour, that the flame temperature is considerably reduced. If the compression ratio is then raised until the flame temperature under these conditions is just under the pre-ignition limit, then considerably increased power output can be obtained. For various reasons, among them the inability to start cold and to run throttled, pure alcohol is not a suitable fuel for a racing engine. 23 567

Fig. 18. Chassis of Napier-Railton Car

RACING MOTOR CAR DESIGN

307

In practice it is usually blended with suitable proportions of ether, benzol, and other constituents, according to the requirements of the particular engine. Incidentally, it is the presence, or absence, of these "other constituents" which constitutes the cherished mystery of the racing motorist. There is now available another agent, a derivative of lead, which enables the detonation point to be delayed, and which acts by controlling the rate of combustion rather than its ultimate temperature. The admixture of even very small quantities to ordinary petrol has a very marked effect in delaying the detonation point. Such mixtures are on sale everywhere under various brands bearing the prefix "Ethyl". This addition is often used in combination with alcohol mixtures in ultra high-compression engines. Finally, in the last eighteen months, great strides have been made in the commercial production of what are called "100-octane" fuels, having a very high resistance to detonation. The properties of such fuels have been known for some time, but it has only recently been found possible to synthesize them on a commercial scale. Such fuels include blends of iso-octane or isopropyl-ether with aviation petrol, and a small addition of tetraethyl lead. They have been intensively developed for the aircraft industry in America, chiefly for the sake of the reduced consumption and the resulting saving in fuel weight. There is little doubt that their use in a suitably designed racing engine will, before long, be the means once again of stepping-up the standard of performance. World Record Cars. We now come to the consideration of those unusual-looking cars which are built from time to time to attack the various world speed records. As was pointed out above, these vehicles are not racing cars at all, being far too ponderous and unhandy to compete with a road-racing car on its own ground. As, however, they represent this country's chief claim to distinction in the subject under review, some description of their general features may be of interest. The construction and running of these cars is nearly always a private venture, and this usually means that the money available is limited. Of the total cost of producing any new car, at least threequarters is absorbed by the manufacture and development of the engine. Naturally, therefore, anyone contemplating the construction of such a car will save a lot of money if he can buy the engine readymade. In this country we are fortunate in having a range of first-class watercooled aero-engines which are reasonably well suited by their general layout for installation in a motor car, They can be obtained in various 569

308

RACING MOTOR CAR DESIGN

shapes and sizes from 500 to 2,500 h.p. units, the smaller ones being suitable for the long-distance records (from 1 to 24 hours) and the large ones for the flying mile, or the so-called "land speed record". Full advantage has been taken of this source of supply (unique to this country), with the result that Great Britain has, for some years, figured prominently in the record book. It is true that a car (such as that shown in Fig. 18, p. 306, and in Fig. 19, Plate 3) built round these engines is not the ideal vehicle for the job; but it is safe to say that the ideal vehicle, involving as it would the development of a special power unit, would cost at least four times as much, and probably more. For this application the chief disadvantage of the aero-engine is that, owing to the necessarily low optimum airscrew speed and the high degree of reliability required, it usually has a comparatively low crankshaft speed. The drawbacks of this in a motor car are twofold, as it means (1) that the motor is rather bulky, increasing the size and weight of the car, and (2) that the transmission must also run at a slow speed and is therefore heavy. Weight by itself has only two disadvantages, but they are serious ones. It shortens the life of the tyres, and of the track upon which the car runs. The latter statement may sound absurd, but it is a fact that the chief limiting factor in longdistance record-breaking to-day is the ability of the track surface to withstand the destructive effect to which it is subjected. Its truth is demonstrated by Figs. 20 and 21, Plate 3. Fig. 20 shows a portion of the natural salt track in Utah after a 24-hour record run. The hole in the foreground is over 4 inches deep. Fig. 21 shows another and larger part, where the white salt surface has been completely blackened by rubber worn off the tyres. Long-Distance Record. The world's one-hour record now stands at 170 m.p.h. and the 24-hour record at 153 m.p.h. In the last two years these records have changed hands frequently between the three rivals (two British and one American) who are at present competing in that field. All three use cars having aero-engines of about 25 litres capacity. Two of the three are cars of normal construction, one being shown in Figs. 18 and 19. The other has front-wheel drive, which for this type of work appears to have certain advantages. All three are probably capable of raising the 24-hour record by another 20 rn.p.h. if a better track could be found. It is sometimes asked why the Germans and Italians with their 180 m.p.h. road-racing cars do not go out in their spare time and capture the long-distance records. There is no doubt that they have the speed, and their light weight would make them easy on the track 570

RACING MOTOR CAR DESIGN

309

and on tyres. On the other hand, their motors are not designed for prolonged full-throttle running (which is never required in a road race) and it is doubtful whether they would stand the strain of a 24-hour record attempt. The" Land Speed Record" Still further removed from the true racing car are those vehicles which, for many years now, have held the world's land speed record. Hitherto, considerations of cost have meant the use of an existing aero-engine and, as explained above, the resulting vehicle has been comparatively bulky and heavy by reason of its slowspeed engine. Criticism is sometimes levelled at cars like Sir Malcolm Campbell's Blue Bird (see Fig. 22) in that they are only (sic) a hundred miles an hour or so faster than other cars with a quarter their power and bulk. It is, of course, obvious that the use of high-speed engines of the same power specially designed to fit, would result in a considerably smaller, lighter, and therefore faster car. At the same time, it should not be forgotten that, other things being equal, the larger the car, the faster it will be. In other words, a 2,000 h.p. motor car with a "bulk" of (say) 200 cu. ft. will, other things being equal, be faster than an exactly similar car having 1,000 h.p. and 100 cu. ft. bulk. The same is, of course, true of ships and aeroplanes, and is exemplified by the fact that a model ship or aeroplane is necessarily slower than its prototype. The subject of the tyre equipment of these cars is perhaps the most interesting of all. The tyres used nowadays for this purpose are about 38 inches in diameter, and revolve at a maximum speed of about 2,700 r.p.m. They have each to support a load of up to 1 ton, with momentary shock loads of two or three times that figure. Neither the speed nor the load is individually a very serious problem. The combination of the two, however, does produce very grave difficulties. The weight on the wheel causes a deformation of the tyre casing at the point of contact with the ground, and, as the wheel revolves, this deformation travels round the tread. This continuous deformation of the casing represents a definite amount of work done. This work is converted into heat, not at the surface, but more or less uniformly throughout the fabric between the inner tube and the tread. As may be imagined, at speeds over 2,000 r.p.m., the amount of this work is very considerable, and, owing to the low heat conductivity of the material, the temperature of the fabric rises very rapidly and may soon reach a point at which the strength of the rubber is affected. This does not seriously affect the strength of the tyre as a whole, this being controlled by the cord structure which is unaffected by these 571

Fig. 22. Blue Bird, 1933 A B C D

Air intake to carburettors Upper water pipe Water tank 12-cylinder, 36,582 cu. cm. supercharged Rolls-Royce engine

E F G H J

Head rest Petrol tank filler Stabilizing fin Jack housing Torque reaction member

K Engine subframe L Dual Maries steering M Three-speed indirect-drive gearbox N Brake compensation O Propeller shaft driven by layshaft

RACING MOTOR CAR DESIGN

311

temperatures. On the other hand, the rubber tread, outside the cord structure, is held in position solely by its own adhesion, and if the strength of its inner surface deteriorates through heat, the whole tread will probably fly off under the very heavy centrifugal loading. For these very high-speed tyres, this trouble has been eliminated by the evolution of a grade of rubber showing a remarkable strength at high temperatures, and by the expedient of reducing the thickness of the tread to an absolute minimum, consistent with the necessary durability. Actually, the tread is less than 1 mm. thick. In spite of this, the quality of the rubber is such that, after use, the treads are often absolutely unmarked. Further, in order to reduce the deflexion to a minimum, the tyres are run at very high pressures—usually about 120 Ib. per sq. in. It is impossible to stress too highly the wonderful work which has been carried out in recent years by British manufacturers in the production of racing tyres. It is entirely due to this that the present record speeds are possible at all, and that tyre failures, in either races or record attempts, are now practically unheard-of. From having been at one time the most serious limiting factor in the design of racing cars, the tyres are now about the last item to cause anxiety. The chief problem with these ultra high-speed cars is how to reduce the aerodynamic drag to a minimum. By comparison with the aeroplane the problem is a difficult one, and the best of streamlined bodies produced so far are greatly inferior to a good modern aeroplane, when measured by the standard of the theoretical minimum drag. To begin with, a motor car must, by definition, have at least four wheels on the ground, and it is obviously difficult to do anything about the violent air disturbances which certainly exist in the region of these four zones of contact. Further, the close proximity of the ground to the under side of the car undoubtedly has an adverse effect of whose amount little is known, due chiefly to the difficulty of imitating the conditions exactly in a wind tunnel. Wind tunnel tests are, nevertheless, always made when developing these bodies, and, as between models of the same general type, they furnish quite a useful basis of comparison. Some idea of the magnitude of the wind resistance can be inferred from the fact that, at 300 m.p.h., even with a carefully streamlined car, the resistance approaches the limit of adhesion between the rear driving wheels and the ground. The air resistance is about equal to the resolved gravitational pull on a gradient of 1 in 5. The fact that a 5-ton car like the Blue Bird could (in vacua) climb a 1 in 5 gradient at 300 m.p.h., furnishes a graphic illustration of what a 2,000 h.p. motor car really means. Furthermore, it is easily seen that under such con573

312

RACING MOTOR CAR DESIGN

ditions any slight imperfections of the road would quickly induce wheel spin. For these conditions the obvious solution is to drive all four wheels, instead of the rear two only as has been done hitherto. There are considerable mechanical difficulties in the way of effecting a fourwheel drive with engines of existing design, but there is no doubt that it will be essential on these cars before long.

Fig. 24. Curve showing Speed against Length of Run during Acceleration Period, Blue Bird, Daytona, 1933 Finally, it is no use producing a car capable of beating the record if it cannot be accelerated up to, and slowed down from, its top speed, inside the total length of straight track available. Probably the best of the known tracks is the natural salt bed in Utah, U.S.A., where Sir Malcolm Campbell established the present record in 1935. Car and track are shown in Fig. 23, Plate 4. Here there is a 13-mile straight run, absolutely level and having a surface particularly kind to tyres, 574

RACING MOTOR CAR DESIGN

313

The "measured mile" is, of course, in the middle, which allows 6 miles for accelerating and another 6 for slowing down. This is ample for speeds up to 300 m.p.h., but would be distinctly on the low side for speeds in the region of 350 m.p.h., which will shortly be the objective. Such a car would have to reach 300 m.p.h. in about 3 miles, as it would need just about another 3 miles, or 6 miles in all, to reach 350 m.p.h. at the point where it entered the measured mile. Illustrating this point, Fig. 24 is the speed-distance curve of Blue Bird's run at Daytona in 1933. Here, the measured mile was between the fifth and sixth mile posts, and the average speed over it wbout 272 m.p.h. The braking arrangements would also need serious consideration. While it would not be difficult to provide the necessary braking effort, the weight of whatever mechanism were employed would, of course, affect the acceleration adversely. Speeds in the region of 230 m.p.h. have recently been accomplished on a new motor road in Germany which is dead straight for 30 or 40 miles. The car used is shown in Fig. 25, Plate 4. Provided it were sufficiently level, the use of such a road would simplify enormously the design of a 350 m.p.h. car. The author, however, fears that even if the road were sufficiently level to start with (and a car is sensitive in this respect in proportion to the square of its speed), the inevitable settling of its foundations would, in a short time, make it unsafe for such speeds. The question is often asked "What is the limit of speed on land ?" As the author has tried to show, the limit to-day and for many years to come is likely to be governed only by the roads or tracks available, and by the "state of the art" as it concerns the design of engines, tyres, streamlined shapes, etc. Looking further ahead, it can be said that the problem will become much more difficult, when, and if, the speed of sound (about 750 m.p.h.) is approached; and it seems likely that, as in the case of aircraft, the ultimate limit will be considerably below that figure. So far as the immediate future is concerned, there is considerable activity at the moment. Two or three cars are being prepared for attempts on the land speed record, and it is not unlikely that, in the course of the next three years, we may hear that a motor car has covered the flying mile at a speed of 6 miles a minute. For permission to reproduce drawings and photographs, the author wishes to express his grateful thanks to the proprietors of The Autocar, The Motor, The Automobile Engineer, and Speed.

575

RACING MOTOR CAR DESIGN

Fig. 1. All Four Wheels in the Air at 140 m.p.h.

Fig. 2. Mercedes Front Suspension [I.Mech.E., 1937]

576

Plate 1

Plate 2

RACING MOTOR CAR DESIGN

Fig. 6. Example of Tubular Frame.

Fig. 11. Modern Hooke Joint

Fig. 16. Centrifugal Blower (Offenhauser) [I.Mech.E., 1937] 577

RACING MOTOR CAR DESIGN

Plate 3

Fig. 19. Typical Long-Distance Record Car (Napier-Railton)

Fig. 20. Part of Track after 24 hours at 150 m.p.h. [I.Mech.E., 1937]

578

Fig. 21. Part of Track "Painted51 Black by Rubber worn off Tyres

Plate 4

RACING MOTOR CAR DESIGN

Fig. 23. Blue Bird on Bonneville Salt Flats, 1935

Fig. 25. Fully Streamlined Mercedes Car [I.Mech.E., 1937] 579

Bluebird.

580

Frank Whittle (1907-1996) Probably the most widely recognised of 20th century engineers, Whittle was born in Coventry. His father was the owner of a small engineering concern and the young Whittle showed a keen interest in technical matters generally, but aviation in particular. He attempted to enter RAF Halton as an aircraft apprentice in 1923, but was refused, not once but twice, on the grounds of his slight physique. Third time lucky, Whittle reported to Cranwell, commencing a glittering forces career. Whittle passed out as a pilot in 1928 and became a daring test-flyer. Written work was expected of officer cadets, and he obliged by a thesis on Future developments in aircraft design (1928). Here, Whittle predicted the requirement for high speed flight by means of rockets or propeller-driving gas turbines. These ideas were developed to the stage where Whittle could drop the propeller in favour of direct jet action. Approaches made to the Air Ministry for support met with rejection, but with the help of fellow pilot, W E P Johnson, Whittle patented his designs in the years 1930-1932. Recognising the special nature of their officer, the Royal Air Force sent Whittle to Cambridge University to study engineering, the Mechanical Sciences Tripos. He founded the company Power Jets Limited in 1935. Armed with a contract from the Thomson-Houston Company for the building of a gas turbine test engine, Whittle's vision could at last progress from the drawing-board. The W.I engine, as it became known, was flown from 15 May 1941 in the experimental Gloster-Whittle E.28/39 aircraft. Its research and development was fully described in Air Commodore Whittle's IMechE paper of 1945, delivered as the first James Clayton lecture. Such was the importance of the subject that the lecture was specially re-arranged to prevent priority of the announcement going to the United States. It was repeated several times, by Whittle, to record audiences. After the war. Whittle retired from the RAF. He became a consultant to BOAC and contributed to various engineering projects overseen by BristolSiddeley and Rolls-Royce, including a turbo drill for oilfields. Later, he served on the faculty of the US Naval Academy at Annapolis.

581

419

THE FIRST JAMES CLAYTON LECTURE

The Early History of the Whittle Jet Propulsion Gas Turbine By Air Commodore F. Whittle, C.B.E., R.A.F., M.A., Hon. M.I.Mech.E.* "I do not endeavour either by triumphs of confutation or pleadings of antiquity, or assumption of authority, or even by the veil of obscurity, to invest these inventions of mine with any majesty . . .". Francis Bacon. PART I. INTRODUCTION AND GENERAL OUTLINE Introduction. My purpose is to tell you briefly the early history of the development of the jet propulsion gas turbine in Great Britain, concentrating chiefly on the story up to the first flight tests in May 1941. The preparation of this lecture has not been easy, for in it I have tried to give you as clear a picture as possible of many years of work. In order not to fog you with detail, much of the story must be omitted, whilst in some passages the work of months must be dismissed in a single paragraph in order to preserve an accurate general impression. I will try to give some idea of the difficulties encountered and be frank about the mistakes that we made. Difficulties were by no means confined to the engineering side of the story; but of the large financial, administrative, and political problems which faced us from time to time I propose to say practically nothing. I shall say little about the aircraft side of the work, or of the more serious manufacturing problems which had to be solved. There have been many attempts to solve the gas turbine problem in the past. I will not attempt now to give the background of the early history of the gas turbine as a prime mover, for that has already been done many times. The records in the Patent Office would probably show that the problem of the gas turbine has engaged the attention of inventors almost to the same extent as perpetual motion. It will suffice to say that the constant pressure gas turbine is an old idea, and that there have been many attempts to produce a practicable engine of this sort, but at the time I started thinking about the subject, i.e. in 1928-9, the many failures had led to a general belief in the engineering world that it had no future. I now know that, apart from myself, there were others who refused to conform to the prevalent view and believed that the problems of the gas turbine were not insuperable. They included Dr. A. A. Griffith, Mr. H. Constant, and certain others at the Royal Aircraft Establishment, and engineers of the BrownBoveri Company led by Dr. Meyer. The main argument against the gas turbine was that the maximum temperatures permissible with materials available, or likely to be available, was such that the ratio of positive to negative work in the constant-pressure cycle could not be great enough to allow of a reasonable margin of useful work to be obtained, after allowing for the losses in the turbine and compressor. There seemed to be a curious tendency to take it for granted that the low efficiencies of turbines and compressors commonly cited were inevitable. I did not share the prevalent pessimism because I was convinced that big improvements in these efficiencies were possible, and, in the application of jet The MS. of this lecture was first received at the Institution on 21st June 1945. For the Minutes of Proceedings of the meeting on 5th October 1945, at which this lecture was delivered, see p. 379 ante. * Special Duty List, R.A.F., attached to Power Jets (Research and Development), Ltd., Whetstone, Leicester. The statements and opinions expressed herein are the personal views of the lecturer, and are not to be taken as in any way representing the opinions of the Air Ministry or the Ministry of Aircraft Production.

propulsion to aircraft, I realized that there were certain favourable factors not present in other applications, namely:— (1) The fact that the low temperatures at high altitudes made possible a greater ratio of positive to negative work for a given maximum cycle temperature. (2) A certain proportion of the compression could be obtained at high efficiency by the ram effect of forward speed, thereby raising the average efficiency of the whole compression process. (3) The expansion taking place in the turbine element of such an engine was only that which was necessary to drive the compressor; and therefore only part of the expansion process was subject to turbine losses. In order to make my story as clear as possible, I propose now to tell it in outline first, and then to fill in certain parts of it in greater detail by describing the design and testing of the three editions of the first experimental engine; then dealing with the design and testing of the first flight engine (the Wl), and an experimental version of it (the W1X); and ending with a few brief notes on two or three of the later engines. Outline. I first started thinking about this general subject in 1928, in my fourth term as a Flight Cadet at the R.A.F. College, Cranwell. Each term we had to write a science thesis, and in my fourth term I chose for my subject the future developments of aircraft. Amongst other things, I discussed the possibilities of jet propulsion and of gas turbines; but it was not until eighteen months later, when on an Instructors' Course at the Central Flying School, Wittering, that I conceived the idea of using a gas turbine for jet propulsion. I applied for my first patent in January 1930. The principal drawing of the patent specification as filed is reproduced in Fig. 1. It may be seen that I tried to include the propulsive duct, or "athodyd" as it has since been called, but this had been anticipated at least twice, so the upper drawing and relevant descriptive matter had to be deleted from the specification. The idea was submitted to the Air Ministry, but was turned down on the ground that as it was a gas turbine the practical difficulties in the way of the development were too great. During 1930 I tried to interest various firms in the scheme, but met with no success; for the most part they thought the same way as the Air Ministry. It is probably also true that in their view the prevalent industrial depression made it anything but a favourable moment for expensive ideas of this sort. Nothing very much happened for a few years. I gave up hope of ever getting the idea to the practical stage, but continued to do paper work at intervals, until, in May 1935, when I was at Cambridge as an Engineer Officer taking the Tripos Course, I was approached by two ex-R.A.F. officers (Mr. R. D. Williams and Mr. J. C. B. Timing), who suggested that they should try to get something started. Though I had allowed the original patent to lapse through failure to pay the renewal fee, and though I regarded them as extremely optimistic, I agreed to co-operate. I thought that there was just a bare chance that something might come of it.

583

JAMES CLAYTON LECTURE 420 We eventually succeeded in coming to an arrangement with a chamber, instruments, and some accessories) with The British firm of investment bankers (Messrs. O. T. Falk and Partners) Thomson-Houston Company in June 1936. The engine was to be a simple jet propulsion gas turbine which led to the formation of Power Jets, Ltd., in March 1936. Before Power Jets was formed, O. T. Falk and Partners ob- having a single-stage centrifugal compressor with bilateral tained the opinion of a consulting engineer (Mr. M. L. Bramson) intakes, driven by a single-stage turbine directly coupled. who gave a wholly favourable report. The initial sum subscribed Combustion was to take place in a single combustion chamber through which the working fluid passed from the compressor was £2,000, and with this we cheerfully went ahead. to the turbine. We were going beyond all previous engineering experience Fig.l. in each of the major organs. We were aiming at a pressure ratio of about 4/1 in a single-stage centrifugal blower when at the time, so far as we knew, a ratio of 21/2/1 had not been exceeded. We were aiming at a breathing capacity in proportion to size substantially greater than had previously been attempted. The combustion intensity we aimed to achieve was far beyond anything previously attempted. Finally, we had to get over 3,000 s.h.p. out of a single-stage turbine wheel of about161/2inches outside diameter, and to do it with high efficiency. At first, our intention was to do the job stage by stage—that is, to make a compressor and test it; then to add a combustion chamber to the compressor; then to test a turbine alone; and finally to build an engine—but we very soon realized that this was likely to be a long and costly process and we decided to go FIG. 2, for a complete engine right away. This first engine was based on a design for flight, but was not intended for flight; and though it was designed to be very light by normal engineering standards, we did not put forth special efforts to make it as light as possible. I was fairly confident in the compressor and turbine elements but felt rather out of my depth with the combustion problem, and so (in 1936) I visited the British Industries Fair with a view to enlisting the aid of one of the oil burner firms, but the requirements I specified were considered to be far too stringent by most of them until I met Mr. Laidlaw, of Laidlaw, Drew and Company. Though he recognized that we were aiming at something far in advance of previous experience in this field he considered the target possible of attainment, and so it was with his help that we attacked the combustion problem. While the engine was in course of design and manufacture, FIG.3 we carried out a number of combustion experiments on the premises of The British Thomson-Houston Company, with 1 At apparatus supplied by Laidlaw, Drew and Company, until we considered that we had enough information to design a combustion chamber. Power Jets therefore placed the contract for the design and manufacture of the combustion chamber with Laidlaw, Drew and Company. By this time the Tripos Examinations at Cambridge were over, and the Air Ministry had agreed that I should stay for a post-graduate year. This was really a device to enable me to continue work on the engine, and so a considerable proportion of my time was spent at The British Thomson-Houston Company's works in Rugby. Testing of the engine commenced on 12th April 1937, and continued intermittently until 23rd August. These early tests made it clear that the combustion problem was by no means solved, and that the compressor performance was far below expectations. Nevertheless they were sufficiently encouraging to show that we were on the right track. Testing was then stopped in order to carry out a major reconstruction; and once more The British Thomson-Houston Fig. 1. Reproduction of Drawings Illustrating British Patent Company did this work for Power Jets. No. 347,206, filed 16th January 1930 At this time it was also decided that no further running The upper drawing—the thermo-propulsive duct—had to be deleted should take place in the Rugby factory, as it was considered too from the specification. dangerous, and so Power Jets came to an arrangement with The British Thomson-Houston Company by which they rented The President of the Air Council was a party to the agreement part of that firm's old foundry at Lutterworth (Ladywood Works) which resulted in the formation of Power Jets, and the Air for their future testing. At this point I think it will be of interest to explain the basis Ministry was a shareholder from the start in that a proportion of the shares allotted to me was held in trust for the Department. on which I was working with The British Thomson-Houston During the negotiations leading to the formation of Power Company. I was allowed virtually complete access to all parts Jets, I was working on the designs of an experimental engine. of their turbine factory and its contributory departments, and Messrs. O.' T. Falk and Partners placed an order with The as time passed I was given an increasingly free hand. I was British Thomson-Houston Company, Ltd., for engineering and provided with an office and had access at all times to Mr. F. design work in accordance with my requirements in advance of Samuelson, M.I.Mech.E., Chief Turbine Engineer, to Mr. R. H. the formation of the new company. Power Jets placed the order Collingham, M.I.Mech.E. (who succeeded Mr. Samuelson as for the manufacture of the engine (except the combustion Chief Turbine Engineer when the latter retired), to his staff

584

THE E A R L Y HISTORY OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E of specialist engineers, and to the Drawing Office. I spent much time following the manufacture and investigating costs (as all the work was being done on a "cost-plus" basis) and I handled the controls whenever the engine was tested. All this was very valuable experience for me and, I believej helped the work along very considerably. The story from this point becomes increasingly complex, so to help me to tell it clearly I will refer you to Fig. 2. This shows the time relationship of some of the major events in the history, After the first series of tests in 1937 the Air Ministry (on the recommendation of the Aeronautical Research Committee) began to take more notice of the work and decided to place contracts with Power Jets for a report on the first series of tests

421

While the reconstruction of the engine was in progress we carried out another series of combustion tests until once more we thought we had solved the combiistion problem, The reconstructed engine was ready to run again on 16th April 1938, and the testing (at Lutterworth) continued intermittently until 6th May, on which day the engine was severely damaged by a turbine blade failure. Though the total test running time was only about five hours, it was evident that the combustion problem was by no means solved. A second major reconstruction was decided upon, and the work was once more carried put by The British ThomsonHouston Company, but this time the Air Ministry agreed to bear the cost (by contract with Power Jets). Once more we

Fig. 2. Diagram Illustrating the Principal Events in the Early History of the Jet Propulsion Gas Turbine and for further research running. However, the work was still regarded as "long-term research". At the same time further private money was being raised; but the total financial resources available were still very small and the work had to be done on the basis of the most stringent economy. This proved expensive in the end, because it meant that we were continually having to use parts which really ought to have been scrapped. So far, there had been no official secrecy requirement on the work; but secrecy was a condition of the Air Ministry contracts, and when it was imposed it naturally made the raising of private money very difficult. (Incidentally, we received rather a shock when in 1939 we found that about six of my patent drawings had been reproduced in an article in the German journal Flugsport, but there was nothing very surprising in this because there was nothing to stop anybody from seeing the published patent drawings. As a matter of interest some of these are reproduced with their German captions in Fig. 3.) The Air Ministry further helped the project by placing me on the Special Duty List at the completion of my post-graduate year in the summer of 1937, to enable me to continue work on the engine on a full-time basis.

embarked on a further series of combustion tests while the reconstruction was in progress. Testing of the third model of the engine commenced at the end of October 1938 and continued intermittently until February 1941 when it was damaged beyond repair by a turbine failure, It had by then served its purpose, and had provided an enormous amount of information on which subsequent designs were based. In the summer of 1939 the Air Ministry ceased to regard the development as a matter of long-term research and came to accept the fact that we had the basis of a practicable aero-engine, As a result, Power Jets received a contract for a flight engine, and a short time afterwards a contract was placed with the Gloster Aircraft Company for the manufacture of an experimental aeroplane to specification E28/39. Thereafter, Power Jets and the Gloster Aircraft Company worked in close collaboration, The Ministry also agreed to purchase the experimental engine, and to cover all test running by contract. Power Jets subcontracted the drawing and manufacture of the flight engine with The British Thomson-Houston Cornpany. The design team now included a few engineers recruited by Power Jets as well as British Thomson-Houston engineers 27

585

422

JAMES C L A Y T O N L E C T U R E

Fig. 14. Whittle 1935. Exhaust Jet Drives a Gas Turbine Compressor

Translation

Figs. 62-65. Whittle 1935: Thermal Jet Propulsion with Axial Turbine

Fig. 66. Whittle 1936: Twin Propulsion System

Fig. 67. Whittle 1936: Thermal Jet Propulsion with Piston Engine and Turbine

Fig. 3. Reproductions of Illustratioas from Flugsport (1939) and myself. This first flight engine was known as the "W1". In the course of manufacture of the Wl, certain major components were considered to be unairworthy on completion, and it was decided to use these—with certain spare components made for the first experimental engine—to build an "early edition" of the Wl. This was known as the "W1X". The W1X was delivered, loosely assembled, in November 1940. Power Jets had by this time begun to build up a small sheetmetal shop as well as the nucleus of an engineering team. They had also installed combustion test apparatus at the Ladywood Works; this work also was covered by contracts from the Ministry of Aircraft Production. The W1X, though falling short of expectations, proved to be far in advance of the experimental engine, and a number of tests were carried out on it which provided experience on which modifications to the Wl were based. When the Wl was delivered,

586

it was stripped by Power Jets, and was modified to incorporat the latest experience gained with the W1X. Eventually, afte certain preliminary testing, the W1 was put through a 25-hour: special category test to clear it for flight. Meanwhile the E2 aeroplane was completed by the Gloster Aircraft Company, an the W1X was installed in it for taxi-ing trials in April 1941. I the course of these trials the aircraft did in fact leave the groun for a short straight hop. Flight trials with the Wl began on 14th May 1941. Th engine had been cleared for ten hours' flying, and the programm of flight testing laid down was completed in fourteen day without special incident. The test pilot was the late Flight Lieutenant P. E. G. Sayer. I have now to take you back a year to explain that at the beginning of 1940 the Air Ministry (later the Ministry of Air craft Production) began to work on the assumption that there was a good chance of getting jet propelled fighter aircraft into

THE EARLY HISTORY OF THE WHITTLE JET P R O P U L S I O N GAS TURBINE production in time for use in the war. As a result the following steps were taken:— (1) Power Jets were authorized to go ahead with a more advanced engine (the "W2"). (2) The Gloster Aircraft Company was authorized to proceed with the design of a twin-engined interceptor fighter (the "F9/40"—prototype of the Meteor). (3) Direct contracts were placed with The British ThomsonHouston Company and other firms for the manufacture and development of jet propulsion gas turbines. (4) It was decided that Power Jets should become a research and development organization and that they were to supply all other firms engaged with all necessary drawings and any other information they needed to assist them in their work. I refer you once more to Fig. 2 for the relationship in time of the first few of the types subsequently made, and would add the following series of notes to amplify Fig. 2:— (1) By arrangement between the British and American Governments, the W1X, a set of drawings of the W2B, and a small team of Power Jets' engineers, were flown over to America in the autumn of 1941, and this initiated the intensive development of the jet propulsion gas turbine at the General Electric Company's works at Lynn, Mass. (2) The "W1(T)" was a modified Wl built for bench development from spares. (3) The "Wl(3)" was also a modified Wl. (4) The "W1A" was an engine incorporating most of the features of the Wl but it also had in it certain of the special features of the W2 which it was desired to test in advance. This engine was manufactured by The British ThomsonHouston Company to Power Jets' requirements. (Power Jets were contractors to the Ministry of Aircraft Production.) It was flown in the E28. (5) The "W1A(2)" was a modified W1A. (6) "W2". Drawings of this engine were handed over to the Rover Company who made it as direct contractors to the Ministry of Aircraft Production. As made, it differed from the Power Jets' design in certain mechanical features. We realized before this engine was completed (as a result of doing more detailed calculations) that if the design assumptions in respect of component efficiencies were not achieved, the penalty would be very severe. This in fact proved to be the case. It was a failure. (7) The "W2 Mk. 4". Power Jets, having received a direct contract for the W2, subcontracted the manufacture to The British Thomson-Houston Company, but when it was realized that the design was extremely sensitive to the design assumptions, it was changed by stages to bring it nearer to that of the W2B, and when it was delivered by the subcontractors it was further modified by Power Jets so that when testing commenced it differed only in detail from the W2B design. This engine was totally wrecked by the bursting of a faulty impeller forging, though not before a fair amount of useful testing had been done. (8) The "W2B" was originally designed by Power Jets, and complete sets of Power Jets' drawings were passed to the several firms by then engaged. The first and second W2Bs to be tested by Power Jets were manufactured by the Rover Company, and (like the W2) differed from the Power Jets' design in certain mechanical features. The W2B was the prototype of the Welland engines (which subsequently powered the Meteor I in this country) and of the "Type I" (the corresponding engine made by the American General Electric Company which powered the Bell Airacomet, and which was the forerunner of a series of similar engines made by the American General Electric Company). (9) The "W2B Mk. 2" was a modified W2B, and was the first engine in which Power Jets did the bulk of their own manufacture. The whole project received a powerful stimulus from the successful flight trials of the E28. The Ministry of Aircraft Production decided to plan for production of the W2B and the Meteor. Many sceptics were converted. The firms already en-

423 gaged increased their activities considerably, and other firms which had practically ignored the gas turbine suddenly evinced a lively interest. A noteworthy event was the formation by the Ministry of Aircraft Production of the Gas Turbine Collaboration Committee in November 1941. The Ministry of Aircraft Production, the Admiralty, the Royal Aircraft Establishment, the National Physical Laboratory, and the firms engaged on aircraft gas turbines were all represented on this Committee. Dr. H. Roxbee Cox, who as an official of the Ministry of Aircraft Production had been closely associated with gas turbine development from early in 1940, was primarily responsible for the formation of this Committee and was its chairman from its formation until he became Chairman and Managing Director of Power Jets (Research and Development), Ltd., in April 1944. The Committee proved to be an institution of great value, and it served to ensure a large measure of co-operation amongst the firms engaged. This leavening of competition with co-operation is a very healthy thing and worthy of preservation in peace time. I will conclude this outline by saying that the types of engines which trace their descent directly from the W2 and W2B have undergone continuous development, and the original design performance for that size of engine has been far exceeded. PART II. THE DESIGN AND TESTING OF THE EXPERIMENTAL ENGINE

The Design and Testing of the First Model. (a) Design. The first engine was designed with a specific target in mind. It was a very optimistic one, but nevertheless it formed the starting point and is worth recording. The aim was to propel a very "clean" little aeroplane of about 2,000 Ib. "all up" weight at a speed of 500 rn.p.h., at a height of the order of 70,000 ft. This speed was estimated to correspond to that of minimum drag at that height, i.e. this high speed was also the most economical speed for the height. It was estimated that a net thrust at this height of 111 Ib. would be required. The size of engine corresponding to these data was considered to be the smallest in which the necessary accuracy of machining could be obtained without excessive manufacturing costs. The design assumptions and leading particulars are given in Table 1, and in the pressure-volume cycle shown in Fig. 4. TABLE 1. LEADING PARTICULARS OF FIRST EDITION OF EXPERIMENTAL ENGINE Dimensions are in inches. Compressor impeller— Tip diameter . T i p width . . . Outer diameter of eye Inner diameter of eye Number of blades Material . . .

. . . . . . .

Compressor casing—

Inner diameter of scroll Material . . . .

Turbine—

. . . .

.19 2 10.75 5.5 30 Hiduminium RR56 31 Hiduminium RR55, DTD.133B.

,

Mean diameter of blades Blade length . . Material of blades . Material of disk . . Blade chord . , Number of blades. . Maximum speed, r.p.m. .

. 14 . 2.4 . Firth-Vickers Stayblade . Firth-Vickers Stayblade .0-8 . 66 .

17,750

Figs. 5, 6, 10, and 12 illustrate various features of the design, which are further amplified in Figs. 7, 8, 9, 11, and 13, Plates 1 and 2. For the most part, I rely on Table 1 and Figs. 4-13 to convey the necessary information on the design; but my comments on certain special points are given below. The assumption of 80 per cent adiabatic efficiency for a

587

424

JAMES CLAYTON L E C T U R E centrifugal compressor running at a tip speed of 1,470 feet per sec. was very optimistic indeed, and received a good deal of criticism; but I felt confident that we could design to avoid many of the losses which were occurring in all centrifugal compressors of which I had knowledge at the time. The general

Fig. 5. Assembly of First Model of Experimental Engine

VOLUME—CU. FT.

Fig. 4. The Pressure-Volume Diagram, Design Assumptions, etc., on which the Initial Design of the Experimental Engine was Based.

We aimed at having as many blades on the impeller as manufacturing limitations would permit, in order to keep the blade loading as low as possible. In particular, it was hoped that by keeping the pitch-chord ratio of the rotating guide vanes small we should avoid stalling at the intake, as I believed then—and still believe—that this is one of the main sources of loss in centrifugal compressors.

Sea level cycle. Assumptions:—Compressor efficiency, 80 per cent. Turbine (shaft) efficiency, 70 per cent. Axial velocity at turbine exhaust, 800 feet per sec. = 14 per cent of heat drop. Efficiency at final expansion, 97 per cent. Weight of air, 26 Ib. per sec. Weight of fuel, 0.3635 Ib. per sec. = 168 gal. per hr. approx. Power to drive compressor, 3,010 h.p. Static thrust, 1,389 1b. Products of combustion, per sec. N2 = 19.963 1b. = 0.713 1b. mol. O2 = 4.763 „ =0-149 „ CO2 = 1.122 „ = 0.0255 „ H2O = 0.516 „ = 0.0286 „ Total

26.364 „

0.9161

Fig. 6. Test Assembly of First Model of Experimental Engine

„

Mean molecular weight = 28.78 Value of R (gas constant) = 96.5. g = 1.-4 for compression and 1.379 for expansion. Kp = 0.24 for compression and 0.25 for expansion and combustion. Latent heat of fuel, 75 C.H.U. per 1b. Calorific value, 10,500 C.H.U. per 1b. State point

Pressure, 1b.per sq. inch

Temperature, deg. C. abs.

Volume, cu. ft. .

Velocity, ft. per sec.

A B C D E F after reheat G

14.7 64.6 64.6 64.6 23.45 23.45 14.7

288.0 439.5 477.5 1,052.0 795.8 836.7 737.0

339.5 118.0 128.2 287.9 6000.0 631.0 885.0

0 0 200 200 2,400 900 1,720

view was that we should be fortunate if we got 65 per cent adiabatic efficiency. We went for the double-sided compressor because we wanted to get the greatest possible breathing capacity in proportion to size. I also counted on this feature to give a reduced proportion of skin friction losses.

588

No diffuser blades were fitted to the blower casing at first. Two stages of diffusion were aimed at. The intention was to obtain partial diffusion in the bladeless vortex space between the impeller tips and the scroll and to convert most of the remaining kinetic energy into pressure in the "honeycomb" diffuser through which the air passed from the compressor scroll to the combustion chamber. The turbine nozzle arrangement was very unorthodox, as no nozzle blades were used. The idea was that most of the expansion took place in the convergent-divergent entry to the

Fig. 10. Combustion Chamber Arrangement at Commencement of Testing of First Model of Experimental Engine The pilot jet and the igniter are omitted.

THE EARLY HISTORY OF THE WHITTLE JET PROPULSION GAS TURBINE nozzle scroll which was shaped to cause the discharge of the gases through its annular mouth with constant axial velocity, the whirl velocity corresponding to that of a free vortex, i.e. constant angular momentum. This nozzle design was the source of considerable controversy; and though I am not very proud of it now, I thought it a good idea at the time. It is of considerable interest in retrospect, because it became evident, later, that I had not succeeded in conveying to others what I had in mind. Air tests were made on a half-scale model of this nozzle, and though very rough, they seemed to show that it would behave approximately as expected. The design of the rotor assembly needed much careful thought. It was considered necessary to use unbored forgings for both the turbine wheel and compressor impeller, also to use an overhung turbine rotor, as it was thought that the provision of a satisfactory bearing support on the exhaust side of the turbine would be very difficult. The bearing housing assembly which contained one of the two water jackets for the turbine cooling could not be split in the diametral plane. These considerations governed the rotor design. I have already referred to my meeting with Mr. Laidlaw and to our collaboration in the design of the first combustion chamber and also in the preliminary testing which provided the basis

425

slight damage was made good and the test was subsequently repeated satisfactorily. It was, however, a rather depressing start. As already mentioned, the testing of the engine under its own power began on 12th April 1937. The very first attempt to start was successful in that the engine ran under its own power, but it accelerated out of control up to about half its designed full speed. This happened several times; and altogether it was a very alarming business, so much so that in the early days the individuals in the vicinity did more running than the engine. The starting procedure was as follows. The engine was motored at about 1,000 r.p.m. and the pilot jet (which injected an atomized spray) was switched on and ignited by means of a sparking plug and hand magneto. The motoring speed was then raised to about 3,000 r.p.m. and the main control opened slowly. The engine would then accelerate under its own power; but, as I say, not always under control. The explanation of the first few uncontrolled accelerations was simple when we found it, and may be understood by reference to the diagram of the early fuel system shown in Fig. 12. If the spill line from the burner was not full of fuel the needle valve of the burner would be forced into the fully open position when the fuel pump ran. We were frequently breaking the fuel line and doing various motoring tests so that often, unknown to us, there was a "lake"

Fig. 15. Experimental Engine (First Model): Arrangement of Combustion Chamber when converted to Downstream Injection The pilot jet and the igniter are not shown. Fig. 12. The Fuel System used in the Early Tests of the First Model of the Experimental Engine for the design. These tests were done on a site just outside the British Thomson-Houston turbine factory, using air supplied by a larger blower in the factory. The "set-up" for these tests is shown in Fig. 14, Plate 2. The tests were very crude but they did at least show that the required combustion rate could be obtained, though (as it proved) they were by no means sufficient to avoid combustion trouble in the engine subsequently. (b) Testing. For test purposes the engine was mounted on a four-wheeled trailer. This trailer also carried the starter motor, instrument panel, and controls, making the whole set selfcontained except for the fuel supply from the fuel tank and the water supply. It was intended to measure the thrust by a spring balance connecting the trailer to a fixed abutment, but in the testing of the first model no thrust measurements were taken. The original intention was to start the engine by means of a two-cylinder air-cooled aero-engine, but the torque fluctuations with this engine were found to be too severe, and a 20 kW. d.c. motor was used instead. This was mounted on the trailer, and because of its weight the wheels of the trailer were removed and the axle supported on blocks. The drive was direct to the main rotor and operated through an automatically releasable dogclutch in which a driving pin was forced out of engagement as soon as the engine overran the starter motor. Preliminary motoring tests were carried out by using a special air nozzle to drive the turbine with the compressed air supply available in the British Thomson-Houston turbine factory. We obtained sufficient power this way to motor the engine up to 10,000 r.p.m. with the compressor intakes blanked off. In the first of these tests the blower impeller fouled its casing, but the

of fuel in the combustion chamber. Other uncontrolled accelerations were caused by the sudden opening of the burner needle valve after initial sticking; by loss of temper in the burner spring through overheating, etc. Fortunately none of these uncontrolled accelerations took the engine beyond about 9,000 r.p.m. No trouble was ever experienced in starting except when the ignition failed through cracked electrodes, or when mistakes were made in the assembly of the fuel lines. The overheating of the burner already referred to was a serious problem; and as a result of it a fairly drastic change in the combustion system was made. We changed over to downstream injection as shown in Fig. 15. Five runs were then made with this system up to maximum speeds of about 8,500 r.p.m., but the combustion was so bad that this speed could not be exceeded. Any further opening of the control seemed to result only in the burning of more fuel aft of the turbine. One attempt was made to use the "Primus" principle in the combustion system, i.e. the combustion chamber was fitted with a vaporizer, and we changed from Diesel fuel to kerosene. The pilot jet was relied on for the initial heating of the vaporizer. Only one attempt to run was made with this arrangement and once more there was uncontrolled acceleration, believed to be due to priming of the vaporizer through insufficient heating surface. Though we changed back to liquid injection after this one effort at vaporized injection, we continued to use kerosene as fuel. Many attempts to improve the combustion were made by a series of modifications to the combustion chamber; and some improvement was achieved—we managed to get up to a speed of 11,000 r.p.m. for ten minutes. This series of runs ended when the compressor impeller fouled the casing after running for four minutes at 12,000 r.p.m. The damage to the compressor and impeller casing was only

589

426

JAMES CLAYTON LECTURE slight; but as it had now become clear that the delivery pressure blower, combustion chamber, and the flow path between the from the compressor was much below expectations, we decided compressor scroll and the combustion chamber was made, most to fit diffuser blades (Fig. 16) in an effort to improve the blower of them with little improvement, but subsequently we succeeded performance before further testing. We then managed to attain in reaching a speed of 13,600 r.p.m. with the blower fitted with a a speed of 13,000 r.p.m., but the compressor performance still modified set of diffuser blades. At this point in the tests the Chief Engineer of The British left much to be desired. Moreover, the combustion had Thomson-Houston Company considered it unwise to run at speeds higher than 12,000 r.p.m. in the open factory, and this was the speed limit for the remaining tests of the first model. A return to upstream injection was made, the burner now being insulated against overheating by using a fuel cooling arrangement as shown diagrammatically in Fig. 17. Combustion appeared to be improved and for the first time no part of the casings reached glow heat at speeds of up to 12,000 r.p.m. Testing was now suspended for the following reasons: first, because the speed restriction of 12,000 r.p.m. made it necessary to find a new site for running at higher speeds; and second, because it was decided to make major modifications to the general arrangement. The tests so far had been very disappointing, but there were many encouraging features. We had demonstrated that there was no particular difficulty in starting or in control. There was also plenty of evidence which suggested that the whole scheme was well worth developing, though it had become obvious that much hard work lay ahead. The principal defects of that particular arrangement were shown to be:— (1) Poor compressor efficiency. (2) Excessive preheating of the air to the rear intake, owing to the disposition and temperature of the combustion chamber. (3) Very unsatisfactory combustion. (4) Excessive frictional loss in the unorthodox turbine nozzle scroll. No readings of sufficient reliability had been obtained from which any estimate of the efficiency of the turbine could be Fig. 16. Diffuser Blades fitted in Course of Tests on First made, but it seemed practically certain that it was well below Model of Experimental Engine that assumed in the design. deteriorated, and as this was believed to be due to the nature of the flow from the compressor scroll to the combustion chamber, many modifications to improve this flow were made, but without noticeable improvement. There was evidence of a flow reversal in the elbow, but we did not realize how severe this was until (in one test late in the series) flames were seen through a small hole which had been drilled in the neck of the blower casing scroll. A series of rapid modifications to the diffuser system of the

Fig. 17. Burner for Upstream Injection, insulated from Overheating by Cooling Jacket of Fuel: First Model of Experimental Engine

590

The Design and Testing of the Second Model. (a) Design. The main aims of the reconstruction were to obtain a single straight combustion chamber and to provide a much improved diffuser system for the blower. In particular complete symmetry about the axis was aimed at. Features of the new arrangement are shown in Figs. 18 and 20, and are further amplified by Figs. 19, 21, and 22, Plate 3. It should be noted especially that the flow through the turbine was reversed and that we hoped to get a certain degree of heat exchange by passing the ten exhaust pipes through the annular space between the outer casing and the combustion tube. We used as much as possible of the original engine because we could not afford to do anything else. For example, the same compressor casing was used with the scroll machined off and replaced by ten fabricated discharge pipes (Fig. 21). I have no doubt now that it would have been cheaper and quicker to have had a new blower casing. We had decided to use the "Primus" principle for combustion, and while the reconstruction was proceeding we carried out a further series of tests on combustion chambers at the British Thomson-Houston factory, once more with the assistance of Mr. Laidlaw. Many arrangements were tried before we decided to proceed with one having a vaporizing coil with a cluster of ten vapour nozzles injecting upstream towards the mouth of the flame tube. The primary air entered the flame tube through two concentric rings of swirl vanes of opposite hand. Excess air entered the flame tube through a system of holes in its side. There was another important difference in the turbine design, apart from the fact that the direction of flow through it had been reversed. Hitherto I had left the detail design of the turbine blades almost entirely to the British Thomson-Houston engineers but now, by a strange chance, I discovered that there was a fundamental difference between my ideas and those of the steam turbine world in general. I have already said that the original bladeless turbine nozzle scroll was intended by me to provide a rotating annular flow of

THE E A R L Y H I S T O R Y OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E

427

Fig. 18. Assembly of Second Model of Experimental Engine constant angular momentum. I believed that this was the effect aimed at in any case in normal practice. Working from first principles I had taken it for granted that it was well known that, in theory at least, the ideal flow from a complete nozzle ring should be one of constant angular momentum. I was astounded to find that on the contrary it was the general practice to assume that the flow from a turbine nozzle ring was considered to be a series of straight jets of substantially constant velocity and

Fig. 20. Revised Pressure-volume Diagram and Design Assumptions on which the Second Model of the Experimental Engine was Based Sea level cycle. Assumptions:—Compressor efficiency, 70 per cent. Turbine (shaft) efficiency, 70 per cent. Axial velocity at turbine exhaust, 1,020 feet per second. Efficiency of final expansion, 90 per cent. Weight of air, 26 1b. per sec. Weight of fuel, 0.3635 1b. per sec. g — 1.4 for compression = 1.379 for expansion. Kp = 0.24 for compression = 0.25 for combustion and expansion. Static thrust, 1,240 1b.

pressure. My discovery of this difference in outlook made much clearer to me why there had been so much argument over the original nozzle scroll. It appeared, therefore, that the blades of the first turbine had been provided with angles based on standard practice; and that though they had a substantial twist, this was only to allow for the radial variation of blade speed. I therefore took care to ensure that the nozzle and turbine blades used in the second model were designed to conform to a flow of constant

angular momentum from the nozzle ring. This meant that there was about twice as much twist on the blades as there was before. I do not want to elaborate on this very important point; it will suffice to say that the concept of vortex flow is now fairly well established and it has remained the foundation of Power Jets' practice in turbine design. It is only fair to add that I subsequently found that the engineers of the Royal Aircraft Establishment had already accepted this concept, and there is now evidence that more than one turbine firm was thinking similarly at that time. In the redesign of the engine, the design assumptions were substantially modified. A blower efficiency of 70 per cent was now assumed, and the revised pressure-volume diagram is shown in Fig. 20. The revised figure for the thrust was 1,240 1b. at full speed. (b} Testing. All further testing was done at the Ladywood Works at Lutterworth, though Power Jets continued to rely on The British Thomson-Houston Company to a very large extent for labour and modifications. I frequently drove over to the British Thomson-Houston factory to collect test hands, fitters, sheetmetal workers, etc., as the situation required from time to time (though by now Power Jets employed an assistant engineer and a watchman). We were now using a 10 h.p. motor car engine as starter motor, which was mounted on the trailer complete with its own radiator, petrol tank, etc. Only nine test runs were made on the second model of the experimental engine, because the ninth test was brought to an end by a turbine failure which caused fairly extensive damage. For eight of these test runs the speed never exceeded about 8,500 r.p.m., but in the last run of the series a speed of 13,000 r.p.m. was maintained for half an hour, at the end of which the failure occurred. A thrust of 480 1b. was recorded. Many troubles were encountered, many of them due to defects in the fuel system. There were failures of the separately driven fuel pump, hunting due to faulty relief and control valves, etc., but the main trouble was due to a serious fault in the design for which I was wholly to blame. The nozzle ring assembly was not adequately supported against the difference of pressure between the combustion chamber and the turbine housing, and though modifications to stiffen it were made in the course of the tests these were inadequate, and when the engine was dismantled after the turbine failure it was found that the nozzle assembly had rubbed very heavily on the turbine wheel at the blade roots. It was thought at first that the blade failure was due to the heavy rub mentioned above, but subsequent examination of the blades led to the conclusion that they had been subjected to very uneven gas temperature, of such a nature that there was a narrow annular belt of extremely hot gas impinging on the blades at the point of failure, and that the failure was therefore due to the running stress combined with excessive temperature.

591

428

JAMES C L A Y T O N L E C T U R E Until this uneven temperature distribution became evident, Testing began at the end of October 1938 and continued with the combustion had seemed to be fairly satisfactory. The flame increasing intensity until the engine was wrecked by the failure never appeared to be more than about 2 feet long, and com- of the turbine disk in February 1941. bustion seemed to be complete; but it was now evident that the Practically the whole of the testing was dominated by the mixing of cool secondary air and hot combustion products was combustion problem. Not only were a great many combustion very inadequate. modifications tested in the engine, but these were supplemented Apart from a considerable amount of secondary damage by a much greater number of rig tests which took place in resulting from this failure, it was also found that there were a parallel, at first (as before) on the premises of The British large number of cracks in the buttress ribs of the impeller blades. Thomson-Houston Company and later at the Ladywood Works at Lutterworth. The Design and Testing of the Third Model, (a) Design. For two years we struggled with a combustion system based Another major change in the arrangement of the engine was upon the pre-vaporization of the fuel, but eventually had to decided upon. This was in many ways a compromise between abandon it. I cannot possibly give an adequate picture of that the two arrangements already tried. The main feature was the heartbreaking period. What I have to say below is a bare outline use of multiple combustion chambers. The general design of only. Even if the whole lecture were confined to this period of the engine is shown in Fig. 23, and also in Figs. 24 and 25, testing, it would still be no more than an outline. Plate 4. We began with a coil type vaporizer mounted in the exhaust The compressor remained the same as in the second model, pipe. The vapour was led to a cluster of vapour nozzles in each

Fig. 23. Assembly of Third Model of Experimental Engine except that the impeller was modified by machining away the cracked buttress ribs and by thinning off the blades. Once more we were being penny wise and pound foolish. Each of the ten combustion chambers was basically similar to the one combustion chamber of the second edition, but approximately 1/ 10 times the size. The counterflow arrangement was preserved for the following reasons:— (i) The outer casings were insulated against high temperatures by relatively cool air from the blower. (ii) Serious modifications to the rotor assembly were avoided. (iii) A number of expansion and assembly problems were avoided. An important feature of the combustion chamber arrangement was the provision of interconnecting ducts between both the air casings and the flame tubes. The main purpose of these was to make it possible to light up all combustion chambers by providing ignition in one only. It was also hoped at that time that they would act to some extent as pressure balancers. The turbine design remained, in principle, the same as in the second model, but of course the direction of flow was reversed. The exhaust assembly of the first model was brought back into use with (at the beginning of testing) a vaporizer mounted within it. (6) Testing. To describe the testing of the third model in detail would take a long time, and I do not propose to attempt it.

592

combustion chamber. Each combustion chamber was provided with a pilot jet of atomized spray type for the initial heating of the vaporizer. However, it was soon found that to obtain the necessary amount of heating surface resulted in a serious restriction in the exhaust, and we were forced to fit each combustion chamber with its own vaporizer. The effort directed to the combustion problem grew steadily in intensity until by August 1940 four combustion rigs were in operation at Ladywood. Power Jets had recruited more engineers, and had built up a small sheetmetal shop. They were aided by the loan of Dr. Hawthorne and two other young engineers from the Royal Aircraft Establishment. We paid many penalties in the early days for the primitive nature of the tests. It was very much a case of "more haste, less speed". The solution always seemed to be just around the corner, and I have to admit that we were very slow to realize just how big was the problem we were attacking. The problem was one in which reason did not help much; there were too many variables. It was a case for continued trial and re-trial. Our efforts fell under the following main heads:— (i) Tests of many different types of vaporizer, (ii) Modifications to the flame tubes. (iii) Modifications to the baffle system. (iv) Modifications to the position and direction of spray of the vapour nozzles. To give an idea of the rate at which changes were being made,

THE EARLY HISTORY OF THE WHITTLE JET PROPULSION GAS T U R B I N E even in the early part of the testing, it is sufficient if I tell you that in January 1939 ten types of vaporizer were tried in the combustion rig, also nine flame tube modifications (most of these modifications consisting of variations in the position and size of secondary air holes). These were by no means the only changes tried. A factor which dominated the testing for a very long time was that seemingly good results achieved on the combustion test rig could not be repeated in the engine. For some time we attributed this to the fact that the combustion test rigs ran at, substantially, atmospheric pressure, whereas in the engine the pressure was much higher, and so was the temperature of the air entering the combustion chamber. The need for a highpressure combustion test rig was ultimately met when the Ministry of Aircraft Production made arrangements by which we were able to do combustion experiments with an air supply from a large compressor used in the construction of the Dartford tunnel. This testing was done at Dartford. These experiments confirmed the results in the low-pressure rigs rather than in the engine, so that for some time we were still left with the problem of why the combustion in the engine was so much worse than it was in rig tests. We now have a fairly clear idea of the reasons for the difference. I summarize them as follows:— (1) We now know that the combustion problem is very largely an aerodynamic one, and despite the fact that on rig tests we endeavoured to simulate the ducting from the compressor to the combustion chamber, the way in which the air flowed into the combustion chamber in the engine was not reproduced on the rigs sufficiently closely. (2) Small differences as between a given combustion chamber in the engine and on the rigs were having a much greater effect on combustion behaviour than we realized. (3) Lack of accuracy in manufacture and distortion in service caused differences in the combustion behaviour of the different combustion chambers of the engine, also to a much greater degree than we realized. This factor was, of course, aggravated as lack of uniformity increased through distortion and repeated modification. The troubles we encountered from time to time were:— (1) Severe surging in the fuel system. (2) Severe out-of-balance in the fuel flow to each of the ten vaporizers. (3) Blockage of the vaporizers and vapour nozzles with carbon. (4) Local overheating and burning of the vaporizers. (5) Local overheating of the casings through bad temperature distribution. (6) Heavy deposits of carbon in the flame tube. The burning-out of the vaporizers and blocking of the vapour nozzles was very unsystematic. We would, for example, find practically all the vapour nozzles blocked in one combustion chamber, while in another none would be blocked. Some of the vaporizers had a life of about half an hour only; others would last much longer. About thirty-one different types of vaporizer were tested in the rigs, many of them with several subvariations, and nine of these types were tried in the engine. A typical arrangement used early in this period is shown in Fig. 26, whilst the last of the vaporizer types is illustrated in Fig. 27, Plate 5. This last was the most successful, and with it we were able to get the engine nearly up to its design speed, but it suffered from the troubles of the earlier types in lesser degree. A good deal of the trouble with the vaporizers was attributed to the "cracking" of the fuel, and we tried to overcome this by using different fuels. In this we received advice and assistance from the Asiatic Petroleum Company, who were very helpful in making special fuels available. It became clear from rig tests that a factor contributing to local overheating of the vaporizers and to bad temperature distribution of the gases at the discharge of the combustion chamber was the very poor distribution of the air flow into the combustion chamber. The bad temperature distribution was

429

serious from the turbine point of view, because almost invariably the hotter gases passed through the roots of the turbine blades. Many attempts were made to rectify this by fitting various kinds of baffles between the flame tube and the outer casing and at other places in the path of the air flow into the flame tube. Innumerable alterations were also made in the number and distribution of the secondary air holes in the wall of the flame tubes. The flame tube shown in Fig. 27, Plate 5, is typical of the type used throughout this phase of the testing. The system of

Fig. 26. Typical Combustion Chamber with Vaporizer swirl vanes through which the primary air entered the combustion chamber remained substantially unchanged. It consisted of two concentric rings of swirl vanes of opposite hand, the vapour nozzles being set to inject approximately along the plane of separation of the oppositely rotating streams of primary air. Many changes in the fuel lines were made. We tried different types of fuel pump, different types of relief valves, control valves, etc. Frequently we were in trouble with dirty fuel or through foreign matter getting into the fuel line during modifications. For a large proportion of the testing the fuel pump was

Fig. 28. The Shell Type Combustion Chamber adapted for Use in Engine not mounted on the engine, but was driven by a separate electric motor; but it was subsequently mounted on the engine without producing any abnormal features. Among the more important modifications to the fuel line were:— (1) The introduction of an air bottle to damp out pressure fluctuations and to provide initial "boost pressure" for starting. (This was later replaced by a Dowty accumulator, but the scheme was eventually abandoned by Power Jets.) (2) In order to try to provide equal flow of fuel to each of the ten combustion chambers, a flow-balancing device was introduced into the fuel line. Fortunately the need to continue our struggle with this system of combustion disappeared in the autumn of 1940. For a long

593

430 JAMES C L A Y T O N L E C T U R E time Mr. I. Lubbock, of the Asiatic Petroleum Company, had (4) We scalloped the tips of the impeller blades as shown in been helping us by advice and by obtaining special fuels, etc., Fig. 31, with the object of imparting more energy to the but at this stage in the development he did something much air near the walls of the casing, in order to neutralize more important ; he and members of his team produced a comskin friction effects. bustion chamber using atomized spray injection and tested it at All except the last of these modifications seemed to produce the Fulham Laboratory of the Asiatic Petroleum Company. The improvement, but the last one caused a substantial re"Shell" combustion chamber is shown in Fig. 28, and this was some the starting point of a new phase in the development. (We had duction in delivery pressure for a given rotational speed, as it made several attempts, on the combustion rigs, to get a satisfactory system using liquid injection in parallel with the development of the vaporizer system but with little success.) Power Jets now commenced intensive development both on the controllable atomizing burner and on the other features of the "Shell" combustion chamber. Of the several arrangements tried on the rigs and in the engine, the most successful (the one in use in the engine at the end of its life) is shown in Fig. 29, Plate 5. Though the first tests on the engine with this system were disappointing, it nevertheless ultimately proved to be the turning point. Once more many of the combustion troubles were subsequently found to be due to lack of accuracy, especially in the manufacture of burners. We continued to use kerosene as the fuel, though gas oil was tried. The main reason for persisting with kerosene was its low freezing point. Though the combustion problem was by far the greatest, and though much work still lay ahead, we were over the worst by the time the experimental engine gave up the ghost. At least it was possible to say that the combustion problem had ceased to be an obstacle to development. The atomized liquid spray system had outstanding advantages over vapour injection in respect of the simplicity of the fuel system and of the simplicity of the starting procedure. A disadvantage was that much higher fuel pressure had to be used. Needless to say, our troubles were not confined to combustion. Fig. 30. Guide Vanes fitted to Blower Intakes during In March 1939 one of the impeller blades cracked near the tip. Testing of Third Model of Experimental Engine In order to keep running, we cut off the tip of that blade and also the tip of the opposite blade, to maintain balance. We only constituted an effective reduction in diameter. Uncertainty as managed to get one more test run out of this impeller before to the exact values of the temperature rise made it difficult to many other blades failed in a similar manner, one blade tip say what the effect on efficiency was. passing right through the engine. Fortunately the damage to the A complete new rotor assembly was fitted in July 1940; but turbine blades was relatively slight. the impeller of this assembly was withdrawn from service after It was now considered that the combination of ten diffuser a few runs and replaced by the old one, as the new one was blades and thirty impeller blades was unfortunate; so the new required for the flight engine, the impeller for which had been impeller, which had been ordered from The British Thomson- scrapped in manufacture. Houston Company, was to have twenty-nine blades; also the We had to fit the new rotor assembly when a turbine blade central disk was not to be cut away between the blades at the tips. While waiting for the new impeller, the opportunity was taken to modify the diffusers, to try to improve the compressor performance, which was still much below expectations. This modification did in fact produce a marked improvement in blower performance, in spite of the fact that, due to faulty assembly, there was a substantial variation in the throat widths of the diffuser passages. A crack appeared at the tip of one of the blades of the new impeller where it joined the central disk, at a place where there had been damage in heat treatment; and because of this and a rub on the casing, a further new impeller was ordered. Nevertheless this impeller continued in use for all further testing with this engine (except for a short period) because the crack did not extend beyond11/2inches in length. We made a further attempt to improve the blower performance at the end of 1939 by fitting guide vanes in the blower Fig. 31. Modification to Tips of Impeller Blades in Third intake to help the air round the corner (see Fig. 30). This Model of Experimental Engine modification did not increase the delivery pressure but seemed to have improved the compressor efficiency as the measured compression temperature rise was reduced. New diffuser blades failed by tearing off about 1 inch from the tip. Until then, little were fitted to the blower in March 1940, in a further attempt had been done to the turbine in the course of testing (except to improve the blower performance. Many other modifications light overhauls after numerous incidents inflicting minor to the same end were made, briefly as follows:— damage). This turbine wheel had had a rough time in the course (1) We attempted to prevent warm air from being drawn of its history. Apart from the effects of bad combustion, there into the rear intake of the blower by fitting heat shields were many tip rubs, partly through distortion and partly through between the elbows of the "starfish" assembly and the yielding of the disk, and both nozzle blades and turbine blades had sustained progressive damage through impact with foreign rear intake of the blower. bodies, including pieces of impeller blade. The turbine had also (2) We modified the side clearances. (3) We attempted to stop leaks in the blower casing which survived two main bearing failures (each due to foundry sand were present as a result of progressive distortion (mainly in the lubricating oil: each time the engine ran there was a fine rain of sand from the roof of the building). caused by numerous rubs).

594

431 THE E A R L Y H I S T O R Y OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E Testing of the WlX. When the engine was delivered by The British Thomson-Houston Company, the combustion system was modified in accordance with the latest experience on the experimental engine. For test purposes the engine was now carried on a frame which was suspended from the roof of the test house by four swinging rods. At this time a spring balance was still used for measuring thrust, though it was later replaced by hydraulic thrustmeasuring devices. Testing was carried out in the new test houses which had been built at the Ladywood Works. We succeeded in reaching a speed of 16,500 r.p.m. within a day or two of the beginning of tests on the WlX, and though the performance was somewhat below expectations there was a very great improvement as compared with the experimental engine.

The replacement turbine was of the same design, except that the blades were made of "Rex 78" (Firth-Vickers) instead of "Stayblade". A new nozzle ring was fitted at the same time as the new rotor assembly. There was gradual yielding of the disk of the second wheel, of the same kind as with the first. This yielding occurred mainly at the shoulders of the bulb roots of the blades, which became progressively looser in their sockets as a result. For some time before the new wheel was used it had become obvious that the De Laval type of root fixing was unsatisfactory, and a wheel was ordered with the "fir tree" fixing, as used in the Wl, but it was never used in the experimental engine. The disk ultimately failed completely and wrecked the engine in February 1941 after a very useful running time of 169 hours. The failure is shown in Figs. 32 and 33, Plates 5 and 6. Further points of interest connected with the test running of the experimental engine are:— (1) Some useful experiments were made on starter motors. Successful starting was achieved with a small singlecylinder two-stroke petrol engine, and also with the electric starter motor of the 10 h.p. motor car engine which had been used for starting during most of the testing. (2) We found that we could eliminate the rear water jacket without deleterious effects; and as a result, the W1X and Wl engines were modified to dispense with the rear water jacket. (3) Towards the end of its life, several non-stop runs, lasting up to ten hours, were done at cruising speed (14,000 r.p.m.). P A R T I I I . THE DESIGN AND TESTING OF THE WlX AND Wl ENGINES

The leading particulars of the Wl engine are given in Table 2, and the general design is illustrated in Fig. 34, and in Figs. 35-38, Plates 6 and 7. As already mentioned, the WlX was built from components ordered for the experimental engine and from certain parts TABLE 2. LEADING PARTICULARS OF Wl Dimensions are in inches. Compressor impeller— Tip diameter . T i p width . . . Outer diameter of eye Inner diameter of eye Number of blades Material . . .

. . . . . .

.19 2 . 10.75 . 5.5 . 29 Hiduminium RR59

Compressor casing, Material

.

Turbine— Mean diameter of blades Blade length . . Material of blades . Material of disk . . Blade chord . . Number of blades .

. 14 .2-4 . Firth-Vickers Rex 78 . Firth-Vickers Stayblade . 0.8 . 72

Maximum spend, r.p.m.

.

.

Mg. Alloy No. 299 DTD.350

17,750

made for the W1 which were rejected as "unairworthy". The WlX differed from the Wl only in respect of the necessary adaptations. The general design of the Wl was very similar to the third model of the experimental engine, except that much lighter scantlings were used. Other important differences were as follows:— (1) A very different and much lighter design of auxiliary drive box was used. (2) The number of turbine blades was increased from sixtysix to seventy-two, and the fir-tree root fixing was used. (3) The blower casing carried four mounting points for attachment to the tubular engine bearer structure of the aircraft.

Fig. 34. Assembly of Wl Engine Combustion chamber details not shown. Bearing failures were our main trouble in the early testing of the WlX. These failures were not due to the inability of the bearings to carry the load imposed, but to faults in the installation, either because the design was unsatisfactory or because the assembly was faulty. One was due to malalignment of the bearing on the shaft; another to a failure in the lubrication system; another to foreign matter getting into the bearing housing; one at least was due to an oversight through sheer fatigue on the part of the people doing the work (we were working at very high pressure and for very long hours). The intensive testing of the WlX resulted in a series of modifications to the Wl (and the other engines then in course of design and manufacture). It was a most useful engine, and we did 132 hours of test running with it in nine months. After that it was sent to the General Electric Company of America, where it was used for demonstration purposes. The Testing of the Wl. There is little to be said about the testing of the Wl without going into excessive detail. There were early "teething troubles" of course, but serious troubles at speeds up to the rating at which it was decided to fly were avoided by virtue of our experience with the WlX; and though testing of the engine did not begin until 12th April 1941, we succeeded in getting through a 25-hour special category test to clear it for flight, installed it in the aircraft, and completed the ten hours of flight trials, all in 46 days. The engine was cleared for flight at a static thrust of 850 1b. at 16,500 r.p.m. Typical performance curves are given in Figs. 39 and 40. A hydraulic starter motor was designed for the Wl and WlX engines, but was never used. The starter motor for both bench and flight tests was a 7 h.p. motor car engine mounted on a hand truck and fitted with a long flexible drive which could be

595

432

JAMES CLAYTON LECTURE connected to and disconnected from the engine by hand. It was also intended that each of these two engines should be fitted with a governor to prevent overspeeding. Such a governor is necessary, because the fuel required for full speed diminishes with height; hence for a given setting of throttle—which itself consists of a needle valve—engine speed increases with height. However, a satisfactory governor was not ready in time for the flight trials and so we did without. All later engines were provided with a governor and a barostat relief valve (by means of which the maximum fuel-line pressure was made a function of the height—as height was increased, the fuel pressure was reduced). We received much valuable help from Dr. H. R. Ricardo, F.R.S., M.I.Mech.E. (Past-President), and his team in the design and development of these and other fuel-line accessories. The general arrangement of the aeroplane is shown in Fig. 41, Plate 7 , a n d i n Fig.

4

2

.

P

A

R

T

ENGINES

Fig. 39. The Wl Engine: Curves of Thrust, Specific Fuel Consumption, and Exhaust Temperatures plotted against Speed Test results. "Design" performance.

Fig. 40. The Wl Engine: Curves of Blower Delivery Pressure and Air Mass Flow Air flow readings are now known to be low. Test results. '-"Design" performance.

596

The WlA Engine. The WlA engine was designed for a thrust of 1,450 1b. The general design is illustrated in Fig. 43, also in Figs. 44 and 45, Plate 7. The general arrangement, as may be seen, has many features in common with the Wl, but air cooling was used for the turbine instead of water cooling, the fins of the air cooling fan being machined integral with the turbine disk. Another important difference was that stationary guide vanes were mounted at the entry to the blower intakes to give the air a "pre-whirl" in the direction of rotation. Initial tests with the WlA were very disappointing, primarily because of low turbine efficiency. Surging of the blower was experienced for the first time, but in the case of this engine it was cured by dispensing with the air cooling system at the rear side of the turbine wheel. (The cooling air to the rear side of the wheel discharged into the exhaust annulus and being relatively "stagnant" it caused "choking" of the exhaust. This lowered the capacity of the expansion system below the surge point of the blower.) A substantial improvement in turbine efficiency was achieved by fitting a redesigned set of blades. The unsuccessful blades had profiles based on aerofoil sections. With the new ones we returned to more conventional profiles, though still conforming to angles appropriate to vortex flow. Eventually, though this engine fell short of its design performance, it was relatively successful and it powered the E28 aeroplane in flight trials. The W2 Engine. An illustration of this engine is shown in Fig. 46, Plate 8. As I have already said, this engine was a failure. It was too ambitious. Of the many serious faults in the design, perhaps the worst was that we had designed for an exit velocity from the turbine much too near the speed of sound, so that when the assumed component efficiencies were not realized the exhaust velocity reached the critical value at speeds well below the design full speed. Both the compressor and turbine efficiencies were substantially lower than the expected figures, the net result being that surging of the blower and high exhaust temperatures made it impossible to run at more than 75 per cent of the design r.p.m. The corner vane system at the blower casing discharge ports was a much-criticized feature of this design, but this same feature has been embodied in later designs with Considerable success. We had realized before we started to test the W2 that the design was very critical in respect of design assumptions, and therefore no serious attempt was made by Power Jets to develop it. We relied on the revised design known as the W2B to give us what we wanted. The general design of this engine is illustrated in Fig. 47 and in Fig. 48, Plate 8. A great deal of development work was required before the W2B gave its design performance, but that is a story which must be told later. As already mentioned, this was the parent design of the Rolis-Royce Welland engines which powered the Meteor fighter, and of the American General Electric Company's engines which powered the Bell P.59 A (or Airacomet).

I

THE EARLY HISTORY OF THE WHITTLE JET P R O P U L S I O N GAS T U R B I N E

Fig. 42. Fuselage Arrangement of the Experimental Aeroplane E28/39 CONCLUSIONS

If the foregoing account has caused mental indigestion, my excuse is that it could scarcely do otherwise when the intensive work of about six years has to be compressed into so short a

Company, and the other firms engaged. I have not been able to mention many individuals and firms who have given great help. Acknowledgement will, I hope, be made in due course. I certainly hope I have not conveyed the impression that this has been a one-man job. No big engineering task ever is a oneman job. As every engineer knows, it is always the work of a team. At first the team working in this field of engineering consisted of The British Thomson-Houston engineers, Messrs. Laidlaw, Drew and Company, and myself. Later it was aug-

Fig. 43. Assembly of the W1A Engine

Fig. 47. Assembly of the W2B Engine

description. It is far from complete in many ways. I have been able only to hint at the complex relationships which existed trom time to time between Power Jets, the Air Ministry (or Ministry of Aircraft Production), The British Thomson-Houston

mented by the excellent team that I helped to build up at the Power Jets' Works; and later still, as other firms began to share in the work, it would be true to say that the fruits were the product of a group of teams working in collaboration.

597

JAMES CLAYTON LECTURE I feel that I must point out that a factor which has contributed greatly to the work has been the very fine training I have received in the Royal Air Force. Of my211/2years' service, I have spent ten under training, mostly engineering. In this I regard myself as having been very fortunate, and I hope that the outcome will do something to suggest that such a degree of training should not be as exceptional in the future as it is at present. 434

A NOTE ON GERMAN WORK

Evidence from Germany recently obtained indicates that the Germans started practical work on jet propulsion gas turbines at about the same time as ourselves, i.e. 1936, but that they succeeded in flying with this system of propulsion before we did (believed to be a few days before the outbreak of war). However, the engine they used was of a type subsequently abandoned, and successful flight trials with the type of engine which they ultimately used in quantity did not take place until some time after the British flight trials in May 1941. The engine with which our first flight trials were carried out was the obvious parent of a successful series of British and American engines and so I think we can claim with justice that the E28/39 was the first successful aeroplane using the gas turbine jet propulsion engine. The power plant of the Me.262—an aeroplane which the Germans used in considerable numbers towards the end of the war in Europe—was the Junkers 004 jet propulsion gas turbine having an axial flow compressor. They put this engine into production at a stage in its development history which would be considered far too early in this country, and it was technically a long way behind the British engines, being much heavier in proportion to its power, having a much bigger fuel consumption, and only a fraction of the life. Moreover it had the serious disadvantage of being very sluggish in response to the throttle. This, combined with the low reliability of the engine and the high wing loading of the aeroplane, represented a very dangerous combination, and it is believed that fatal accidents were so numerous as to affect the morale of the pilots very seriously. A NOTE ON THE FUTURE

So far I have confined myself to the past, but a brief word about the future may not be out of place. Jet propulsion is only one way of using the gas turbine in aircraft propulsion (incidentally, the gas turbine method is only one of about five systems of jet propulsion which have been developed to, or nearly to, the practical stage during the war). Up to now far too much emphasis has been placed on the jet propulsion aspect of this development. This particular mode of using the gas turbine has come first because it happens to be the simplest, but there is no doubt that the gas turbine will also be used in combination with a normal propeller (Fig. 59), and also with what has come to be called the "ducted fan", i.e. a low-lift axial flow compressor in a duct. The use of the jet propulsion gas turbine will probably be confined mostly to aircraft designed for high speeds and relatively short ranges. The other applications of the gas turbine will cover speed and range requirements unsuited to the plain jet. So wide is the possible range of application of the gas turbine in one form or another that it seems reasonably certain that within (say) 10-15 years the piston engine will have given way to the gas turbine in all aircraft except possibly the light aeroplane. I make the reservation about the light aeroplane because at present it appears to be far more difficult to design for the low power necessary than it is to design for much higher powers than we use at present. The uses of the gas turbine will not of course be confined to aircraft. It has already been successfully used in locomotives in Switzerland, and in other ways. I expect to see it reach some prominence in the field of marine propulsion. Acknowledgeinents. I am indebted to the British ThomsonHouston Company, Rugby, for permission to reproduce the photographs used in Figs. 7, 8, 9,11,13,14, 21,22 and 35, and to the Asiatic Petroleum Company for Fig. 28.

598

THE E A R L Y H I S T O R Y OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E 435 ADDENDUM efficiency is relatively poor, though at 600 m.p.h. the efficiency Since the foregoing lecture went to press, details of several of the propeller is dropping to values little more than this. more recent types of turbo-jet engines and aircraft have been The efficiency of the engine as a heat engine for generating officially released for publication. Several of these were described kinetic energy in a jet is about 34 per cent at heights where the in a subsequent lecture given before the Royal Institution, temperature is very low, so that the absolute overall efficiency London, on 7th December 1945. Figs. 49-53, Plates 9 and 10, is of the order of 17 per cent. show views of the Power Jets W2/700 engine, the Rolls-Royce Some more recent types of turbo-jet engined aircraft are Derzvent I, the Metropolitan-Vickers F2, and the de Havilland illustrated in Figs. 56-58, Plates 11 and 12, which show the de Goblin; while Figs. 54 and 55, Plate 11, illustrate typical rotors Havilland Vampire, and the Gloster Meteor III. The former has from centrifugal type (Power Jets W2/700) and axial type its two air intakes in the leading edges of the wings. The Meteor (Metropolitan-Vickers F2) turbo-jet engines respectively. III is fitted with Rolls-Royce Derwent I engines. The Meteor IV, In these turbo-jet engines the jet velocities are of the order which at present holds the world speed record of 606 m.p.h., of 1,600 to over 2,000 ft. per sec., i.e. of the order of three has longer nacelles and is powered with two Rolls-Royce times the forward speed of the aeroplane. The Froude efficiency Dement V engines. Another well-known single-engined intercorresponding to a 3/1 ratio is 50 per cent, so that the propulsive ceptor is the Lockheed Shooting Star.

599

THE E A R L Y H I S T O R Y OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E

Fig. 7. First Model of the First Experimental Engine

Fig. 8. Test Assembly of First Model of Experimental Engine The truck and starter motor have been "blanked out" of the photograph.

Fig. 9. Rotor Assembly of First Experimental Engine [I.Mech.E., 1946]

600

Plate 1

Plate 2

JAMES CLAYTON LECTURE

Fig. 11. The Unconventional Bladeless Turbine Nozzle Scroll in the First Model of the Experimental Engine

Fig. 13. The "Honeycomb" Diffuser used in the First Model of the Experimental Engine

Fig. 14. Combustion Test Arrangements outside the British Thomson-Houston Factory, as used before the Completion of the Experimental Engine [I.Mech.E., 1946]

601

THE EARLY HISTORY OF THE WHITTLE JET P R O P U L S I O N GAS T U R B I N E

Plate 3

Fig. 19. Second Model of Experimental Engine : Wooden Model

Fig. 21. Modification to Blower Casing, Second Model of Experimental Engine [I.Mech.E., 1946]

602

Fig. 22. Rebladed Turbine (Blades Designed for Vortex Flow) for Second Model of Experimental Engine

Plate 4

JAMES CLAYTON LECTURE

Fig. 24. Test Installation of Third Model of Experimental Engine at Ladywood Works, Lutterworth A. 10 h.p. engine for starting. B. Radiator for starting motor.

Fig. 25. Another View of Test Installation of Third Model of Experimental Engine at Ladywood Works, Lutterworth [I.Mech.E., 1946]

603

THE E A R L Y HISTORY OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E

Fig. 27. Vaporizer Type 31, with Flame Tube

Fig. 32. Failure of Turbine Disk of Third Model of Experimental Engine (February 1941)

Fig. 29. Combustion Chamber Type 75 Type in use at end of testing. [I.Mech.E., 1946]

604

Plate 5

Fig. 33. Yielding of Turbine Disk of Third Model of Experimental Engine (February 1941) Fig. 35. View, from Gearcase End, of the Wl, the First Flight Engine A. Fuel pump. B. Tachometer generator. C. Oil starter motor. D. Lubricating oil pump. E. Water pump. G. Oil tank.

Fig. 36. View, from Exhaust End, of the Wl, the First Flight Engine

Fig. 37. Rotor Assembly of the Wl Engine

Fig. 38. "Fir Tree" Blade Root Fixing on the Wl Engine

Fig. 44. View, from Gearcase End, of the W1A Engine

Fig. 41. The Gloster-Whittle Experimental Aeroplane E28/39 (R.A.F. Official, Crown Copyright Reserved)

Fig. 45. View, from Exhaust End, of the W1A Engine

Plate

8

JAMES CLAYTON LECTURE

Fig. 46. View, from Exhaust End, of the W2 Engine

Fig. 48. View, from Gearcase End, of the W2B Engine

607

THE E A R L Y HISTORY OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E

Fig. 49. Power Jets W2/700 Engine

Fig. 50. Rolls-Royce Derwent I Engine

Fig. 51. De Havilland Goblin Engine

608

Plate 9

Plate 10

JAMES CLAYTON LECTURE

Fig. 52. Rolls-Royce Derwent V Engine

Fig. 53. Metropolitan-Vickers F2 Engine

609

THE E A R L Y H I S T O R Y OF THE W H I T T L E JET P R O P U L S I O N GAS T U R B I N E

Fig. 54. Rotor of the Power Jets W2/700 Engine

Fig. 55. Rotor of the Metropolitan-Vickers F2 Engine

[I.Mech.E., 1946]

610

Fig. 56. De Havilland Vampire Jet-Propelled Interceptor Fighter with Goblin Engine

Plate 11

Plate 12

JAMES CLAYTON LECTURE

Fig. 57. R.A.F. Gloster Meteor Jet-Propelled Single-Seat Fighter (M.A.P. Crown copyright reserved.) Power unit: two Welland or Derwent engines; armour, four 20 mm. cannon in the nose. Span, 43 feet; length, 41 feet; height, 13 feet; wing area, 374 sq. ft.

[I.Mech.E., 1946]

Fig. 58. The R.A.F. Gloster Meteor Fighter in Flight (M.A.P. Crown copyright reserved.)

611

This page intentionally left blank

Harry Ralph Ricardo FRS (1885-1974) "One of the brainiest men I have met" and the "one of the soundest...on the theory of the petrol engine" were two contemporary opinions of Ricardo. He was already a skilled mechanic before graduating from Trinity College, Cambridge. There, he conducted research into the internal combustion engine under the direction of Professor Bertram Hopkinson. His early life and career were the subject of Memories and machines (1968), one of the most readable of engineering autobiographies. Ricardo began to make engines in a small way to power fishing boats in Shoreham-by-Sea. Subsequently, he acted a consulting engineer to many engineering companies, the two most notable being Rendel, Palmer, and Tritton and Armstrong, Whitworth, and Company. The Great War proved a formative experience. Ricardo was inspecting Hugo Junkers' (1859-1935) laboratories in Aachen only days before hostilities commenced. On his return to Britain, the engineer immediately began work on aeroplane engine development. He was asked to assist the Mechanical Warfare Department. By 1916, he was working on tanks, designing 150 hp engines fitted to the British Army's Mark IV and Mark V vehicles. Ricardo formed his own company in 1917, concerned initially with fuels research. This involved such eminent contemporaries as Henry Tizard (1885-1959) and David Pye (1886-1960). This collaboration resulted in Ricardo advising Allcock and Brown on the fuel mixture for their important and pioneering transatlantic crossing. The company came to be consulted by many of the great engineering concerns of the day. RollsRoyce, Napier, Vauxhall and others were all clients. Ricardo's researches brought him into contact with most types of powerplant and his book, The internal combustion engine (1922), was an admirable summary. The second volume, on high-speed engines, was revised through many editions. His designs incorporated many technical improvements and Ricardo contributed greatly to the sleeve-valve aeroengine. He also made elements of the fuel systems of the first jet engine for Frank Whittle.

613

193

AEROPLANE PROPELLING MACHINERY, 10TH JUNE Chairman: Sir Ben Lockspeiser, B.A., M.I.Mech.E.

Piston Aero-Engines By Harry R. Ricardo, B.A., LL.D., M.I.Mech.E., F.R.S. (Past-President)* It has fallen to my lot to speak of the development of the piston aero-engine during recent years and its possible development in the future. So far as military aviation is concerned, mine is likely to be the swan song of the piston engine—for the turbine, whether its energy be used to drive an airscrew or as a simple jet, will probably supersede it for most military duties within the next few years. It must be borne in mind that all research and development work on aero-engines during recent years has been carried out either during war itself or under the shadow of gathering war clouds, with the result that military requirements have always dominated. Military requirements call for the utmost possible power output from the minimum weight or bulk of engine while such considerations as fuel economy or even safety, both so vital in civil aviation, must take but a second place. To the Service pilot, safety depends first and foremost on his ability to out-manoeuvre his enemy, and to this everything else is subordinate. Fuel economy, of course, becomes important where very long range is required, but for the most part it is but a secondary consideration. I have said that the goal to be aimed at is minimum weight or bulk of engine, and it is here that we in this country adopted a policy differing from that of all the other belligerent nations: for we, having always in view the need for very high speeds, concentrated primarily on small bulk, or more particularly on small frontal area. To this end we devoted our energies to the development of small but very high-pressure high-speed engines, while all the other countries concentrated on the relatively large, lowpressure, low-speed type. We gained but little thereby in specific weight, for the smaller engines must be proportionately stiffer to withstand the higher gas pressures and dynamic loadings, but we gained greatly in reduced frontal area. In so doing, we chose the more difficult path, but our policy was, I think, vindicated amply during the Battle of Britain. Broadly speaking, the cubic capacity of our engines was only about 70 per cent of that of others of the same type and power, and the frontal area even less. This we achieved by the employment of much higher mean effective pressures of the order of 350 to 400 1b. per sq. in. in the case of operational types of liquid-cooled engines, and 250 to 280 1b. per sq. in. in that of air-cooled engines, and by employing piston speeds of the order of 3,000 ft. per min. and over. The famous Rolls-Royce Merlin engine (Figs. 1 and 3, Plate 1, and Fig. 2) which powered most of our fighters and many of the American and Russian aircraft also, was perhaps the most outstanding product of this school of thought. In its latest form, this engine of only 27 litres cylinder capacity has passed a special type test with a combat power of no less than 2,340 b.h.p., corresponding to an indicated mean pressure of 475 1b. per sq. in. and an "all out" maximum power of 2,620 b.h.p., or very nearly 100 h.p. per litre of cylinder capacity. It would be hard indeed to point to any one feature or discovery which, by itself, has revolutionized the performance of the aero-engine. Progressive improvements in fuel, with a view to reducing the tendency to detonate, have done more than anything else to render possible the high performance of the modern engine. Since the realization, more than thirty years ago, that the incidence of detonation set a limit, and in those days an early * Consulting Engineer.

limit, to the power output obtainable, research on fuels has been carried on intensively, and the engineer has at every step taken full advantage of it, at first to raise his compression ratio and thereby gain in thermal efficiency and, when that had reached the practicable limit, to increase further his mean effective pressure by supercharging. An increase from 66 to 100 in octane number permits of almost a threefold increase in mean effective pressure but at the cost of more than doubling both the maximum gas pressures and the intensity of heat flow. Throughout the

00 Fig. 2. Installation of Merlin Engine in Mosquito (a) and Hornet (b) Aircraft Showing how drag was further reduced in later developments. last twenty years it has been a neck and neck race between the chemist and the engineer; at times the chemist has been ahead and the engineer at frantic pains to stiffen the structure and working parts and improve the cooling of his engine in order to take full advantage of the improved fuel, at others he has taken the lead. It proved a wise recommendation of some fifteen years ago, following research work on the compression-ignition engine, that development tests on aero-engines should be carried out on specially prepared fuels many octane numbers ahead of what was available in contemporary service. From a mechanical point of view, progress has taken the form of strengthening step by step each weak link as it gave under the ever-increasing strain. At one time exhaust valves were the limiting factor, but the introduction of sodium cooling and the use of stellite or other similar materials for facing the valves banished this limit; then came bearings, when the intensity of loading exceeded the capacity of ordinary anti-friction linings and special materials

615

194

CENTENARY LECTURES ON AEROPLANE PROPELLING MACHINERY

such as copper-lead or cadmium-nickel involving new technique had to be substituted. Throughout, the piston itself has always been a weak point; here detail design in the way of better disposal of material—with a view both to stress distribution and heat dissipation—and the use of oil cooling have done much to improve conditions, but, so far as the piston is concerned, the greatest gain of all has, I think, been the discovery by Napiers some fifteen years ..go of the value of the wedge-shaped piston ring, which has raised the limiting temperature set by ring sticking, followed by blow-by and eventual seizure, by some 50 deg. C. (90 deg. F.). The above are but a few of the thousand and one steps that have been taken to improve the performance of any given engine. It is probably safe to say that within the past ten years almost every single stressed part of every aero-engine in service has been redesigned not once but several times, and this, more especially in war time, has had to be done by a process of infiltration in order to interfere as little as possible with maintenance or the flow of production. Air versus Liquid Cooling. From the day when the first aeroplane left the ground, controversy has raged as to the relative merits of air and liquid cooling. The exponents of the former have asked, logically enough, What is the sense of introducing an intermediary, with its increased vulnerability and all its associated plumbing troubles, since the waste heat must be removed by air in any case? On the other side, it can be argued that since the intensity of heat flow in an internal combustion engine cylinder is highly localized, only the latent heat of evaporation of a liquid can cope with it satisfactorily; that if sufficient energy is devoted to cooling the local hot spots by air blast alone then either the cooling drag will become excessive or the performance of the engine must be derated considerably, as compared with that of a liquid-cooled engine. To-day it has generally been accepted that, where very high speeds are called for, the liquid-cooled engine with its much higher performance rating and lower cooling-drag is preferable despite the objections of vulnerability, plumbing, and freezing, but that for moderate speed machines, where a relatively large frontal area and coolingdrag can be tolerated, air-cooling is to be preferred. As to overall specific weight there is but little to choose; on the whole the specific weight of the liquid-cooled engine, together with its radiator and cooling liquid is, if anything, the lower. Sleeve Valves. Unlike any of the other belligerents we, in this country, have adopted the sleeve valve for both air- and liquid-cooled engines, the former as exemplified by the range of Bristol engines Hercules, Centaurus, and others, and the latter by the Napier Sabre, the Rolls-Royce X and the new RollsRoyce Eagle engines; in fact, the Rolls-Royce Merlin and Griffon represent about the only surviving poppet valve engines in military service in this country to-day. Some twenty-five years ago, when active research on fuels and combustion chamber forms was in progress, it was realized that the sleeve valve—with its freedom from hot exhaust valves, its relative immunity from the effects of lead poisoning, and its compact form of combustion chamber— would permit of a considerably higher output within the limits set by the incidence of detonation. It would also have the important advantage that, with less top-hamper above the piston, frontal area could be reduced substantially. A thorough analysis of the various possible forms of sleeve valve was made and the single sleeve with a combined reciprocating and rotary motion, as patented some forty years ago by Burt and McCollum, was considered the most promising. Preliminary tests showed that, on the same octane fuel, a sleeve valve engine could operate with one ratio higher compression than an otherwise similar poppet valve engine, or alternatively could cope with a correspondingly higher supercharge, and it could digest a higher proportion of lead. In those days the octane number of fuels was very low, and lead almost the only means of raising it appreciably, so that the advantage gained was very material. Further, the sleeve valve was shown to have a number of secondary advantages—but time will not permit me to discuss these. The first complete sleeve valve aero-engine to be put into service was the Bristol Perseus—a single-row nine-cylinder aircooled radial. After the usual initial teething troubles, this gave a very good performance, and it was succeeded by a range of double-row radial engines: the Taurus, Hercules,

616

and Centaurus (Fig. 4, Plate 1). Personally I had felt always that the sleeve valve would show to better advantage in a liquidcooled engine owing to the difficulty, with air cooling, of scooping out the heat from the deeply re-entrant cylinder heads. The Bristol Company, however, went a long way towards solving this difficult problem by the development of a two-part head using copper conducting fins (Fig. 5, Plate 1). Next in the field were the Napier Company, with the Sabre, a 24-cylinder liquid-cooled engine designed by Major Halford to take full advantage of the sleeve valve in the direction of extreme compactness (Fig. 6). This engine powered the Typhoon, Tempest, and Fury aircraft, the latter having a speed of well over 480 m.p.h. at 20,000 feet. It gives a take-off and combat power of no less than 3,050 b.h.p. when running at 3,850 r.p.m. and has shown itself capable of a sustained output of 3,600 h.p.; thus its output per litre of cylinder capacity is about equal to that of the much more highly developed RollsRoyce Merlin. Lastly, I should mention the new 3,500 h.p. Rolls-Royce Eagle, an engine of generally similar design but of somewhat larger dimensions than the Sabre. Geometrical Arrangement of Cylinders. For aero-engines of large power two geometrical arrangements had become almost standardized throughout the world for the last twenty years, namely the radial cylinder lay-out for air-cooled engines, and the liquid-cooled 12-cyUnder for V-engines. At first glance it would appear that the former with its very short crankshaft and compact crankchamber should afford the lightest possible construction ; in practice it has always proved disappointing—it remains the lightest in terms of 1b. per unit of cylinder volume but not per h.p., since the limitations imposed, on the one hand by aircooling and on the other by the concentrated loading on the crankpins, have compelled such engines to operate at lower pressures. The British policy of concentrating on high pressures and high speeds favoured the use of large numbers of small cylinders and resulted in the development of the so-called H type engine— in effect two superimposed 12-cylinder horizontal engines sharing a common crankcase. This arrangement, when combined with the use of sleeve valves, provides a remarkably compact and rigid form with a very small frontal area, e.g. the Napier Sabre and Rolls-Royce Eagle engines (Figs. 7 and 8, Plate 2). Compression-Ignition Engines. Very shortly after the 1914-18 war, the Air Ministry instigated a research into the possibilities for aircraft propulsion of the compression-ignition engine using heavy oil, and by 1921 the Royal Aircraft Establishment had converted their largest single-cylinder aero-engine unit and succeeded in reaching a power output at a piston speed of 2,400 ft. per min. closely comparable with that of contemporary aeroengines, and that with a fuel consumption as low as or lower than that of large stationary or marine Diesel engines. This really remarkable achievement never received the appreciation or publicity it deserved, for it demonstrated, for the first time, that a Diesel engine could both be built with light scantlings, and operate very efficiently at a piston speed far in advance of anything that had been achieved hitherto. It served, however, to inspire and encourage others to pursue the same aim and by about 1928 research and development work had reached the stage when several full-sized experimental aeroengines, both air-cooled radial and liquid-cooled V-design, were designed and built (Fig. 9, Plate 2). At that time the performance of contemporary petrol engines was limited severely by the low octane number of the available fuel, and the heavy-oil, compression-ignition engines looked like being closely competitive even on the basis of specific weight, and were, of course, immeasurably superior on the score of fuel economy, but by the time these engines were completed and had been nursed through their initial teething troubles, the octane number of petrol had so much improved, and with it the performance of the petrol engine, that much of the advantage had disappeared. There followed, for the next five years, a neck and neck race, with improvements in the performance of the compression-ignition engine just about keeping pace with those of the petrol engine and its fuel. During the later 1930's, however, the rate of increase, actual

PISTON AERO-ENGINES

195

Fig. 6. Cross-section of Napier Sabre Engine

617

196

CENTENARY LECTURES ON AEROPLANE P R O P E L L I N G M A C H I N E R Y

or potential, of the octane number of petrol steepened, with the result that the performance of the petrol engine gained a lead such as the compression-ignition engine could not reasonably hope to catch up with, for it must be remembered that, unlike the petrol engine, the compression-ignition engine, with its immunity from detonation, had little to gain from any improvement in its fuel. With this situation in view, and with the war clouds gathering ever closer, further development of the compression-ignition engine for aircraft was either abandoned or relegated to a low priority. The same sequence occurred in other countries, notably in America, where the Packard Company had developed, and gone into small-scale production of, a very neat air-cooled radial; and in Germany, where the Junkers Company had devoted a great deal of energy to the development and relatively large-scale production of an opposed-piston, two-cycle version. Fuel Injection. The early work carried out on the compression-ignition engine, though abortive so far as aircraft was concerned, proved extremely valuable in other fields. The higher gas pressures involved served to show up and focus attention upon weak links in the mechanical design of petrol engines, and these weaknesses were corrected by the time improved fuels subjected them to as high and even higher pressures. Again the technique of fuel injection, upon which the success of the compression-ignition engine depends, was evolved and developed during the early phases of the development of the high-speed compression-ignition engine and was applied by the Germans to their petrol-engined aircraft. Here again, controversy has raged as to the relative merits of fuel injection, metered individually to each cylinder, and the alternative of carburation. Fuel injection confers certain advantages, e.g. a more accurate distribution of the fuel as between the several cylinders and, almost equally important, an accurate apportionment of lead, thus ensuring a lower overall fuel consumption and less liability to excessive lead deposits in one or more cylinders. It has the further advantage that since air only is supplied by the supercharger, it becomes possible by wide valve overlap to scavenge the cylinders and thus obtain an increase in power output of the order of 10 per cent. Against these advantages must be offset the fact that the admission to, and evaporation of, fuel in the supercharger improves greatly its performance—a very important consideration when as in British practice very high ratios of supercharge are employed. Again, with an externally carburetted mixture, it becomes possible to resort, when desired, to the use of very rich fuel/air ratios and thereby gain a substantial increase in power for take-off or combat use, due in part to the additional internal cooling, and in part to the fact that all modern high octane fuels show to best advantage as regards detonation when used very rich. It would seem, therefore, that on balance, the externally carburetted system is preferable for military usage where the highest possible power output is required, even though it is at the cost of an increase in fuel consumption, but that for civil aviation the balance of advantage lies with fuel injection. Exhaust versus Mechanically Driven Superchargers. The British practice has throughout been to employ mechanicallydriven superchargers in preference to exhaust-driven. In view of the very large amount of potential energy in the exhaust gases it would, at first sight, appear logical to make use of this to drive the supercharger, but to this, as applied to four-cycle petrol engines, there are a number of serious objections. In the first place the temperature of the exhaust from a highly boosted petrol engine is too high for the turbine's digestion, and some means must be found for dissipating heat between the exhaust valve and the turbine entry. It should be emphasized that this, probably the overriding consideration, applies only to four-cycle petrol engines; in the case of either two-cycle petrol engines or compression-ignition engines of any type, there is sufficient dilution air available at all times to bring the exhaust temperature down to a level acceptable to the turbine. In the second place, the use of an exhaust-driven turbo-blower introduces some tiresome control problems, as does also the time-lag in accelerating the turbine. Thirdly, the four-cycle unlike the two-cycle engine objects to exhaust back pressure. If it be asked, Why throw away exhaust energy which might

618

be put to useful purpose ? It may be replied that it is not thrown away, but is usefully employed to produce jet thrust. In a highspeed machine the additional thrust obtained from rearwardfacing exhaust outlets provides nearly, if not quite, as useful an employment for the exhaust energy as does the turbo-blower. Temporary Augmentation of Power. In military usage more especially, it is very desirable to be able, for short periods, to augment the power of the engine either for take-off or for combat. At take-off and at relatively low altitudes the supercharger can always provide the engine with more oxygen than it can safely consume within the limits set either by detonation or by thermal considerations, or both. Under such conditions a temporary increase of power can be obtained by the injection of water or of a water-methanol mixture. In this case the high latent heat of the injected liquid serves to provide internal cooling both to the supercharger and to the engine cylinders, while the steam produced serves as a very effective anti-detonant. By such simple means it is possible to augment the power by about 20 per cent without increasing either the heat stresses or the maximum peak pressures. Although this method has been well known and used in other fields for some twenty-five years, there has always been a curious opposition to its use for aircraft, and it was only the exigencies of actual warfare that induced the service authorities to accept it. At high altitudes where, even with the supercharger all out, the engine is still starved for lack of oxygen, temporary power augmentation can be achieved only by supplying additional oxygen in some form or other. In the first attempts liquid oxygen was injected into the eye of the supercharger : this achieved the desired result and was used in operational service but was open to the objections that the use of liquid oxygen involved difficult supply problems; that owing to the increased flame temperature and greatly increased tendency to detonate, it could safely be used only at altitudes well above the rated altitude; and, lastly, that freezing difficulties tended to introduce a serious time-lag in its introduction. Later Sir Ralph Merton proposed, as an alternative, the use of nitrous oxide which could be stored and carried as a liquid, in light cylinders, at normal temperature, and under quite a moderate pressure. This had very great advantages, not the least of which was that to the surprise of everyone it proved to be a very effective anti-knock or to be more exact, it permitted of a large increase in power without any increase in detonation. By the use of nitrous oxide at high altitudes, it was found possible to augment the power by as much as 40 to 50 per cent at a consumption of approximately 4 1b. of nitrous oxide per additional hundred h.p. per min. Since the time during which such power augmentation was required was generally only a matter of seconds, i.e. in order to close with or break-away from the enemy, this was not a serious objection. Civil Aviation Engines. Thus far I have been considering only engines for military aircraft, to the development of which the bulk of all our research has been directed. The needs of civilian aircraft are somewhat different in that safety, economy, and reliability become much more important than ultimate speed or rate of climb. I am not, however, of those who consider that to meet the needs of civil aviation we require an engine differing radically from the military type. I think we should modify the already highly-developed military designs to fit the needs of civil aviation, and there is no need for the modifications to be very extensive. Safety and reliability are, I think, best achieved by reducing slightly the rating of well-proved engines, while improved fuel economy can be obtained by lowering the supercharger and increasing the cylinder compression-ratios, combined with the use of fuel injection to individual cylinders. I can see no case for developing a new and exclusively civilian engine for large aircraft, where military engines of appropriate power are available. All military engines, however, are to-day of high power, whereas for a large proportion of civil aircraft a very much smaller engine will suffice, and it is in this field, I think, that new developments may find a place. I have said earlier that the compression-ignition engine could not hope to compete for military usage with modern high octane petrol engines, but I am not so sure that this applies in the case of

197 PISTON AERO-ENGINES the smaller classes of civil aircraft. Though it may be some- structure of the engine, an arrangement lending itself also for what heavier, it has the advantage of a lower consumption of driving contra-rotating air screws. a cheaper fuel, very greatly reduced fire risk, and freedom Another possible line of development for long range civil or from electric ignition with all its maintenance troubles. More military aircraft is that of the compound engine, using a very especially, I think, does this apply to the quite small privately- small, high-pressure, valveless, two-stroke piston engine of the owned plane: to the owner of such a plane pay-load, as such, simplest form as the high-pressure element supplying a lowmeans relatively little; safety, and in particular safety from risk pressure turbine. In such a system the piston engine becomes of fire, means a great deal; while the constantly recurring fuel in effect the combustion chamber of the turbine, but with this bill, though it may be a relatively small item, is none the less a difference: that the maximum temperatures of the cycle can be very insistent one. Unlike the military user, he is not interested used to advantage by converting the high-temperature heat into in performance at very high altitudes, nor in very high speed. useful work rather than by quenching it down by the addition For such a user the compression-ignition engine, either in the of a very large proportion of dilution air (all of which must be two- or four-cycle form, may well prove very attractive. The handled by a compressor whose efficiency is less than 100 per chief practical objection to such an engine will be the severe cent) as must be done in the case of the straight turbine. By such torque-recoil due to the high compression, and it may well be means it should be possible to attain an overall efficiency subthat the solution will be found in the use of a reverse-rotation stantially higher than any in sight by either the turbine or the engine in which the torque reaction is cancelled out within the piston engine alone.

619

PISTON AERO-ENGINES

Fig. 1. Front View of Rolls-Royce Merlin Engine

Fig. 4. Front View of Bristol Hercules 14-cylindcr Engine

Fig. 3. Three-quarter View of Rolls-Royce Merlin Engine

[I.Mech.E., 1947]

620

Plate 1

Fig. 5. Composite Cylinder Head of Bristol Hercules Engine

Fig. 8. Three-quarter View of Rolls-Royce Eagle Engine

Fig. 7. Front View of Napier Sabre Engine

Fig. 9. Experimental Diesel Aero-engine designed in 1929

This page intentionally left blank

John Douglas Cockcroft FRS (1897-1967) Descended from several generations of millwrights, J D Cockcroffs first glimpse of power engineering was a very traditional one, the water wheels and steam engines of the textile industry. Educated in Manchester, he served as an artillery signaller during the Great War. In peacetime, he progressed to Cambridge, where he became a professional researcher under the direction of Ernest Rutherford. The search for ever-more precise and exotic equipment for Cavendish Laboratory experiments brought out Cockcroft's inventive flair. He collaborated with several scientists to develop their experimental machinery. With E T S Walton, he developed a particle accelerator for firing protons at lithium, boron, and other targets. Cockcroft and Walton studied these transmuted elements and the energies released. The extension of this work earned Cockcroft and his partner the Nobel Prize for Physics in 1951. In the approach to wartime, Cockcroft's talents were applied to the establishment of Britain's radar defences. By 1940 he was part of Henry Tizard's mission to the United States dedicated to the exchange of technical data. His own contribution to this rapport came with his realisation that an atomic weapon should be made. In 1944, therefore, Cockcroft travelled to Canada to direct the work of the NRX reactor at Chalk River, which produced heavy water. This experience, and his earlier work with Rutherford, meant that he was perfectly placed to take a leading role in the peacetime nuclear effort. Cockcroft became Director of the UK Atomic Energy Research Establishment at Harwell in 1946. Convinced that cheap nuclear power was a real possibility, research into gas-cooled reactors proceeded under his authority. Cockcroft was equally at home on the international stage as he was in the research laboratory. He was highly influential in United Nations sponsored activities for the peaceful use and exchange of nuclear information. At Harwell, he pioneered the next generation of fusion-based research with the ZETA machine.

623

206

THE THIRD JAMES CLAYTON LECTURE

The Possibilities of Nuclear Energy for Heat and Power Production By Professor J. D. Cockcroft, C.B.E., M.A., F.R.S.* The First Nuclear Reactor. The first man-controlled release of "atomic energy" was achieved in December 1942, when Professor Fermi and his co-workers assembled the first atomic pile or nuclear reactor in the squash court of the University of Chicago. This pile consisted of several hundred tons of rather pure graphite into which was inserted lumps of uranium metal and blocks of uranium oxide. When about six tons of uranium metal had been inserted a nuclear chain reaction started and the reactor began to develop heat, derived from the splitting up of uranium atoms. Man had tapped a new power source of enormous potentialities.

The fission process is in general accompanied by the emission of between one and three secondary neutrons (Fig. 2). It results also in the formation of two radio-active atoms. During the succeeding sequence of radio-active changes,

The Fission Process. The work of Hahn and later of Frisch and Meitner showed that neutrons could split up the nucleus of uranium into two fragments, a typical process being the upper figures denoting the masses and the lower figures the nuclear charges. The fission fragments fly apart (Fig. 1) with speeds of the order of 109 cm. per sec. and have a kinetic energy of about 180,000,000 electron volts (1 electron volt = l.6x 10-12 ergs).

Fig. 2. Liberation of Neutrons on Fission of Uranium electrons and gamma rays are emitted, the total average energy being about 20,000,000 volts per fission process. This energy may be developed over a long period of time—up to a year—and is one of the important features in the design of nuclear reactors. The fission of U 238 can be produced only by rather fast neutrons—those having energies over 1,000,000 volts. The chance that a fast neutron will produce a fission in U 238 is rather small—it might have a few chances in a hundred of producing a fission in passing through a centimetre thickness of U 238. The fission of the lighter isotope, U 235, present in a relative abundance of 1/140, can be produced by both fast and slow neutrons, slow neutrons being at least 100 times as effective as fast neutrons.

Fig. 1. The Fission of Uranium The fission fragments are usually unstable nuclei. Thus in the above reaction the Kr97 is fourteen mass units heavier than the heaviest stable isotope of krypton. It will therefore return to stability either by, emitting neutrons or by emitting electrons, and so increasing its nuclear charge. The fission process leads, therefore, to a sequence of radioactive changes in the fission products such as The MS. of this lecture was received at the Institution on 2nd May 1947. * Director, Atomic Energy Research Establishment, Ministry of Supply, Harwell, near Didcot, Berks.

Other Nuclear Processes. In addition to the fission process a number of other nuclear reactions are important in the operation of a pile. First, neutrons can be captured by a nucleus with the emission of gamma rays of energies approaching 6 to 8 megavolts. The chance of such a capture increases as the neutron slows down—there may be particular energies of the neutron when it is in resonance with the nucleus and its chance of capture is specially high. A particular important example of this is the capture of neutrons of a few volts' energy in U 238 to form U 239. This U 239 is unstable and turns into neptunium—a new element of nuclear charge 93—by emission of an electron. Neptunium in turn is unstable and turns into a second new element—plutonium—of nuclear charge 94. The importance of plutonium is that it has nuclear properties very similar to U 235. It has a high probability of fission by slow neutrons and a probability of fission by fast neutrons rather similar to U 235. Slow neutrons are absorbed by most elements and their absorption by the structural materials of a pile is very important. The absorption properties of different materials vary by factors

625

N U C L E A R E N E R G Y FOR HEAT AND P O W E R P R O D U C T I O N 207 of at least a thousand million. Thus an element such as carbon The second requirement for a moderator is that it should have has a very low absorption for slow neutrons. Aluminium absorbs a low absorption for neutrons. them about 50 times as strongly, steel 500 times as strongly, The third requirement is that it should have a high density, boron 140,000 times as strongly. since otherwise the neutron would have to travel too far between collisions and the size of the reactor would be very large. Principle of Operation of a Nuclear Reactor. Nuclear reactors Ordinary water would be the simplest moderator to try, but are of two general types known as fast and slow reactors. hydrogen absorbs neutrons too strongly and the chain reaction The neutrons emitted in the fission process have energies would not develop. between a few hundred thousand and several million volts. When Heavy water is the next possibility and was suggested for this a fission occurs in a large mass of uranium the neutrons produced purpose by Professor Joliot-Curie and his colleagues in 1940. will have a chance of producing further fissions by the fast fission Heavy hydrogen was found to have a very small absorption for reaction. They have also a fairly strong probability of being neutrons and heavy water has been proved to be a practicable absorbed in U 238 to produce plutonium as an alternative to moderator. Unfortunately it is very costly—its pre-war price producing fission. was £500,000 per ton. Nevertheless many tons of heavy water

Fig. 3. Nuclear Chain Reaction Due to this latter factor a fast neutron chain reaction cannot start in a mass of natural uranium. If, however, the lighter constituent of uranium, U 235, is separated from the absorbing U 238, a fast neutron chain reaction can be set up in the U 235 (Fig. 3). This is the principle of the atomic bomb. The slow reactor on the other hand makes use of the fission of U 235 by slow neutrons. A slow reactor consists usually of bars of uranium metal in its natural proportions embedded in a mass of material known as the moderator whose function is to slow down the neutrons to the velocity where they can produce fission in U 235 by the slow reactor—about 2 X105 cm. per sec. (Fig. 4). The first requirement for a moderator is that it must be made up of light atoms because a neutron is slowed down by making elastic—billiard-ball—collisions with the atoms of the moderator. The lighter the atom it collides with the greater the slowing down effect.

Fig. 4. Use of a Moderator to Slow Neutrons to Speeds more likely to cause Fission have been produced—in Canada, United States, and Norway— and at least two heavy water nuclear reactors are now operating. The next possible moderators are beryllium and carbon. Beryllium is very scarce in nature but carbon in the form of graphite has been found to be very suitable, and most nuclear reactors so far built have used graphite as a moderator. The Multiplication Constant and Critical Size of a Pile. The development of a chain reaction in a nuclear reactor requires that one neutron producing a fission should produce more neutrons of which at least one should survive to produce a second fission. A small number of neutrons are continuously produced in uranium by spontaneous fission.

208 C E N T E N A R Y C E L E B R A T I O N S 1947 Assume that one of these neutrons enters a U 235 nucleus in as a foot in a graphite pile between birth and death—this distance the metal rod of the pile and produces a fission. This leads to M, is called the migration length (Fig. 6). the emission of h neutrons where h may be greater than 2. A pile must therefore be over a certain critical size before the These fission neutrons are fast and have energies between a chain reaction can develop. few hundred thousand volts and three or four million volts. By taking account of the fact that neutrons born at a distance Before they leave the metal rods they have a small chance of producing fission by the fast neutron fission process. Their number will therefore on the average be increased to he. They now enter the graphite and, in colliding with the carbon atoms, lose1/6of their energy in each collision. After about 200 collisions they are slowed down to a speed of about 2 x 105 cm. per sec.— the so-called region of thermal energies (the average energy of a molecule of hydrogen gas at room temperature). When they reach these speeds the neutrons are able to produce fission by the slow fission process in U 235 atoms. Not all of them reach these energies, however. Some of them are swallowed up by any structural materials in the pile. If a fraction pt survive this, we are left with hxp1. Others of them are captured by U 238 to form plutonium. If a fraction p2 survive this capture process, we are left with 1 2 = k neutrons produced by one initial neutron. Suppose that the multiplication constant k, is just less than unity. Then one primary neutron produces

hep p

Thus as k approaches unity one neutron will produce a chain of very many more, but the chain is convergent. Fig. 5 shows how the neutron intensity increased in the Chicago pile as it was being assembled.

Fig. 6. Diagram to illustrate "Migration Length" of Neutron Paths M from the face of the pile will be likely to escape, it is possible to show very simply that the critical linear dimension a is proportional to

At the critical size the proportion of neutrons escaping is such that the effective k is The pile is then in equilibrium and the neutron density will remain constant. Control. The next important question is the method of control of a pile. We may imagine that for a pile in equilibrium —with k just equal to unity—some movement of the system may change the amount of absorption and therefore make k — 1 + 8 where d is small. This would lead to neutron multiplication. At first sight one might think this would be very rapid since the time of a neutron generation—birth to death—is of the order of 10-3 seconds, and so in a second we should get 1,000 generations of neutrons and a consequent very rapid build-up of flux. Very fortunately, however, control is made possible by a lag in the emission of neutrons after fission. About 1 per cent of the neutrons are delayed by quite appreciable times—of the order of a minute in some cases. If, therefore, k jumps to 1.01, the additional 1 per cent which brings multiplication will be born up to a minute late and this provides the system with a most desirable inertia. If, in fact, one winds out the control rods of a pile just a little, the power level of the pile will increase exponentially but with a very long period—of some minutes. Having raised the power level to its new value, the control rods are wound back and the power level stabilized. In equilibrium the level of power can easily be kept constant to within 1 part in 10,000.

Fig. 5. Growth of Neutron Activity of the Pile as Layers are Added If k> 1 then the series is divergent and one neutron would produce an infinite number of neutrons in a pile of very large size. If, however, the pile is small many of the neutrons will escape before they can produce more fissions. They may travel as much

Safety. In addition to control rods to provide for stable operation, it is necessary to provide emergency measures to prevent the pile from running away due to faulty operation. The level of power at which the pile is operating can be measured in a variety of ways—by an ionization chamber to measure the neutron flux or by a thermometer to measure the heat developed. Indicators of this type are connected up by relays to operate safety rods containing a strong neutron absorber such as boron; these rods being driven quickly into the pile by springs or compressed air at any prefixed level of power.

627

N U C L E A R E N E R G Y FOR HEAT AND POWER P R O D U C T I O N

209

Fig. 7. Aerial View of Plant at Clinton, U.S.A., where an Air-cooled Pile is operating Heat Generated in a Pile. The energy released in a pile appears in the first instance as the kinetic energy of the flying fission fragments together with the energy of gamma rays emitted in the fission process and subsequently from the fission products. These fission fragments come to rest in the uranium metal rods and their energy is there converted into heat. The gamma rays are absorbed in the metal or graphite and contribute their quota of heat. If a pile is burning up 1 gram of U 235 a day by fission and each fission releases 200 megavolts of energy, a simple calculation shows that the pile generates heat at a rate equivalent to about 1,000 kW. The power level at which a pile can operate depends solely on the rate at which heat can be removed. If the pile is uncooled save by convection then the power level could not exceed about 50 kW. without overheating of the uranium metal rods. The first pile to operate at Harwell will work at about this level. The power extraction can be increased by blowing air past the rods—in this way the power can be put up to several thousands of kilowatts. The second Harwell pile will operate at this level. A next step is obviously to cool by compressed gas. An alternative, and the method the United States adopted in high-power piles at Hanford is to water cool the uranium rods. Fig. 7 shows the layout of the Clinton plant, where an air-cooled pile is operating. Radio-activity. A pile generating 1,000 kW. and burning up 1 gram of U 235 per day is destroying by fission 3 x 1016 atoms per second. This means that at least this number of gamma ray quanta are being emitted, and at the same time 6 x 1016 fission product atoms are born, most of which are radio-active, In the course of time as many of these atoms are destroyed by emitting electrons as are created. Immediately after shut-down of a pile which has been running for a long time at a power of 1,000 kW. we will therefore have 6x 1016 atoms per second breaking up by emission of electrons and quanta. This is an activity equal to that of 2,000,000 grams of radium and it has three results. First, we must surround the pile with a shield of concrete or other material to absorb the harmful radiations, and protect the operators (Fig. 8). One foot of concrete reduces the intensity of neutrons and gamma rays by a factor of 10.

628

Second, even after shut-down we have a high radio-activity and generation of energy, and cooling cannot be immediately dispensed with. Third, the radio-activity of the fission products dies away gradually and even after a year will still be of the order of 20,000 grams of radium. Other parts of the pile structure become radioactive and it is not possible to obtain access to the interior after it has operated.

Fig. 8. Pile Shielding 1 watt power level requires 1 foot of concrete. 106 „ „ 7 feet „ Types of Piles. Piles so far built have used either graphite or heavy water for slowing down the neutrons. Heavy water piles have a larger k than graphite piles and so, since the linear dimension a

they are smaller and give

a larger neutron flux for a given size. Much more compact piles can be built by the use of uranium in which the U 235 isotope is enriched. In such a system k can be substantially increased. The piles can be made smaller, and greater freedom in design results. The limit occurs if one has available pure U 235 or pure plutonium.

210

C E N T E N A R Y C E L E B R A T I O N S 1947

In this case a pile can be made by a few hundred grams of the material in a few litres of water. This small quantity of the fissile element might then develop 100 kW. for ten years. Such piles will probably find a considerable use as research reactors in the future owing to their high neutron flux, which depends on power per gram of fissile material. The Use of Natural Uranium Piles for Power Generation. If we consider a pile constructed of natural uranium, it might at first sight appear that the pile would continue to operate until all the U 235 was burnt up. We cannot, however, assume that we can burn up just the U 235, but for generality we should assume that X times its heat content can be obtained. The important question arises as to what is X. It is determined by four main factors:— (a) A U 235 atom destroyed by fission emits between one and three neutrons. One of these replaces the original neutron; the remaining neutrons are either wastefully absorbed in structural materials and impurities or produce plutonium atoms. On the average each primary fuel atom of U 235 is replaced by Y secondary fuel atoms of plutonium, where Y will on the average be less than unity. This secondary fuel atom can in turn be1 destroyed and the surplus neutrons again used to produce Y tertiary fuel atoms and so on. These partial replacements of fuel atoms mean that in principle we can utilize more than the fission energy of the U235. (b) Fission products accumulate in the uranium metal and their additional absorption will decrease the multiplication constant, k. (c) Changes may occur in the crystal structure of uranium metal and moderator.

(d) Because of (b) and (c) the uranium metal rods have to be withdrawn at intervals, the fission products extracted and the uranium remade into metal (Fig. 9). During this process some portion of the fuel atoms will be lost. We do not at present know how many cycles of purification the metal will have to undergo. For these reasons X is at present uncertain within wide limits and it may well be greater than unity. Principles of Construction. In most nuclear reactors built so far the rods of uranium metal have been sheathed by aluminium to prevent corrosion of the metal and to prevent escape of the radio-active fission products into the coolant. Aluminium has been chosen because of its low neutron absorption. Piles built for power generation will probably use a coolant of high-pressure gas in a closed-circulation system. This gas must have a low neutron absorption and might be either helium or carbon dioxide. The coolant would be passed through a boiler to raise steam, directly through a gas turbine, or used to heat a secondary fluid in passing through a heat exchanger. For these purposes a high exit gas temperature from the pile is desirable. Fig. 10 is a schematic drawing, issued by the U.S. project authorities, illustrating such a system.

Fig. 10. American Official Diagram of an Atomic Pile for Industrial Use If aluminium sheathing is used, the temperature will be limited by the safe stress of the metal from the point of view of creep. The maximum exit gas temperature would then be of the order of 300-400 deg. C. (570-750 deg. F.). It is probable, however, that future developments will remove this limitation by using sheaths of other low neutron absorbing materials such as refractories, which will withstand higher temperatures. It is not unreasonable, therefore, to assume that a thermal efficiency of the order of 30 per cent will be reached.

Fig. 9. Processes involved in Remaking Uranium after Extraction of Fission Products

Capital and Generating Cost. The capital cost of a nuclear power generating system will almost certainly be greater than the cost of a conventional coal-fired system since the nuclear reactor will be more expensive than a coal-fired boiler. Provision would also have to be made for a plant for purifying and recycling the uranium metal. A report submitted to the United Nations Atomic Energy Commission by the U.S. Atomic Energy Project estimated the construction costs of a 75,000 kW. power station to be 25,000,000 dollars and 10,000,000 dollars for a nuclear and coal-fired plant respectively. The cost of the electrical generators and associated equipment would be about 5,000,000 dollars in both systems so that the cost of the nuclear boiler plus chemical and metalmaking plant is estimated at four times the cost of the coal-fired boiler. Only experience can show how sound an estimate this is. Fuel costs in the nuclear power system are subject to similar uncertainties. The pre-war cost of uranium was about £2,000 per ton. To this has to be added the cost of metal making and the cost of chemical extraction of fission products in the number of cycles

629

NUCLEAR E N E R G Y FOR HEAT AND P O W E R P R O D U C T I O N 211 required before the U 235 is extracted. If this increases the quantity of uranium required to satisfy the thermal needs of the United Kingdom and the world. primary cost by a factor Z the cost of U 235 per gram is

One gram generates 1,000 x 24 xy x 0.3 kilowatt hours = 7,200y kilowatt hours.

Z Fuel cost is therefore 0.0094Z/yd. per kilowatt hour. A value of Z\y of 30 would therefore bring fuel posts into line with coal-fired systems. It should be borne in mind, however, that the factor Z must also provide for increased costs of uranium over pre-war prices and for the increased cost of working lowergrade ores. Considerations such as these no doubt led the U.S. report cited above to estimate that nuclear costs might be about 0.80 cents per kilowatt hour or about 25 per cent higher than the cost of a coal system. With the present state of knowledge we can only say that nuclear costs should not be an order of magnitude greater than coal costs. Applications to Mobile Power Units. The application of nuclear energy to mobile power units will be limited by the weight of shielding required to protect operators. This will be of the order of 100 tons for a unit developing a few thousand horse-power. Nuclear propulsion seems therefore to be impossible for planes weighing under several hundred tons, and is quite impossible for road vehicles. Ship propulsion would seem to offer a more favourable field. Practical Realization. It seems probable that a nuclear power generating station designed round a natural uranium pile using aluminium sheathing could be built within three or four years to generate power at an efficiency of 15 to 20 per cent. We could make a reasonable estimate of the capital cost but could not estimate the operating life of such a system. Operating costs are subject to even greater uncertainties owing to our lack of knowledge of the amount of metal recycling required. Only operational experience can provide the answer to these questions. We might, however, reasonably expect that a pile of this kind will be the first step towards the practical utilization of nuclear energy. Subsequent piles might firstly increase the operating temperature, and secondly to improve the overall utilization of uranium. In the course of this work we shall have to "solve the problem of the disposal of radio-active fission products which would be produced as a by-product in very large intensities. These products can be a serious danger to health if they are dispersed in concentrated form. A good deal has been determined, by medical research in the U.S.A., about the safe concentration, but much remains to be determined by medical research. The chemical engineer will no doubt make an attempt to concentrate these products and to use their heat and radiative energy. It seems reasonable to expect that we shall require at least a decade to investigate and, we hope, to solve these technical problems, and to acquire the operational experience necessary for a sound evaluation of the future of nuclear energy's contribution to world power requirements. If we assume that all the difficulties will be overcome and that operating costs are reasonable, we shall be interested in the

630

Uranium Requirement to satisfy Thermal Needs of the United Kingdom. The fission energy of 1 gram of U 235 has been shown on p. 209 to be 91,000 kW.-days. The fission energy of 1 ton of U 235 is 24 x 10 kW-hr. = 24Q. We already have shown (p. 210) that we may release X times the fission energy of U 235. Thus the fission energy of 1 ton of natural uranium = The total annual thermal requirements of the United Kingdom are 300Q, of which 38Q represent the electrical units generated. Thus the uranium requirements to satisfy the total thermal needs of the United Kingdom are

The requirements of the United Kingdom are about one-tenth of world thermal requirements. The total coal reserves of the United Kingdom have been estimated to he The uranium tonnage required to equal this is

World coal reserves have been estimated to be about 80 times the coal reserves of the United Kingdom. We do not at present have information on the total available uranium and thorium. We may perhaps conclude that until we know the total available world supply and the efficiency of utilization X, we cannot assess the overall importance of nuclear fission to world power. We see, therefore, that atomic energy presents a great challenge to the engineer. Before we can realize its potentialities, many hurdles—technical, economic, and political—have to be overcome. All we can say now is that it is worth a determined effort, and that effort we will make. I think that a lecturer standing before you in five years' time will be able to give a much better estimate of its potentialities than I can give you to-night. Professor J. D. COCKCROFT, C.B.E., M.A., F.R.S., Director of the Atomic Energy Research Establishment, at Harwell, near Didcot, Berks, received his technical education at the Universities of Manchester and Cambridge, taking degrees in physics and engineering. He also took a College Apprenticeship course with MetropolitanVickers Electrical Company, Ltd. This course gave scope to his active interest in engineering, which has continued during his eminent work in physics. He formerly occupied the Jacksonian Chair of Physics in the University of Cambridge, where he was also a Fellow of St. John's College. During the war of 1939-45 Professor Cockcroft was Chief Superintendent of the Air Defence Research and Development Establishment of the Ministry of Supply.