THE BRITISH OVERSEAS RAILWAYS HISTORICAL TRUST   A UK Registered Educational Charity

Kevin Jones' Steam Index

Journal Institution of Locomotive Engineers
Volume 38 (1948)

The IMechE virtual library is accessible (full papers, all diagrams, photographs, extensive tables, etc).via SAGE

Issue No. 201

Loach, J.C. (Paper No. 472)
Bogies and pony trucks: their behaviour on the locomotive and the track. 4-22. Disc.: 22-79.
Second Ordioary General Meeting of the Session 1947-8 was held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Thursday 9 October 1947 at 5.30 p.m. In the unavoidable absence of the President, Mr. H. Holcroft occupied the chair.
Includes discussion of the swing-link type used on the GNR and LNER; LMS 2-6-2T and 2-6-4T and the redistribution of weight, especially on obtuse angle crossings. The improved performance of GWR pony trucks with Cartazzi inclined slides which provide the centering forces as well as a certain amount of friction is noted.
H. Holcroft (Chairman) (22-4) noted :On the question of pony trucks, he had done a good deal of riding on engines with such trucks, including the 2-6-0 type with coupled wheels 6 ft. in diameter, and at speeds up to 70 m.p.h., and he had every confidence in them when fitted with Cartazzi side control. There seemed to be a little uneasinesss in some quarters about using a 2-wheel truck on the front of an engine working fast trains, but personally he saw no reason for that feeling. Comparing the wear on the coupled wheels of 2-6-0 and 4-6-0 locomotives he had not noticed any greater flange wear on the leading coupled wheels of 2-6-0 locomotives. There was, however, one weak point about the 2-wheel truck; while it did not appear to subject the leading coupled flange to a greater average side wear and gave good riding on normal track, there were times, as under oscillation produced by track irregularities, when the pressure on the leading wheel was momentarily very much heavier than it would be with a leading bogie under the same circumstances. His remarks on the behaviour of the pony truck at a diamond crossing it would be seen that there was a gap in the track unprotected by a check rail through which the pony truck wheel could, under side thrust, either strike the nose of the diamond or get on the wrong side of it. He had had a fair amount of experience in investigating derailments of that sort, and he had found by repeated trial that they only occurred at a very low speed— a crawl. At anything like walking pace or faster, the inertia of the truck itself entered in, and that prevented appreciable side movement taking place in the short interval of. time available. He also noted that he had never observed a curved diamond crossing. H.C. Barton (22-3) noted the bogie centering devices fitted to bogie electric locomotives and to the "Java" trucks used on GIPR 4-6-2 electric locomotives. D.C. Brown  (24-5) hoped that the author would continue his study to observe the relationship between the trailing truck and the leading truck.
F.J.R. Watts (25-6) expressed particular interest in Mr. Brown's remarks, and said that most readers of the Paper must have been struck by the fact that very little attention appeared to have been paid to the effect of the trailing truck. A good example of that effect could be seen from a study of the derailment of the V2 2-6-2 locomotive on the L.N.E.R. at Hatfield in 1946. On that locomotive were leading and trailing 2-wheel trucks; the leading truck had a swing-link arrangement of the type shown in Fig. 8, while the trailing truck had the Cartazzi arrangement. The arrangement of the leading truck was such that in the central position there was no side controlling force; the side controlling force increased from the zero point at nil deflection to a maximum at maximum deflection. At the rear end, with the Cartazzi arrange nent, if one ignored the redistribution of weight one had a constant restoring force, irrespective of deflection. It seemed to him that there was a case there where the joint effect of the leading and trailing trucks had either been overlooked or ignored.

This V2 locomotive was also interesting in respect of the redistribution of weight. At both ends of the engine there was an arrangement which tended to redistribute weight with sideways movement of the truck, and there was no compensating gear between the trucks and the driving wheels. The derailment near Hatfield took place under conditions where ·there was proved to be an irregularity in the vertical set-up of the rails, and it seemed to him that in those circumstances it would be possible for there to be quite a considerable variation in the controlling force of the leading truck. There was a very small deflection; it was a 76-chain curve. Any controlling force which one might have due to the swing links, which would be small at that deflection, might be completely lost due to track irregularity.
N.W. Swinnerton (26-7) noted that the LMS had many crossovers on curves: suggestion had been made to reduce flange-way gap by 1/8 in. The GWR deepened the flanges on pony trucks by 3/16 inch. Author replied that modifications had been made to crossing nose by reducing it to ½ inch. The concept of deeper flanges was good but insufficient to cover gap. The author (page 37) mentioned a crossing at Melton Mowbray where 2-6-2Ts had to be banned following two derailments. R.E. Pennoyer (27-8) had experience of Zara and Krauss-Helmholtz trucks, and had been impressed. F.C. Johansen (28-9): effect of vibration on friction characteristics of springs. Raised check rails were useful, but caused problems with Mansell wheels. K. Cantlie (29) mainly on Zara truck. P.W. Bollen (29): experience of Zara truck on 1-Do-1 electric locomotives in Germany where he found a "somewhat jerky ride".

B.W. Anwell (29-30)

Centering effort of springs

*ferro-asbestos rubbing plates

A.S. Gillitt (30) observed that Zara trucks were used on many Italian types, especially 2-6-2 passenger locomotives. H. Holcroft (30-1) drew a distinction between Zara and Krauss-Helmholtz: former combined leading coupled wheels with "pony" on a single frame; latter the leading coupled wheels moved laterally under the influence of the pony truck.

Poultney (written pp 44-5) stated that "Looking back over the years many different four-wheeled bogies can be brought to mind as having been used in British practice, but none so extensively as the Adams type which still holds the field. That it has been modified to some extent is true, but the principle remains the same. Generally, the Adams bogie has been accepted without question, but not in all instances.

An alternative which appears to have given good service over many years is seen in the "Webb" double radial truck. So far as the writer knows, this was the only leading "bogie?" ever used on the LNWR. This design used curved guides for the transverse movement and did not turn about a pivot centre. It only moved in the lateral direction. T. W. Worsdell also used curved guides in conjunction with a pivot centre, and as check springs those of the laminated pattern were used. Sir Vincent Raven used the Adams bogie but with independent bearing springs so arranged that the weight on the leading wheels was rather less than on the trailing pair. From the Paper he gathered that the Author prefered an equal-weight distribution. Patrick Stirling used a four-wheeled bogie in which the pivot centre was behind the longitudinal centre of the bogie, thus taking weight off the leading wheels, and H. A. Ivatt using a swing 1ink bogie again adopted Stirling's off-centre pivot and fitted independent bearing springs. Sir John Aspinall used the swing link bogie for the 4-4-2 "1400 class" but later a return was made to the Adams type. The Gresiey Pacifics have a swing link bogie and a very short wheelbase, only 6 ft. 3 in. The Collett, "Kings" on the G.W.R. have a bogie wheel. base of 7 ft. 8in. In both cases independent springing is used. Is there apything in having a long wheelbase and are laminated springs better than coiled for the control of the side movement? Enclosed is an illustration of a geared centring introduced some years ago by the American Locomotive Co., which may be of interest. (Fig, 26.) It is understood this arrangement has been. rather widely used. Regarding pony trucks, the writer does not really like them but realises that they can be and in fact are very useful.

The Author's remarks on the reasons instrumental in the design of the latest LMS pony truck are of very great interest: it is gathered that the Author is afraid the wheels might slip sideways when going over a crossing if not prevented from doing so by a check rail. The extra load placed on them via the action of the swing links might resist this tendency. The friction pads will of course reduce the centring effort set up by the swing links, but if this is considered in excess of requirements the proper thing to do is to connect the pony truck spring suspension with that of the leading coupled wheels; as suggested by the Author. This can readily be done in the case of locomotives with two outside cylinders only. If the cylinders are between the frames it would not be easy. It is believed usual practice to compensate the pony springs with those of the leading coupled axle and in U.S.A. 2-8-0, 2-8-2 and other types having a two-wheeled truck leading compensation is considered essential. By way of illustration fig. 27 shows a drawing of the 2-8-2 design extracted from ' The Railway Mechanical Engineer."

Regarding the swing link four-wheeled truck. The Pennsylvania adopted compensation between the truck and the coupled wheel spring gear 'for the E.6.S. Atlantics and also it is believed for the K.4 Pacifics. This is, however, quite exceptional. the writer cannot recall any similiar application. Churchward's Standard engines were largely based on American designs, and the two-wheeled truck spring compensating arrangements illustrated by ie Author supports this statement.

Author's reply: Mr. E. C. Poultney has raised some very interesting points. The "Webb" four-wheeled radial truck was used on the 4-4-0 compound engines; it was prevented from rotating about its own centre by two bolts, one 9½ in. in front of the centre and the other 9½ in. behind the centre. Although the same general design was followed for many years, holes in the intermediate member through which the bolts passed x~ere elongated and increased in width so that the four-wheeled truck could then rotate about its own centre; thus it became a bogie working in curved guides. It is obscure when the change was first made, possibly about 1904 but certainly by 1906, so the desirability for rotational freedom was soon found. The bogie, as shown in the drawing of the 4-6-0 "Claughton" class was even perpetuated on some of the, earlier L.M.S. class ₅X 3-cylinder 4-6-0 engines (12 with parallel boilers and 53 with taper boilers); the Author, however, fails to see any merit in retaining curved guides when the bogie is capable of rotating about its own' centre. Note Atkins West Coast 4-6-0s at work refutes this statement on page 77.

Sir Vincent Raven's practice of arranging for less weight to be carried on the leading wheels of a bogie was not good insofar as it gave a larger ratio of flange force to downward load at the leading wheel than the normal arrangement.

Lengths of bogie wheelbases are usually settled by considerations outside the bogie itself; its steadiness of riding, guiding properties and flange forces are not nearly so much affected by length of wheelbase as by springing and side controlling arrange ments. Laminated springs in bogies for giving centring efforts introduce a little more friction in to the system and possibly give a flatter line than helical springs, but the convenience of design is usually the governing feature; the Author has the impression that laminated springs fit more conveniently into bogies on metre and 3 ft. 6 in. gauges.

The geared centring device introduced by the American Locomotive Co. uses inclined planes as in a Cartazzi arrangement~ but the friction is eliminated. It is difficult 'to see what advantage is gained by eliminating the friction, although American advertise ments make a point of the fact that geared rollers do so.

W.O. Skeat, (45-8.) wrote that in view of the repeated references to the Krauss-Helmholtz and Zara leading trucks during the verbal discussion, it would perhaps have added to the due of the Paper if diagrams and brief descriptions of those devices could have been included. Unfortunately the only English work, so far as he knew, which, mentioned the Krauss-Helmholtz was Ahrons' Development of English locomotive design. Skeat included three diagrams including on of the Kolomna arrangement (Russian).

M.A. Henstock (58) noted the controversy of high flange forces on the leading or trailing wheels of bogies that he had changed many leading bogie wheels over to the trailers because the wear was greater on the leaders. If wear was any criterion of flange pressures the conclusions would, appear to be clear. He had seen a number of cases where the weights. on· the pony trucks were found to be reduced from what they were designed to be, possibly due to the settling down of laminated springs. This would appear to be a good argument for a compensating device on the lines of the Krauss truck to ensure that the correct weight was always on pony trucks. Whether to go over to complete compensation he did not know, as there were definite complications concerning the practical aspect. The riding of LMS engines with the modern design of bogie was a great improvement due to the stronger side control springs together with ferro-asbestos pads which provided a moderate amount of friction opposing the lateral movement of the bogie. Anyone who had ridden on the Royal Scot with the original bogie and the latest design was in no doubt which was the better. The LMS 4-6-2 engines were also similar at high speeds with the weaker controlling springs originally fitted. On a particular run when one of these engines was running at about 90 m.p.h. he had experienced considerable swaying at the front end, but immediately the brake was applied the oscillation stopped and did not start again. He did not know whether the oscillations were started up by the permanent way, but as no such experience ever occurred to his knowledge when the same engines were fitted with stronger controlling springs, it was fairly certain that adequate side control plus some frictional device was highly desirable.

J. K. Williams (58-9) wondered if a hydraulic shock absorber as used on cars could be adapted for locomotive bogies. E.R. Durnford (59) considered that improved springing was required for pony trucks. H.B. Bailey (59) reported that C.E. Fairburn had noted that the Krauss trucks fitted to 2-Do-2 electric locomotives in Germany hunted rather badly. R.G. Jarvis (59) spoke of the excellence of the ride of German-built 2-10-0s in Turkey which were fitted with Krauss leading trucks and Cartazzi control on the rear axle. D.W. Peacock (59-60) asked about the effect on the calculations of the coning of the wheels and was informed that it did make such calculations more complex.

Darlington 5 November 1947

J.D. Lewis (Chairman 64) requested best form of side control gear: informed that through helical springs designed to be as flexible as possible. G.M. Pargiter (64-5) difficulties, or otherwise, of lubricating steel laminated springs leading to cocoa rust; and supposed difficulties of 2-6-4Ts on reverse curves. D.R. Carling (65): choice of value for coefficient of friction, effect of vibration upon laminated springs, and difficulties with US 2-8-0s during WW2. R.H. Nicholson (65-6) difficulties of lubricating pins in swing link pony trucks and problems of measuring flanger forces. F. Johnson (66) friction of slides did not affect centering action of pony trucks. H. Ormiston (66): cant and coning of wheels and use of raised check rails: Cox was critical of LNER cant formula. T. Robson (66-7) modification to swing link pony trucks on US 2-8-0 locomotives prevented derailments.

Manchester on 17 November

J. Hadfield (74-5) had stated that the sliding type of centre is preferable to the swing link type and had asked the author about control forces. In reply the author noted that there were several theories about flange forces, hut the only one which he regarded as complete, logical and apparently correct was that by Uberlacker. In principle, it meant that a condition was found in which all the lateral forces on the engine balanced each other; in the same way all forces in the longitudinal direction were balanced, and finally all the couples acting in one direction about a selected point were balanced by couples acting the other way; the vehicle was in equilibrium and the flange forces established. Nobody had yet measured a flange force, though there had been attempts to do so by measuring the lateral force between axleboxes and frames; this method, however, did not overcome friction at the treads of the two wheels on the axle. The subject had been fully discussed at the meeting in Derby to the report on which reference should be made. With a 4-6-0 engine, a test had been made with the bogie fitted with calibrated compressive struts ("weigh-bars") for measuring the lateral forces. When the engine was run on to a curve those struts were measuring loads even after the engine had stopped. Then, by taking careful measurements of the positions of the wheels and bogie deflection, it had been found that the lateral force measurements agreed well with values calculated according to Uberlacker.

B.C. McPherson (75) had requested an assessment of the relative merit of side bearers versus centre pivots and in reply stated that the question of transmitting the weight from the main frame to the bogie was very debatable and had been discussed at the London meeting. Experience with the former Midland Railway 0-6-4 tank engines, two or three of which derailed about 1933, the engines tended to nose when running at speed. Stanier converted one engine, putting on side pads and "through spring" controlling arrangements; the running of that engine was improved tremendously, but how much was due to the better check spring arrangements and how much was due to the use of side pads the Author could not say. The conversion of the class was not pursued because it was decided to scrap engines of that type.

D. Patrick (75) had requested further information about Figure 18 showing the spring loading rubbing plates: this was part of the drawing of the leading pony truck on the small [Ivatt] 2-6-2T and 2-6-0 engines the same device was also on the trailing truck of the 2-6-2T. Practically all the pony trucks on the LMS had been fitted with it either when built or subsently and "it worked remarkably well".

Meeting in Doncaster 6th November 1947
F. H. Eggleshaw (70) (Chairman) said the Figures showed that a considerable amount of care had been shown analysing the forces set up on locomotives, especially going round the curves. He confessed that the rocker type of bogie was quite new to him. Another thing which struck him in the Paper was the amount of care which had to be taken in designing pony trucks as compared with bogies. Nearly all appeared to be in connection with pony trucks whereas bogies seemed to get off very lightly. The main thing to him as an old timer was the complications which have been introduced into both bogies and pony trucks. He could remember the old Stirling bogie which was practically a wagon with one long tapered centre pin. He did not think there was any arrangement made for side motion except that the pin was tapered and 3 in. out of centre of the bogie. He feared the pony truck was a very simple affair compared with the pony truck which had been illustrated.

E. Windle (70-1) felt the Paper dealt with the subject on a purely mathematical basis, which did not always work out in practice. They had had various types of pony truck control but had not got down to agreeing with the figures. There were many variables. The variable of tyre wear — immediately the engine went out the tyres began to wear — friction, rail wear, weather; all affected that matter and in looking at the diagram, Fig. 3, he wondered if the Author had brought it to the notice of his engineer. 13.14 tons on one wheel was presumably a calculated figure and he wondered whether in going round at 70 miles an hour he would get 17 tons wheel load. He had shown the easiest example, a 4-6-0 type locomotive and it made one wonder what the experience would be with the 4-6-2, 2-6-4 and 2-6-2 tank engines where any controlling ioao on the front truck was partly compensated by the load of the pony truck at the back. He was not sure if the constant resistance system was better for ordinary 4-6-2 locomotives with a tender. Accumulative rises did not occur when the pony truck went right over to its full extent.

A more difficult problem was the 2-6-4 tank engine. One had a controlling movement on the bogie and one on the pony truck and they would like to adjust the controlling load on the bogie to make it suitable for the engine to go both ways. If one put a bigger controlling load on the bogie one should put looser control on the pony truck for backward running. He would like the Author's opinion as to whether they should have equal controlling loads on the tank to ride stable and stop it from nosing. In the first place, the Author divided the Paper into two. He would rather like to have him divide it into three. There were three distinct methods of control: (1) control by the ordinary single pin swing link bogie, (2) the spring loaded type and (3) the constant resistance. Would the Author give his ideas of the best method of control?

With regard to Fig. 9, he said they had tried to see what effect this pushing over of the pony truck had but so much depended on the friction of the springs etc. and they never seemed to get what they would like to have.

Would the Author say whether the amounts shown in Fig. 9 were calculated or whether they were actual figures which had been taken from experiments, also whether he had any instruments with which he could ascertain any variation on the wheel as the engine travelled. With regard to the LMS new light tank engine, he would like to ask the reason for incorporating two types of pony truck in that design. On bogies and pony trucks did the LMS still put a thicker tyre on the bogie than on the coupled wheels?

The Author in reply to Mr. Windle: In his opening remarks Mr. Windle spoke about variations of friction and rather suggested that tyre and rail wear affected the flange forces. It was agreed that weather conditions affected the coefficient of friction between the wheel and the rail, and whilst the profile of the tyre might affect the riding of the engine, the flange forces were not appreciably altered by it. The tendency to derail may be increased by a worn tyre or side-cut rail hut that was due to the angle of tbe tangent at the flange-rail contact point.

The figures and summaries in the Paper were done with the object of showing the designer what happened with regard to flange forces as a result of altering bogie check spring strengths. If more than one variable had been introduced, the cause of any change may not have been clear.

The Author chose a 4-6-0 to illustrate the point because there were so many hundreds of 4-6-0 engines at present in Great Britain. Though more complicated, the case of a 4-6-2 was well worth consideration.

The case of an engine with leading and trailing units on a curve was considered at the London meeting and Fig. 23 illustrated the point. The chances of. the engine nosing and riding unsteadily were considerably reduced when an appreciable amount of friction resisted the lateral movements of the units.

He confirmed that it was still L.M.S. practice to have flanges on all bogie wheels and on trailing truck wheels of 4-6-2s thicker than the A profile used on most other wheels, including other pony truck wheels.

B. Spencer (71) said that during the period experiments were being carried out in connection with bogie and trailing Cartazzi truck side control on the streamlined A4 class engines, used on the LNER high speed trains, he had an opportunity of discussing with a continental engineer the question of side control values on the leading and trailing trucks of tank locomotives intended for fast working in both forward and reverse directions. He was informed that the possiblity of providing variable side control on the trucks according to the direction of running, was under consideration. Apparently some form of telemotor control was to be operated from the reversing gear. He had not since seen any reference to this matter and asked whether the Author had any knowledge of such an application. The Author in reply said he had not heard of the ingenious idea but it would lead to some complications the value of which could be questioned when one considered the mechanism of the vehicle on a curve.

Stainsby (72) said he was not quite clear as to the relative advantages of the two forms of bogie control springs, i.e., "through" springs as compared with the arrangement of two springs in opposition. The reply noted that the latter arrangement had certain disadavantages. The two springs were opposing each other when the bogie was in its central pisition, i.e. they were each pressing towards the centre pin; thus with side movement there was additional compression on one spring and less on the other. If the springs were not correctly adjusted in the first place, the bogie might continually run against the rail on one side. With the other arrangement the springs were self centring and correct adjustment was always maintained.

G.M. Riall (72): believed that LMS are very keen on side buffer springs between the engine and tender. The reply noted that side buffer springs between engine and tender were considered to help appreciably in steadying the tender when running. as general; it varied with each type of locomotive. So far as the 4-6-0 which he had considered in the Paper was concerned, the restoring effort of 3.3 tons seemed to be as high as it was desirable to have because of the conditions when the engine stopped on a curve with 6 in. superelevation.
Author's reply

In reply to Mr. E. R. Brown he said a very interesting point had heen raised. There were several theories about flange forces, but the only one which he regarded as complete, logical and apparently correct was that by Uberlacker. In principle, it meant that a condition was found in which all the lateral forces on the engine balanced each other; in the same way all forces in the longitudinal direction were balanced, and finally all the couples acting in one direction about a selected point were balanced by couples acting the other way; the vehicle was in equilibrium and the flange forces established. Mr. Brown had asked whether such forces had actually been measured. The Author stated that nobody had yet measured a flange force, though there had been attempts to do so by measuring the lateral force between axleboxes and frames; this method, however, did not overcome friction at the treads of the two wheels on the axle. The subject was fully discussed at the meeting in Derby to the report on which reference should be made.

With a 4-6-0 engine, a test had been made with the bogie fitted with calibrated compressive struts ("weigh-bars") for measuring the lateral forces. When the engine was run on to a curve those struts were measuring loads even after the engine had stopped. Then, by taking careful measurements of the positions of the wheels and bogie deflection, it had been found that the lateral force measurements agreed well with values calculated according to Uberlacker.

The Author in reply to Mr. McPherson said the question of transmitting the weight from the main frame to the bogie was a very debatable one. It had been discussed at the London meeting to the report of which he hoped reference would be made. He would, however, like to add an experience with the former Midland Railway 0-6-4 tank engines two or three of which derailed about 1933; the engines tended to nose when running at speed. Mr. Stanier converted one engine, putting on side pads and "through spring" controlling arrangements; the running of that engine was improved tremendously, but how much was due to the better check spring arrangements and how much was due to the use 'of side pads the Author could not say. The conversion of the class was not pursued because it was decided to scrap engines of that type.

The Author in reply to Mr. Patrick said Fig. 18 in the Paper showing the spring loading rubbing plates was part of the drawing of the leading pony truck on the small 2-6-2T and 2-6-0 engines which the L.M.S. had recently produced; the same device was also on the trailing truck of the 2-6-2T. Practically all the pony trucks on the L.M.S. had been fitted with it either when built or subse

Cox, E.S. and Johansen, F.C. (Paper No. 473)
Locomotive frames. 81-115. Disc. 115-96. 43 digrs. Bibliog.
Third Ordinary General Meeting held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Wednesday 12 November 1947 at 5.30 p.m., Mr. W. Cyril Williams, Vice- President, occupying the chair.
Paper was mainly concerned with plate frames, although the discussion incorporated a major contribution on cast bed frames based on US practice. In part, the paper reflected a major problem of frame fracture as it occurred on the LMS. The frames of the LMS class 5, LNER B1 and GWR County 4-6-0 were compared. The following parameters were considered: plate thickness, stiffness, horn stays, cross stays, horn blocks, and horn guides. The LNWR Prince of Wales were especially prone to frame fractures and poor welding may have led to further fractures.

The following classification was made:

This short list illustrates some of the features which are so baffling. For example, the Class 5, 2-6-0 has the most rigid, and the Class 4 Compound, one of the most flexibly designed, frames on the LMS but both are "good." The Class 5x 4-6-0 is good, but the Royal Scot is bad, and yet the frame layout is very similar in the two cases. The Caledonian 4-6-0 is the best class as regards cracks on the whole LMS — it doesn't have any. And yet the Highland 4-6-0, also designed in the Scottish tradition, is definitely bad.

Fig. 6 in the paper analysed frame cracking in some of the earlier LMS class 5 4-6-0s.

As regards the increase during the war in the volume of frame repair work noticed in the workshops, to which reference has been made, the survey led to the general conclusions that this was due to:
(a) Increased average value of piston thrust under wartime train loading conditions.
(b) Introduction of larger engine classes with higher piston loads in which at the same time better axieboxes gave rise to increased axlebox load offset, and weight limitation had prevented any increase in frame thickness.
(c) Increase in age of large numbers of new locomotives introduced in years 1934-1937.

Curvature affected frame life
Tests were performed both in service and in the workshop: these showed that there is an overwhelming importance of a tight connection at the bottom of the horngap; frame life is significant; the effect of frame stress due to the offest axlebox load acting on the horn guides; the predominant importance of piston thrust. LMS frame construction was modified to be in more in line with Horwich methods. Improvements in welding techniques. The problem was sufficiently severe for consideration to be given to the adoption of bar frames and the cast steel locomotive bed.

Discussion: 12 November 1947 (London):
J. Hadfield (written 115-16) "We have come to expect a good deal from any paper associated with the name of E.S. Cox, and are certainly not disappointed in the present instance". He then advocated rigid frames, the avoidance of abrupt changes in section, sharp corners, but considered that the cast steel bed was not suitable for British and Colonial conditions. B.R. Byrne (117-25 partly written) contributed on welding techniques, including the use of strain gauges and on the experimental welded horn assembly used on the West Country Pacifics. K.J. Cook (125-6) spoke about the technique of welding in the horn-gap area and noted the short life of such repairs.

C.S. Cocks (126-9 written communication) drew attention to the frames designed by Bulleid for the "Merchant Navy" class, fitted with wide fireboxes which influence frame width. To prevent the frames cracking and reduce the number of hot axleboxes to a minimum, the frames were brought to the middle of the axlebox bearings (description given in "Some notes on the 'Merchant Navy' class locomotives of the Southern Railway Read before the Institute of Mechanical Engineers in December 1945). The frames are made of plate 1 1/8 in. thick and are to B.S.S. part 24/17 and are 3 ft. 11½ in. at their greatest depth. The frames from the front buffer beam to the front of the firebox are stayed in such a manner as to make an almost complete box girder construction. The horizontal plate stretcher at the front buffer beam is riveted to the frames. The vertical plate stretcher is connected through the frames by bolting to the front flanges of the outside cylinders, and the rear flanges of the outside cylinders are bolted through the frames to the front flanges of the inside cylinder. At the bottom of the frames and below the level of the inside cylinder, the bogie stretcher is bolted to the frames. The back flanges of the inside cylinder are bolted through the frames to the outside slide. bar brackets; and the bottom of the frames at the same point is stayed by the casting supporting the brakeshaft. The top edges of the frames are stayed by the motion bracket supporting the inside slidebars and the valve plungers. The leading horn-guides are fastened by bolting through the frames at the bottom front side to the stretcher carrying the brakeshaft and at the top to the motion plate. At the rear of the horn-guides the bottom bolts also fasten, through the frames, the front flanges of the cross stretcher between leading and driving wheels at the bottom of the frames, and the back flanges are bolted to the front flanges of the horn-guides at the driving wheel. The top of the frames, it will be noted, are adequately stayed by two motion stretchers. The rear of the driving horn-guides are bolted through to the front flanges of the casting between the driving and trailing wheels, and the rear of this same casting is attached by the bolts, which hold the forward flanges of the trailing horn-guides. This gives a con tinuous line of support at the bottom of the frames, from the casting in front of the leading wheel to the casting behind the trailing wheel. This latter casting is bolted at its front flanges to the rear of the trailing horn-guides. The top of the frames at the trailing wheel are stayed by three stays, two situated between the driving and trailing wheel and one behind the trailing wheel in front of the firebox. The frames behind the trailing wheel to the hind dragbox are of double thickness, riveted together, the outer piece being 7/8 in. thick. It will be appreciated that the back ends of the frames are cross stayed by the dragbox. Another interesting feature which may have a considerable bearing on the life of the frame is that clasp brakes, giving equal pressure on each side of the wheel, and being diametrically opposite each other, are fitted to all the coupled wheels. It will also be appreciated that the spring load is taken on the frame stretcher castings,.as would be expected, at the middle of the axlebox bearing. Over a period of 7 years, since the introduction of these engines, there has been no fracture of frames nor any indication of trouble in this direction. A further point to note, although not strictly relevant to the Paper, is that hot axieboxes have been reduced to a minimum. The contour of the horn-guides and frame at the top corners, where fractures generally occur, has an 4 in. radius to reduce stress concentration. The edges of the frame, transversely, are radiused with a 7/16 in. radius to reduce stress concentration where any stay or fitting is made to the edge of the frames. Fitted bolts, generally 1 1/8 in. diameter are used throughout for fastening the stays and horn-sheets and are driven home with a hand-hammer. After the nut is fitted, a spot of welding is applied at the end of the bolt, fastening the nut to the bolt to prevent "slacking back" of the nut. It will also be noted that no holes are drilled in the vicinity of the usual zone of frame fracture, which eliminates the possibility of fractures developing across the reduced area of the plate at this point. Wedges are not fitted to the horn cheeks, but loose slides are bolted to them. In connection with the horn stays, experience has shown that no undue amount of wear takes place due to vibration, or to loosening of the stays. The stays are held in position by four studs 4 in. diameter, two at each side of the horn. The piston load on the "Merchant Navy"' is 31.8 tons and the weight at each of the coupled axles is 21 tons at the rail.

The Paper shows in Fig. 17 that the LMS have had the courage to weld a completely new section into an old frame. It is surprising that as this procedure was carried out, an opportunity was not taken to improve the shape of the frame at the top corners of the opening for the axlebox guides. A further improvement could have been made by welding in the horns prepatory to inserting the new pieces of frame plate. If this had been done and the horn cheeks adequately ribbed to the new frameplates, by welding, the whole structure could then have been put in the furance and low temperature stress relieved before it was welded to the frames. There would then have been no bolt holes or bolts required and in consequence, no fear of the horns becoming loose. This would have resulted in a frame in which the dangers of cracking would have been reduced to a minimum or eliminated entirely.

If- the frames of these older engines - after this operation are then not strong enough for the work required of them, consideration might be given to welding a suitable strip of bar material, say  4 in. x  1 in. to the bottom edge of the frame, thereby making the whole frame an inverted T or L section.

A frame designed on these lines and made sufficiently robust, may not even require to be stayed at the bottom of the horns with a hornstay, except that a stay could be fitted to enable the engine to be lifted. There must be a large number of engines actually doing useful work after the hornstay has loosened and therefore the time has surely arrived, when the foregoing should be considered in design, i.e. to make the frame sufficiently strong to withstand this condition. In reply Cox noted that the wide firebox of the Merchant Navy class helpful in frame design and enabled many praiseworthy features to be incorporated.

T. Henry Turner (129-31), said the Paper dealt with :he design of locomtive frames but made no reference to the composition, the chemical analysis and micro-structures, of the steels from which the frames were made. He wished to ask the Authors whether they thought it would be possible to follow aircraft or automobile practice by designing locomotive frames so that the skin stress in the metal was further from the neutral axis of the various members. The solid plate or bar frames located stresses very near to the various neutral axes and the stresses would obviously be less if they could be transferred by using a bulkier construction in light alloy as in aircraft practice or a hollow box steel construction as with automobiles.

W.A. Tuplin (131) was critical of a system which tolerated cracking and repair by welding. F.J.R. Watts (131) considered the shape of the horngaps and considered that sharp angles should be avoided. M. Crane (131-2) considered that the quest for low weight was the root cause and advocated bar frames. J. Blundell (132-3) was critical of the sharp corners and the off-set location of the springs. R.C. Bond (133-4) considered that frame cracking was a source of serious trouble; the design of cross stays; and favoured the cast steel bed. T. Baldwin (134) discussed the problems of machining cross girders.

H. Holcroft (134-8) stated that frame fractures had long existed. When locomotive repairs spread over weeks repair was not a problem, but [now] repairs must be completed in days rather than weeks, frame fracture delayed matters and upset a progressive system, hence the great urgency to obviate the fractures by design to meet the forces acting in the frames. When serving his apprenticeship on the GWR he was much interested in the frame fractures of engines coming into the shops for repairs, and he had made notes of characteristic frame cracking After reading the Authors' Paper he turned up these notes and sketches, made nearly fifty years ago, but there was no parallel at all to the fractures dealt with in the Paper.

Most of the cases concerned 0-6-0 double-framed goods engines in which there was only 11 in. depth of ¾ in. thick plate above the horngap, and in nearly all cases thgre was note which read "horn stay loose," "slack" or even had "1/8 in. play." Notwithstanding this fractures did not occur in the top corners of the horngap, as would be expected by the Authors' experience, but they were to be found at the spring hanger brackets of the overhanging laminated springs, at the angle where the plate was carried down to the horn-stay (Fig. not repro) The depth of plate at fracture was about 15 in. and the fractures usually passed through a rivet hole, progressing diagonally upwards in the direction of the top of the horn gap These fractures were found at leading, driving and trailing wheels, and the same characteristic showed itself in a sandwich frame engine having two ½ in. plates separated by 3 in. of oak.

That these fractures did not arise from propulsive forces was confirmed by the same weakness being shown in the double frames of the 2-2-2 type single wheelers, at the leading and trailing carrying wheels. The horns of the 7 ft. driving wheels being set higher in the frame plate at a hump, no such fractures were to be found there. Frame fractures were due to fatigue of the metal under variable stress after a large number of repetitions, and the, metal offered the least resistance to fracture when the loads producing the stresses alternated from plus to minus in range. These were the forces which. should be looked for first of all, because they were the most damaging. The flange forces transmitted. to the frames were of this nature, acting through the axleboxes, first one way and then the other.

He maintained that here was the probable icause of the fractures which he had illustrated. The legs carrying the horns formed cantilevers which set up a bending moment in the frame under a lateral force from the axlebox situated between the legs, and this was resisted by torsion set up in the upper part of the frame, and thus there was a concentration of stress in the, angle at the spring hanger bracket. In his view flange forces played a much larger part in frame fractures than' the Authors estimated.

In support of his contention he would like to bring in bogie frame fractures, a matter not touched upon by the Authors. After the formation of the Southern Railway in 1923 he was much impressed, when coming into contact with the Brighton engines in the shops, that the bogie frames, almost without exception, had one or more patches to cover fractures, or were even double plated. These bogies were of the individually sprung type, whereas the bogies of the South Western and South Eastern sections were mainly of the Adams type, and it was exceptional to find fractures in their frames. Here the weight was carried on a spring cradle on each side, and the frame plates had mainly to resist flange forces. With the individually sprung bogie, the ends of the frame plates were loaded as cantilevers in the vertical plane and were subjected to flange forces in the transverse. Consequently the combined forces produced a high stress concentration 'at the lower edge of the plate where it was attached to the centre casting.

The SR "Schools" class, 4-4-0 3-cylinder engines, were built in 1930 and had individually sprung bogies. In about three years there was an epidemic of broken frames and a replacement of stronger frames had to be rushed through to meet the situation. Seeing that this bogie was the same as had been running for some time under the "Lord Nelson," 4-6-0, 4-cylinder class without trouble, the query arose why this failure occurred. The probable reason was that the "Nelson" bogie was not so heavily loaded and that part of the flange force was borne by the leading coupled wheels.

These fractures of bogie frame plates, where no propulsive or braking force was present, appeared to give further confirmation of the importance to be given to the .study ot flange forces on frames.

In dealing with bar frames, the Authors had set down the advantages on p. 109, but the claims under .(a) and (b) only applied where a narrow firehox necessitated the frame plates being at about 4 ft. 1 in. apart. Where a wide firebox was concerned there was no necessity to adhere to this dimension. Mr. Bulleid's "Pacifics" on the Southern Railway were multi-cylinder and yet the frames were brought in central with the axleboxes, i.e. about 3 ft. 6 in. apart, so that there was no offset.

B.W. Anwell (138-9) (written communication): In considering the stiffness of frames in a lateral direction, no mention has been made of the effect of the platforms in increasing the stiffness of a frame in this direction. This would seem to have some bearing on the two types of frame in Fig. 3. The platforms of the class 5 2-6-0 run at several different levels, and at some distance above the centre line of the coupled axles, and their effect on the stiffness of the frame would seem to be less than those of the class 4 4-4-0, in which the level is almost#unbroken and is much nearer to the axle centre line. Hence the lack of frame stretchers for the class 4 compared with the class 5 engine is not so surprising.

A certain amount of frame distortion has been experienced on the SR Q.1 class 0-6-0, which, it will be remembered, has no platforms but only a very light angle, well ~above the axles, to support the casing. Although 'the frame· had been' stiffened up internally, compared with that of the conventional Q class engines from which the design was developed, the elimination of the platform was considered to be the cause of the distortion.

In the statistical analysis of defects, no mention has been made of other than six-coupled engines, and it would be of inteiest to know what is the relative distributioh of frame cracks in the LMS class 8F 2-8-0 locomotive.

The tentative .frame design shown in Fig. 14 appears to incorporate a number of valuable features, but is somewhat inconsistent in some respects with comments made, elsewhere in the Paper. The 14 in. x 16 in. joists connecting the frame plates, behind the hornblocks appear to be excellent, but paragraph (b) on p. 102 states that the weight of such stays is generally prohibitive. The hornclips shown in Fig. 14 are of the "clip-up" type in spite of the preference for the " Horwich type mentioned on p. 102. Are the bossed holes below the outer flanges of the joists for the spring hangers? If this is so, the spring hangers would be in compression, and in view of the fact that it is usually preferred to have these in tension, have the Authors any comments to mak on this matter?

The extension above the frames, shown in chain lines woul appear to be needed only to support the platform, and would seem to add a lot of weight to perform a relatively small function

As the frame is well stayed against lateral bending, could not the platform be carried from the boiler, as in the bar-framed locomotives? The method of supporting the boiler is not showix 'Is it intended to use expansion angles, sliding shoes, or vertical boiler carrying plates? The latter would seem to be preferable in view of the relatively shallow depth of the frame plate, as the vertical bending stresses are thereby carried to some extent by the boiler.

The method of repairing cracks in frame plates by cutting out the complete section of frame plate around the horngap and welding in a complete new piece of plate appears to be very sound, and it would be of interest to know if any signs of weakness have shown, themselves at the welds after the frames repaired in this manner have been in service for some time. Where constant cracking is experienced an improvement could possibly be made by inserting a plate of greater thickness than tha't of the original frame plate. This possibly might be taken a stage further by welding up a composite frame plate with sections of varying thickness according to the stresses in the respective sections.

No mention is made in the Paper of the design of frame stretchers. In view of the extended use of stretchers fabricated by welding on the LMS, it, would be of interest to know if the performance of 'these is at 'all inferior to those built up by riveted angles.

E.C. Poultney (139-41) noted that a statement is made quite early in the Paper that "If its (plate frame) inherent flexibility is recognised then it is possible to produce a frame of this kind which will give good service." When, however, the Paper is considered as a whole it would seem that considerable difficulties are to 'be overcome before' such frames can give good service, though at the same time it would appear. that some of the difficulties met with might not have been quite so prevalent had not important departures been made from former standards which hitherto one rather regarded as being axiomatic.
Piston thrusts appear, it is stated, to be the predominant factor affecting frame fractures. This will be agreed. But what are the piston thrusts that would seem to have caused so much trouble. So far as the LMS is concerned, the following tabulation supplies the answers: —

These loads do not appear very formidable and, hardly bear comparison' with those found in conjunction with frames of. other types for two cylinder engines in American practice. Two examples are given below:

4-8-4 Combination, passenger and freight N.Y.C. Class S1 140,000 lb
4-8-2 Same service as above, Penn. Class MIA 142,000 lb.
The N.Y.C. engines have a locomotive bed frame and the Penn. engines cast bar frames.

The reasons given for increasing troubles with certain engines are given as being:

1. Increased average piston thrusts.

2. Longer journals with a correspondingly greater offset. Centre of journal to frame face. No increase in frame thickness.

3. Increase in the age of engines.

It is difficult to follow how (1) can have any effect on the framing. Surely maximum not mean piston loads are the cause of the frame cracks experienced, especially if Fig. 6 illustrates the principal defects. Cause (2) is quite understandable, and is, in fact, very close to being the whole of the story. It supplies the principal reason for the superiority of bar frames, but (3) is not quite so clear. Do the frames get tired?
Passing from the foregoing and considering certain design features. It has never been clear why the old horseshoe form of axlebox guides have not been retained. It would appear that the use of what is called "horn blocks"' in place of the horseshoe type has much to do with the cracks as indicated by Fig. 6. No douht as the Authors rightly say, the whole subject is complex because some engines even if they have what is considered the better type guide still give trouble. Even so, the horseshoe guide in conjunction with a really good hornstay should help materially. The Horwich hornstay as recommended 'by the Authors should be quite satis factory. In the case of most troubles, so far as machinery is concerned, these are two main points, that should first receive attention. These are: (1) design and (2) materials. So far as (1) is concerned, the steps being taken by the Authors are logical; the only thing to say is that it is surprising to find modern engines being built without proper support for. the axleboxes. In regard to material, are the Authors satisfied that the frame plate are of the best class of steel for the job, which appears, to say the least of it, exacting, i.e. for the thickness of the plate used?

A series of Papers recently discussed by the Iron and Steel Institute, see The Engineer 28th November 1947, may be interesting in this connection. One further point recommended for the attention of the Authors. Why not try and distribute the loading due to piston thrust better beween the different axlebox supporting guides? The 4-6-0 class 5 engines have a maximum piston thrust of 60,000 lb. Now by applying articulated main and side rod assemblies, the thrust loading on the main guides could be reduced from 60,000 to 40,000 lb.

The N.Y.C. class' S.i 4-8-4 engines can develop a maximum of 6,6oo I.H.P. but by using articulated main and side rods the maximum main pin loading is only 70,000 lb., 15 per cent, more than the LMS, Class 5-4-6-0 type. It may perhaps be difficult to apply this form of rod assembly to the LMS engines with Walschaert gear on account of the space required fo'r the return cranks, but for engines fitted with poppet valves it may be possible.
One word in defence of the LNWR engines said to cause trouble. These have horseshoe guides, but not perhaps of sufliciently robust design. However, the 4-4-0 and 4-6-0 "Georges" and " Princes " have done a lot of really hard work for their size. In fact, on this basis, no engines have done more. The 6o ton "Georges" regularly worked 400 ton trains, and heavier loads have been noted.. Both designs had a central bearing for the crank axle which must have taken up some of the, piston thrust. When these engines developed major frame defects were these bearings still in place? In regard to the "Royal Scots," their troubles seem to have seriously started about 1935. When were the bogies fitted with side bearers 'to take the load?

William M. Sheeehan (written 148-58), Vice President, General Steel Castings Corp., Pennsylvania, wrote at length of the advantages to be gained from the cast steel locomotive bed. 6000 had been manufactured for the USA and Canada, and they were also in service in Mexico, Australia, France, USSR, Turkey and the Union of South Africa.

Meeting at Derby on 16 December 1947

E.R. Durnford (160-2) noted the advantages of bar frames and "What a grand hornclip one could get with such frames, with the boxes and springs just where they were wanted [but] had the disadvantage that they were not so convenient for the attachment of brackets and fixtures as were plate frames."  Probably, the narrow firebox had been the principal cause of British practice being along the lines of plate frames, and the fact that until comparatively recent times, so many British locomotives were inside cylindered, lending themselves to plate frames, and also to British conservatism to change from established practice. If such frames were to remain standard practice, it appeared that some drastic action was necessary.

"The statement appearing in Table I on p. 96, that a slack of 35 thous. made the hornclip ineffective, thus throwing the stress due to a load of 13 tons on the top corners, was alarming, but what might be expected, and it. was noted that the loosening of the clip-up type hornclip commenced at the inner mating surfaces, due to the offset load—this could not happen with bar frames, while Fig. 10 showed in an exaggerated manner the alarming position a frame tended to develop due to offset loading. This offset had been increased in order to get longer axlebox journals, and so to reduce hot bearings—a case of curing one trouble at the expense of introducing another. The Southern Railway in their "Merchant Navy" class had made a noteworthy effort in eliminating this objectionable feature. Consequently, owing to the difficulty in keeping tight the clip-up type of hornstay, the Horwich type clip was being introduced, but he doubted if it would effectively cure frame fractures, since it gave very little lateral stability. The horseshoe guide, properly fitted, was, as might be expected, of considerable benefit above the frame gap and it was suggested that for outside cylinder engines, as was now general, it might be xeentded to form a box joining the frame plates, as was sometimes done in radial boxes, but it was appreciated it would entail expensive machining and lining up, and in the case of fracture it would he a heavy repair job.

Since there were on the LMS some hundreds of class 5 mixed traffic engines, none of which had frames meeting the opinions expressed by the Authors, surely it was time that a frame for this class of engine was put in hand, based on the results of the experiments carried out by the research department. This should not be difficult, one such tentative design being shown in Fig. 14 on p. 104 although it had the objectionable feature of an offset load. Anything to avoid the heavy outlay on frame repairs by welding in new pieces was well worth following.

J.K. Williams (162) suggested that flange forces contributed to the poblem and suggested fabricating the frame in two halves. D.R. Carling (162-5) supported the rigid frame concept, considered that the boiler should form part of the structure (a dictum followed in the USA) and that bar frames should be adopted. J. Howlett (165) was mainly concerned with the techniques of mathematical analysis for stress cracking using Southwell's method, and made reference to electronic computing machines, such as ENIAC and ACE. T. Rhead (166) also refered to Southwell's method. M.A. Henstock (166) noted that flexible frames were prone to fracture from stresses at the corners of horngaps. L. Barraclough (166-7) suggested plate reinforcement at stress points. R.G. Jarvis (167-8) showed with a diagram how bar frames could accommodate 4-cylinder layout adopted for Duchess Pacifics. J.M. Jarvis (168) woindered if manganese steel axlebox liners wouuld reduce cracking. J.C. Loach (168) sharply observed that within the Royal Scot class each frame averaged 5 fractures per annum, or about once in ten weeks. This was not due to WW2: regular cracking was occurring after eight years in service. The frames were incapable of accommodating the piston loads. H.I. Andrews (168-9) considered the mathematical basis. T. Baldwin (169) noted the high loads on the axlebox guides of the driving axle: 40 tons in the case of the class 5.

Meeting in Newcastle-on-Tyne 18 February 1948

C.C. Jarvis (Chairman 176-7) was critical of lax construction methods and contrasted this with the jigs used in aircaft manuafacture. G.M. Pargiter (177-8) noted that "the line of drag" tended to pull frames apart and considered the effects of notches as at the corners of horngaps; springing; the bending at points and crossings and observed that horns should be of the horseshoe type. A.E. Bright (179) noted that the cause was fatigue to which flexibilty is a contributor. There was no problem with double frames. R.W. Taylor (179) noted the need for large hornstays as adopted in the USA. Frame cracking affected both 2 and 3 cylinder types. T. Robson (179-80) recommended the use of thick Belleville washers to maintain tension on bolts and pins. H.W. Davis (180) received the reply that plain carbon steel was used with a tensile strength of 28-32 tons/in². Few frames were constructed of low alloy steel.

Meeting in Manchester on 17 March 1948

H. Fowler (Chairman 184) introduced speaker. I.C. Forsythe (184-5) noted Webb patent (25 November 1869) for cast steel frames. He asked about knock on the class 5 4-6-0s and was informed by Cox that this was most severe on the trailing boxes, but that the longitudinal forces imposed by the pistons were the most damaging. E.R. Brown (185) commented upon the degree of rigidity of frame structures. J. Hadfield (185) noted that Bulleid had arranged to bring centres of hornguides into line with piston thruts. B.C. McPherson (186) observed that no significant frame crcking had been recorded on the Horwich class 5 2-6-0s. Turner (186) elucidated that no information was available on the extent of crack propagation before repair was needed. G.C. March (186) questionned the effect of braking methods but was informed that was too small to be significant. D. Patrick (186-7) suggested that plate frames were only suitable for light locomotives. The LMS Garratts suffered from having standard LMS hornclips rather than Beyer Peacock's standard horn blocks and this had led to the cracking problem. J.J. Finlayson (187) asked if vertical welding edges would be more efficient than wedges and was informed that Cowlairs was doing it that way. Shaw (187) added that inclined edges were easier to fit and adjust. J. Southern (188) stated that higher strength steels might help. J. Sinclair (188) commented on the severe frame fractures experienced on Scottish class 5 4-6-0s. The "new" (Ivatt) class 4 2-6-0 obviated the stresses induced when the locomotive was lifted. C.R.S. Low (188) considered that the frames should be thickened to strengthen the frames near stress points. H. Fowler (188-9) observed that if fractures were patched on the MR class 3 0-6-0 then the engine ran hot unless a balancing patch was placed on the good frame.

Meeting in Leeds 4 December 1947

A.H. Madden (Chairman 191) queried the efficacy of repair versus replacement and was informed that welding was appropriate when the area of damage was limited. Hawkes (191-2) With reference to the combined hornstay, this was first fitted to the A.4 class engines in 1935, of which there are 34 engines in this class, and up to the present date there has only been one fractured engine frame. The average age of the engines being 9 to 12 years. The second class of engine to be fitted with this stay was the V.2 class, of which there are 151 in the class, the age of which range from 3 to 9 years, and up to the present date there has not been a frame fractured. Considering that the LNER had difficulties with previous classes of engines with fractured frames (KPJ B17 especially prone), he considered it essential to have the horn gaps stayed at the bottom. He also noted from the records of Class 5 engines passing through the shops fractures were more prone to the rear corners of the gaps than the front. On the other hand in the test carried out on a section of a Class 5 frame, he thought it was stated the stress was considerably more in the front corners than in the rear corners to the extent of 11 per cent. Could the Authors give an explanation of these two different findings? [they didn't]. Hawkes had noticed that the LMS were fitting stiffening plates around certain of the horn gaps on their new engines and from past experience he had found that in due course no advantage was gained by the fitting of these plates. Cox noted that this practice had ceased

S. Hancock (192) queried whether bar frames could crack at same points as plate frames, and was informed that they could. J.C. Spark (192-3) queried whether there was any connection between cylinder fracture and frame cracking and was informed that catastrophic cylinder failure had no relationship with fatigue cracking in frames. T, Mathewson-Dick (193) querying location of cracking was informed that this was attributable to a combination of machine linkage and tractive effort loading, which was greater over the leading gaps. E. Windle (193) was surprised that there was little mention of actual rail loads: Cox in reply stated that lateral loads were unlikely to exceed 10 tons - far less than piston thrusts. The latest Doncaster stay fastened the horn blocks and the frame at the same time.

S. Hancock (193) refered to the L&Y 2-4-2Ts and noted problems with slipping, heated bearings and wear on axleboxes and horn cheeks and was informed that this had little to do with frame cracking. R.E. Ketley (194) queried frame life for the 8F class and was informed that it was only medium. Dickson (194) was informed that there was no advantage in the use of alloy steels. H.E. Brown observed that horizontal welding was avoided as far as possible. Hawkes (194) refered to high temperature welding.

Issue No. 202

Lynes, L. and Shephard, C.A. (Paper No. 474)
Southern Railway all-steel suburban electric stock. 205-38. Disc. 238-57. 24 diagrs.
Fifth Ordinary General Meeting of the Session 1947-48 was held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Wednesday 14th January 1948 at 5.30 p.m., Mr. Julian S. Tritton, President of the Institution, occupying the chair. C.M. Cock (238) agreed that experience of the riding of the bogies had shown up the motor bogie to some disadvantage as compared with the trailer bogie. The riding, he said, became progressively worse as the mileage increased, and he agreed that this could not be put down entirely to the effect of the unsprung weight of the traction motor, about which very little precise information was available. It was hoped that in due course tests could be made to establish some definite data regarding this feature. The rough riding was just as probably due to the heavier weight and stiffer springing of the motor bogie. In the course of service the rough riding became progressively worse. It seemed to him that it should be possible to produce some material with better wearing qualities for the axleboxes and the bolster wearing parts than that which was at present used—steel, cast iron and manganese steel—so that the clearances were not so very appreciably increased with service as they were with the metal linings. He hoped that the Authors would be able to give this feature serious consideration with the idea of retaining easy riding of the bogies between the periods of overhaul.

Another and more important feature with regard to the motor bogie, 'and indeed with regard to the coach generally, was the weight. The movement of weight required power, which required fuel; and, taking the Southern Railway as an example, and taking the copsumption of electrical energy as 70 watt-hours per ton-mile (which was the average for all services, including intermediate stopping trains, ~express trains and suburban trains), it would be possible, if the weight of a coach were reduced by 1 ton, to make a saving of 4 tons of coal a year· on the basis of a mileage of 75,000 per year. There would be a saving of 4 tons of coal for every 1 ton reduction in weight.

Taking the railways as a whole, and assuming that 40,000 coaches were running about the country, and that each did an average of 6,000 miles a month or thereabouts, on the basis of the consumption for electric traction and for similar services it would be possible to make a saving of about 140,000 tons of coal a year. It would be necessary, however, to take into account the difference in thermal efficiency of the equipment used for the generation of electrical energy and that of, the steam locomotive, so that for steam traction in this country he thought it would be fair almost to double the figure which he had mentioned. That would mean a saving of coal approaching 250,000 tons per annum.

Stanier (238-9) commented on the transmission losses with electricity, and the many other reductions that it was necessary to make. He thought that the Paper showed the trend of coach design in this country, and that they were gradually getting towards all-steel stock. He supported C.M. Cock's remarks about the saving in coal per ton of reduction of coach weight. Mr. Fairburn said some years ago that if it were possible to save a ton per coach he could save a year in the cost of current, which would, be a very important saving~to make. As a result, the LMS Railway built the comparatively light weight stock of the Liverpool. and Southport and Wirral railways. The Paper showed the trend of British railways towards all-steel stock, but Stanier thought that the weight of the stock was still far too heavy. The weight per coach of the stock described in the Paper was something like 35 tons, 142 tons for four coaches. If he remembered rightly, the Liverpool and Southport trailer coach weighed 22-25 tons.

Mr. L. Lynes intervened to point out that in the 4-coach units described in the Paper the weight was 43 tons for the motor coach and 28 tons for the trailer coach; the motor bogies were heavy.

Sir William Stanier observed that even 28 tons for the trailer coach was heavier than the Liverpool and Southport stock.. He thought that the designers of new coaches ought seriously to consider how they could face the risk of corrosion and yet cut down the weight. The great problem with all-steel coaches was corrosion, and it required a very careful technique and proper ventilation between the panelling and the outside sheeting of the coach to avoid condensation forming. He believed that the British railways had realised the importance of this. He did not suggest that that was due to the recent revolution; he thought that it started some years ago. The time was getting nearer when they would have all-steel stock and be able to reduce the weight, and he looked forward to it.

W.S. Graff-Baker (240-1) stated that stressed structure coach bodies were not novel as they had been used by London Transport. J.W. Eling Smith considered that the buffing arrangements could lead to telescoping. W.L. Topham (242-8) criticised the bogie weight and centre buffers and mentioned light weight Italian rolling stock. S.W. Marsh (243-4) commented upon ride. J. Pelham Maitland noted the strength of the new rolling stock bodies when they had been in collision at Motspur Park. T. Henry Turner (244-5) advocated the use of low alloy steels, but also the need to avoid corrosion. A Campbell (written 250-2) commented on passenger comfort. S.E. Lord (written 252-3) was critical of the separate body and underframe.

Darlington Meeting 31 March 1948.

G.M. Pargiter (254-5) was critical of the centre buffer and its effect on stability, but was relieved that bucket seats (presumably as on Tyneside) had not been fitted. D. King (255-6) critical of lack of strength in design

Koffman, Jury. (Paper 475)
Some aspects of carriage bogie design. 259-307. Disc.: 307-43. Bibliog. 49 diagrs.
Sixth Ordinary General Meeting of the Session 1947-48 was held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Wednesday 11th February 1948 at 5.30 p.m., Sir William Stanier, F.R.S., Past President, occupying the chair.
W.S. Graff-Baker noted the significance of bogie mass: London Transport bogies post-1937 had only one motor. Noted the significance of manganese steel axlebox liners and horn guides as they gave a service life of 100,000 miles. Radius arm bogies had been used on the LCC tramways and on the Chicago, Milwaukee & St Paul RR. He introduced the revolutionary rubber/metal bogie bolster suspension. L. Lynes (312-15) noted the relationship between vibration and passenger comfort, the problem of tyre wear, and the load/deflection characteristics of of helical springs. The suspension arrangements for the Bulleid electric locomotives were described: the centre pivot bearings were fitted to crossbars; there were two main helical springs and two auxiliary rubber springs. T. Henry Turner (315-18) made a plea for a bogie with steerable axles for use on LT tubes, tramways and within steelworks. E.S. Cox (318-19) noted the General Motors high speed bogie (120 mile/h) which incorporated rubber damping and allowed side movement between axle ends and axlebox covers. A.H. Sommer (written 319-21) noted that hornless bogies were in service in petrol railcars by 1934 - these employed Silentbloc bushes. Koffman questionned the use of rubber "there is not a great deal of data on allowable working stress for rubber springs" (this was to become his mantra).W.O. Skeat (written 321) mentioned the design of axleboxes for Kenya Uganda Railway railcars. R.E. Lloyd (written 321-2) discussed freight truck bogie design for North America; James Cochrane (written 322-3) discussed the design of bogies to meet the hostile environment of Argentina.
Meeting in Manchester 27 April 1948
H. Fowler (333) questionned the rate of wear of Silentbloc bushes, and was informed it was very small over ten years. D. Patrick (333-4) questionned whether it was possible to design light vehicles with as good ride as heavy ones and was informed that this is possible. R.P. Carson Dodd (334) was interested in the natural frequency of the human body. A.R. Gawthorpe (334) queried the positioning of bearing springs in relation to ride. N. Thornley (334-5) asked about the length of radius rods and was informed that these should be as long as possible. J.A. Lowe was informed that carriage ride is not influenced by motive power. Whatman (335) discussed the application of Silentbloc bushes to 3 axle bogies.
Meeting in Newcastle-on-Tyne 28 April 1948
G.M. Pargiter (337) was interested in the effect of brake rigging on bogie stability. W.E. Frost (337) considered that the author had over-simplified the question of passenger comfort. T. Robson (337) considered the potential for torsion bar springs. R.W. Taylor (338) if hornguides are replaced by radius arms the length of the arms affects hunting. The bushes were force-fit.
Meeting in York 18 March 1948
T. Boggon (340) refered to buffer heights. F.W. Staveley (340) queried whether hydraulic springs could be used and was informed that not aware of, and cautioned about fluid temperatures. H. Roberts (340) commented on military tanks where torsion bars plus hydraulic shock absorbers were used.  R.E. Ketley (340-1) commented on bogies for Sentinel railcars and Koffman stated that they had very long swing links. J.O.H. Derry (341) spoke about bonded rubber bushes. H.T. Thomas (341) mentioned safety devices for torsion bars. N.H. Booth (341) discussed the assessment of bogie ride quality and C.F. Adams (341) the elimination of hornblocks.

Issue No. 203

Carson Dodd, R.P. (Paper 476)
The start and progress of spring drives in electric locomotive road wheels. 357-416. Disc.: 416-18. 29 diagrs.
Annual General Meeting (Session 1946-47) of the Manchester Centre was held at the College of Technology, Manchester, on Wednesday 19th March 1947, at 6.30 p.m., the chair being taken by Mr. J. Hadfield, M.R.E.
History of drives: rods were used by City & South London Railway. Gearless systems were also used. Quills were developed by Brown Boveri. Author mentioned the springs developed for the NER 4-6-4 electric locomotive. Some of the advantages claimed:.
1..The axles are indixidually driven, which means a locomotive with low friction qualities
2. Power is transmitted to the driving axles without affecting rotational motion
3. There are no reversing thrusts necessitating the keying up or adjustment of bearings.
4. The springs do duty in either direction of motion.
5. There are no wearing surfaces to be lubricated.
6. Practically all the rotating shocks are absorbed by the springs.
7. There is a sufficient number of springs and they are so mounted that the failure of any part of the drive does not disable the locomotive sufficiently to cause a service delay.
The springs are rigid in the brackets at each end and are fixed to the quill and wheel alternatively, but owing to the relative movements necessary between these parts in service, they are subjected to very severe strains from all directions, sometimes separately and sometimes simultaneously.
Also, when the locomotive is running in one direction for any length of time half the springs (in addition to the above strains) are either in constant tension or in constant compression. The above effects caused failures in springs and fixing bolts but the arrangements on the railways for replacement were so made that no great inconvenience was felt, as a set of brackets and spring is easily replaced when the wheel is turned to the lowest position near the‘ rails.
Before deciding upon the adoption of the above drive, tests were made to determine the best proportions of the springs, by subjecting samples to repeated side deflections, etc. As the result of these experiments the first locomotive was equipped with a spring in which capacity was sacrificed for side flexibility, consequently while there was no breakages of springs, there was some bumping in starting heavy trains.
. D. Patrick (416) requested information on the drive system being developed for the LNER Manchester Sheffield Wath system and was informed that Gresley had examined both solid gears and helical springs on existing locomitives in South Africa. The former were favoured but the contractor, Metropolitan Vickers, recommended a sprung system and the bulk of the locomotives will be fitted with Silentbloc resilient gears, but it was agreed to try one locomotive with solid gears and a second with steel resilient gears.

Newsome, N. (Paper No. 477)
The development of L.N.E.R. carriage and wagon design, 1923-1941. 420-73. Disc.: 473-85 + 8 folding plates. 23 illus., 27 diagrs., 2 tables.
Seventh Ordinary General Meeting of the Session 1947/48 was held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Wednesday 10th March 1948 at 5.30 p.m. Mr. H. Halcroft, Vice-president, occupied the chair.
The paper discussed the design of the compound bolster bogie on the GNR, the development of elecreic cooking, the construction of he five-coach articulated train in 1921, rubber springs for buffing and draw-gear which were "extremely reliable in service", the adoption of GNR standards by the LNER (bogies and dimensions for carriages: RCH standards used for freight wagons). New trains were introduced for the Flying Scotsman (with articulated restaurant car) and the Hook Continental: electric cooking was used on both. The quad-articulated sets were developed for the Great Northern line services and were designed for conversion to electric multiple units and quintruple articulated units were introduced on the Great Eastern, but problems were encountered with overloading. In 1926 articulation was applied to sleeping cars where a weight saving of 10 tons was achieved. A shower was fitted to one of these units in 1930. Electric water heating was fitted in sleeping cars to eliminate gas. In 1927 all-steel coaches were purchased from outside manufacturers. In 1928 a new Flying Scotsman train was introduced with interiors designed by Sir Charles Allom. On the freight side large wagons were developed for conveying heavy loads, and new wagons included hoppers for ballast, and bulk grain. Fitted freights and containers were promoted. An experimental brake van was built of concrete. Pressure ventilation was introduced and Rexine was used in sleeping cars and eleswhere, A coach was constructed from aluminium alloy, Alpax, in 1932. Buffet cars were adapted from exisiting vehicles for the King's to Cambridge service. The high speed trains (Silver Jubilee, Coronation and West Riding) clearly involved Newsome and he was closely involved in the high speed trails with Mallard where a Westinghouse modified form of vacuum brake was developed. This was also applied to the East Anglian and following WW2 to the Yorkshire Pullman. An unusual adaption of electric heating was applied to container flats which transported road tanks of edible oil from Selby to Scotland, where the tanks had to be kept warm. All-steel hoppers were developed for coal traffic, but discharge was difficult. Large plate wagons were constructed. Articulated twins were constructed for local services. Anthracite ranges were introduced into catering cars for cross-country services: by 1937 gas had been displaced by electricity in most catering vehicles. Kruckenberg bogie suspension, using rubber balls, was developed with Spencer Moulton. A new Flying Scotsman set was introduced in 1938. The Tyneside electrics were articulated twins and enjoyed bucket seating.
H. Holcroft chaired the London meeting: he noted the significance of Gresley's adoption of articulation and the way in which the Buckeye coupler and Pullman vestibule inhibited damage in accidents. He proposed that concrete should be re-examined for rolling stock as there had been considerable progress in concrete technology. L.J. Le Clair (473-7) gave a vignette on Gresley and the high speed brake trials of the A4 when 126 mile/h was reached. This includes pictures of 4468 Mallard and its crew, and of the rolling stock and of graphs of braking experiments performed near St Neots.
T. Henry Turner (480-1) noted that when a new design of carriage was coming along, Mr. Thom would make a mock-up at Doncaster, and then invite his friends to come and sit in the seats to see if they were the right height. The chief mechanical enginee:r was a tall man and Mr. Thom otherwise constructed, so perhaps the average passenger had benefitted from their varied opinions as to the seats.
Sir Nigel Gresley carried out much industrial research. The public may sometimes think of research only as that which a chemist or physicist carries out in his laboratory, but all that the Author had spoken about constituted a continuous piece of industrial research in which Sir Nigel rode on the test vehicles himself.
The speaker recalled an occasion when Mr. O. V. Bulleid, Sir Charles Allom and himself were invited to travel with the chief from London to Peterborough in a couple of brake vans, the floors of which had been taken up above the bogies. Both the vans were loaded to the same extent with pig iron and old brake blocks, and the total transverse and longtitudinal springing was the same in both cases, but in one case the springs were stronger longitudinally and in the other they were stronger transversely.
It was thus possible to stand on the moving bolster while holding on to the guard rail and so closely to watch the spinning wheels, and as they whirled along the track one could not help marvelling that the tyres stayed on, because they assumed the appearance of flywheels, and one realised that centrifugal forces were counteracting the shrinking on stresses.
The Author had mentioned that when the especially high speeds were reached, trouble was met with in connection with the brake gear.. The old "common iron" brake shoes also broke in service and might have been dangerous, so their dimensions were increased and a change made to the stronger "cylinder" grade of grey cast iron.
There would be, therefore, in the future a problem for the mechanical engineer when high speeds came again, because cast- iron brake blocks would not stand the excessive heat caused by high speed braking.
Shortly before he died, Sir Nigel Gresley arranged for a hundred 20 ton side discharge hopper wagons to be specially built and, at the suggestion of the Iron and Steel Institute's Corrosion Committee, he asked Mr. Cruddas, who was building them, to use as floor plates four different types of steel, namely, ordinary mild steel, copper bearing steel, and two varieties of low-alloy steels. These hundred wagons thus had four different steels exposed to identical wear and tear and atmosphere.
Every half year Dr. Hudson and the speaker had been to look at the wagons and had taken measurements. It was clear from their last report that if it were desired to use less steel in making wagons, and if a little extra trouble could be taken with regard to welding, low-alloy steel was to be recommended.
Those were the only tests, so far as he [Turner] knew, of a systematic character which had been carried out in this country to show the virtue of low-alloy steel in railway wagons.
Meeting at Doncaster 22 April 1948
F.H. Eggleshaw (Chairman 483) introduced speaker. A reply to J.N. Compton (483-4) stated that teak had to be abandoned due to cost. Pressure ventilation required cooling in summer. Manganese liners had been used experimentally in axleboxes. R.E. Ketley (484) was informed that it was impossible to heat carriages electrically due to the load on the batteries. F.W. Staveley (484-5) noted the wear of brake gear on the high speed trains. T. Boggon (485) noted that paint applied to aircraft had an effect upon speed and wondered whether the same technique could be applied to high speed trains. F.W. Staveley (485) observed that steel windows has corroded. R.E. Ketley (485) noted that wet coal led to wagon corrosion.

Issue No. 204

Rudgard, Harold (Presidential address)
The user of locomotives for revenue earning. 494-527. 32 diagrs.
The opening general meeting of the session 19-18/49 was held at the Institution of Mechanical Engineers, Storey's Gate, London, S.W.1, on Wednesday 15th September 1948 at 5.30 p.m., Mr. Julian S. Tritton, retiring President, occupying the chair.
Includes a chronology of locomotive development. Paper was critical of mileage and availability of selected modern locomotives.

The locomotives were not identified but the 2-6-2 could only be the V2 class and there is a probability that "F" was the Western Region.
There is also a wealth of tabulated data on the "recent" (i.e. 1948) performance of locomotives on a regional basis in terms of hot bearings associated with big ends and axle boxes where it is possible to infer Western Region superiority.
Rudgard was eager to use locomotives more intensively and cited the exploits of Webb's Charles Dickens which worked between Manchester and Euston (return 374 miles) six days per week and clocked up 100,000 miles per annum. He stated: "I have attempted to do this with the LMS 4-6-2 type locomotives, but could not achieve an average, mark you, with 45 locomotives, of more than 67,000 miles (one case of 93,253 and four of 85,000 miles being included in the average."
On the former Midland Railway, with which I was particularly associated, a tribute is due to the memory of Sir Henry Fowler. His collaboration with Cecil Paget, and their understanding of the running shed angle, helped to produce a school of design which was pre-eminent amongst all the railways forming the London Midland and Scottish in 1923 for reliability and low repair costs, and which, continuing under the LMS, lent itself admirably to the introduction of developments in motive power practice and organisation; Sir William Stanier carried on the good work.
Another aspect which a study of the work of the locomotive superintendents or chief mechanical engineers during the past century brings out clearly is, that it is the man, rather than any particular system, which has been the more important. At different times and on different railways, the chief locomotive officer may have had a background leaning in one case towards the design or workshop angle, in another case, towards the running and operating sides. Yet again, the mechanical engineering and motive power departments have sometimes been united under a single control, and at other times they have experienced varying degrees of separation. Whatever effect these variations may have wrought in other directions, in the single item of all round effectiveness of the locomotives for the job they had to do, results have been remarkably independent of systems and background.
I regretfully am of the opinion that these good locomotive engineers of recent years were more interested in turning out a newly designed locomotive, and showed little interest, if any, in where the locomotive would live and under what conditions it was housed in order to be an efficient revenue earner. This is amply borne out by any inspection of the motive power depots designed and built in the Churchward, Johnson and Ivatt (Senior) period.
I well remember the remarks of the late Cecil Paget—a locomotive enthusiast at an early age—who was at Harrow School, and who, when the time was available, would go into Watford, and, from a point of vantage on tile bridge near the running Shed, inspect the locomotives in the yard at Watford shed. On his returning to Harrow for a reunion twenty-five years after, Paget's steps took him to the same place and he saw almost exactly thi same things as he had seen twenty-five years previously, namely same men, same open-wick torch lamps, same broken windows the only difference was that the locomotives were larger.
What seems to have been most lacking was a strong well-informed control which knew where it was going, coupled with proper balance between the needs of design and availability of locomotives, shops and sheds, implying the existence of good teamwork.
There are many illustrations of Britsih and overseas locomotives.

Lambe, T.T. (Paper No. 478)
Notes on railway standards. 541-64. Disc.: 565-78.
Indian railway practice. In the discussion (568-70) Cox noted that "Not all railways were so fortunate to have a Churchward succeeded by a Collett, and a Collett by a Hawksworth". Cox also noted that the LMS began with 10,300 locomotives of 393 types. During the first two years the existing types ere augmanted. In 1925 Fowler standardized practically all details on MR practice, but that this had little effect on residual stock. Stanier adopted few Fowler-based details. Thus after 25 years there were two standard groups: one with 2600 members and another with 2200. Cox cast a critical eye over the large number of tyre fastenings: old screw, rivet, Gibson ring and Bulleid which lacked any.  Other discussion: E.V.M. Powell (565); W.A. Nightingale (565-7), M.A. Crane (570-1), K.J. Cook (571), K.H. Leech (571-2) and B.W. Anwell (572).

Journal No. 205

Koffman, Jury (Paper No. 479)
Adhesion and friction in rail traction. 593-640. Disc.: 641-72. 47 diagrs. Bibliog.
General Meeting was held at the Institution of Mechanical Engineers, Storey’s Gate, London, on Wednesday 20th October 1948 at j.30 p.m., Lt.-Col. H. Kudgard, O.B.E., President of the Institution, occupying the Chair.
This review paper covered all aspects of friction and adhesion on a global basis. Rudgard (Chairman 641) agreed that materials other than cvast iron are needed for brake blocks. Care needs to be taken to avoid damage to flange. Cox (641-2) noted that it was difficult to discuss a review paper. claimed that slipping at high speed "did occur, but only as a very rare phenomenon". Suggested that Co-Co was most suitable for general purpose electric locomotive in United Kingdom. Noted that factors other than ratio of adhesion entered into whether steam locomotives slipped; these included steam passages, valve events, sensitivity of regulator and weight distribution. Also  noted that flange pressure had not been investigated. Tellingly Koffman considered that four axles and a 20 ton axleload would suffice. J.S. Tritton (642-3) noted the difficulties encountered in braking long trains due to variations in brake block pressure and added further factors: coupled wheels being loose in axlebox wedges, boxes knocking and slackness of big ends. W.F. McDermid cited his own 1935 paper on brakes for streamlined railway vehicles which had cited Douglas Galton's friction of cast iron brake blocks but Koffman noted the limitations of Galton's curves. D.C. Brown (648-9) gave an excellent appreciation of Koffman's review. N.G. Cadman (649-51) noted that empirical observations had considered that the changeover from a high brake force to a lower one should take place at 28 mile/h and that Koffman's theoretical studies indicated that this should be 25 mile/h. A.W. Simmons (651) considered the interaction between cast iron brake blocks and the wheel and the effect of vibration from points and crossings. T. Henry Turner (651-3) noted the effects of rail burn, rail cracking, the melting point of cast iron brake blocks, roaring wheels, corrugated rails, and engine oil spillage onto brake blocks (as noted at Paddington, not Waterloo). W.A. Agnew (653) advocated non-frictional methods for slowing and materials other than cast iron. W.A. Nightingale (653-4) considered the influences of bored out cylinders and turned down tyres, and the effect of carrying wheels.

Darlington Meeting 24 November 1948

R.W. Taylor (668) noted that high speed slipping took place on the C7 class. Westmorland (668) had observed slipping on D20 and D21 4-4-0s when going downhill and this was cured by opening the regulator. He noted that the LNWR had used wooden brake blocks. King (668-9) mentioned rubber pneumatic tyres and expanding brakes. S.L. Baister (669) noted that railcars have used automotive-type brakes which did not act on the wheel rim. Tattersal (669) proposed braking via the diesel engine. C.C. Jarvis (669) considered porcelain (too brittle), beechwood (risk of combustion) and carbon as brake block materials and recorded the following criteria: availability, effect of rain and snow, friction independence from speed or temperature, adequate mechanical strength, wear resistance, and lack of unpleasant smells.

Cardew, C.A. (Paper No. 480)
The blower, its origin and its functions on the locomotive: with some notes on several different types used, the draught producing value of some of them, and the steam which they consume. 673-728. Disc. 728-45.
Ordinary General Meeting of the New South Wales branch of the Institution was held in Sydney, Australia, on the 22 April 1948, at. 7.45 p.m., 80 (approx.) members and guests being present. The chair was occupied by Mr. H. Young (Chairman of the Branch).
a) when the locomotive is stationary
(1) The blower can be used to expedite the raising of steam pressure, but until there is some 20 to 25 lb. per sq. in. pressure available the vacuum created by it in the smokebox is hardly measureable on the ordinary water column draught gauge.
(2) It can serve to maintain, or increase, the pressure of steam in the boiler when the locomotive is standing at stations, in sidings, or elsewhere, in the course of a journey.
(3) It is also used, when standing at stations, etc., to reduce the large volume of smoke which might be an annoyance in the neighbourhood. Classes of coals, such as our Northern New South Wales grades, often necessitate the use of the blower in this way.
(b) When the locomotive is running
(1) The blower can be employed to maintain, or increase, the pressure of steam in. the boiler, when the engine is running with the regulator closed, or, in the drifting position. Under these conditions, it serves to keep the fire from dying on long, downhill, sections.
(2) This is to serve the same purpose, and it is used in the same way, as indicated in paragraph 3 preceding (i.e. to clear away smoke), but with the engine running. Even if the heavy smoke cannot he entirely consumed by this method, the blower is frequently of service in at least lifting any smoke leaving the chimney top high enough to prevent it from trailing down over the train.
(3) The application of the blower is often necessary to counteract back draughts.
(NOTE: The value of the blower for this particular purpose appears early to have inspired the idea of bringing it into action automatically. Pearson's well known 4-2-4 single driving tank engines for the Bristol and Exeter Railway, when re-designed and built in i868, were fitted with an interlocking device between regulator and blower, whereby the closing of the former admitted steam to operate the latter. The same principle has been used on other locomotives, down to later times, but has never become general practice.)
(4) Though certainly not designed for this purpose, the blower can be, and sometimes is, used to assist the main engine exhaust in its function pf producing draught when the engine is steaming.
(5) Though not, perhaps, in accordance with the active inten tions of either the designer or the operator, the use of the blower, when the engine is running with closed regulator does, with some blower types, tend, somewhat, to counteract any tendency for ashes to be drawn down the blast nozzle should there, at some time, be any vacuum there, due to suction from the pistons.
Based mainly on practice on NSWR, but also refers to USA and to Britain for historival development. Tests in Australia on C36 class. Discussion: R.S. York (728) noted that on the GNR in England drivers forgot to turn off the blower and caused their firemen extra work. The GNR disposed the exhaust from the vacuum brake ejector into the blower. R.T. Russell (731) noted that the NER exhausted the Westinghouse compressor via the blast system

Issue No. 206

Cocks, C.S. (Paper No. 481)
History of Southern Railway locomotives to 1938. 749-822. Disc.: 823-60 + 4 folding plates. 30 illus., 34 diagrs.
Includes a relatively extensive examination of the work of Urie and L.B. Billinton. Each of L.B. Billinton's designs is considered plus the proposed K/² 2-6-2T. Particular attention is paid to the Baltic 4-6-4T both in the paper and in the subesquent discussion. It is noted that the design was developed from the Marsh 4-6-2T designs and was intended to carry sufficient water to run non-stop to Portsmouth: tanks were fitted between the frames. Weir pumps were a feature of LBSCR locomotives. Urie was noteworthy for the introduction of a high degree of standardization. The 4-8-0T and 4-6-2T designs shared a common boiler, and Walschaerts motion was shared with N15. This last had many components in common with the S15. Urie also simplified and modernized the Drummond 4-6-0s. The exposed valve gear, with large piston valves (but short travel) was noteworthy. The N15 boiler was tapered at the front end to save weight. He designed the Eastleigh superheater.
On the SECR the N class introduced by Maunsell was the first to combine high superheat with long travel valves. The leading pony trucks combined a spherical centre with Cartazzi type (the slides were in an oil bath). The tapered boiler barrel and top feeds and superheater were further advances. The K class (2-6-4T) was fitted with a two-wheeled Bissel truck with a 6ft 4 in radial arm. Cocks mentions the Sevenoaks accident without pursuing this topic in depth other than the conversion of the class to 2-6-0s.
On the foundation thus provided at the grouping, Maunsell laid down a set policy for the future design of locomotives, under the following headings:
(1) Ample power for all requirements to enable sectional timing to be maintained with a high degree of efficiency.
(2) Ease of maintenance.
(3) Suitable for operating on all three sections within the limits of bridge loading.
(4) Standard locomotives with as few types as possible to cover the whole of the requirements. Standard details to be as interchangeable as possible, such as boiler, cylinder, motion, tyres, axles, axleboxes, fittings and boiler mountings.
(5) Belpaire type firebox, with grate area and heating surface well proportioned.
(6) Long lap piston valves.
(7) Freer exhaust passages.
(8) Accessibility of all parts ("Make everything 'Get-at-able.' was Maunsell's remark).
(9) Footplate comfort and ease in handling. In this connection wooden models or mock-ups of the cab with the handles for operation in position were made for the enginemen's inspection and suggestions.
(10) Lubrication of all parts to be efficient and simple. To be under the control of the enginemen as far as possible without leaving the footplate.
(11) Large smokeboxes and capacious ashpans.

page 756 notes lack of water troughs on LSWR due to undulating nature, curvature, lack of suitable water supplies and poor drainage. page 771 noted the involvement of Maunsell in ARLE designs: 2-6-0 (on which N class was based) and 2-8-0 were both designed at Ashford.

pages 793/5: In December 1924, Maunsell wrote to the CMEs of the other British railways to ascertain the maximum axle load permitted on their lines. Hughes had submitted three designs on the LMS where the driving axles carried a load of 20 tons which, at the time, was the extreme load permitted on the LMS. Gresley stated that the highest axle load in use on the LNER was 20 tons 16 cwt., and Collett, from Swindon, said that 20 tons was the maximum built to at the time, but new engineering work would permit loads up to 22 tons. Maunsell investigated every channel both at home and abroad to see whether such an engine of the power required could be built within the weight laid down by the Civil Engineer.
The original scheme for an engine which ultimately led to the construction of the Lord Nelson had an axle load of 21 tons 10 cwt., which was 17 cwt. in excess of the 20 tons 13 cwt. of the final Lord Nelson. The cab was rather a departure from the usual Southern Railway cab, being more in keeping with the old North Eastern. To enable the weight to be reduced, the boiler barrel was shortened by approximately 10 in., and this enabled the King Arthur tubes to be used as the distance between the tubeplates was identical.

The improvement made by the alteration to Engine No. 449, particularly with regard to the saving in coal, was so marked that it was decided to incorporate the arrangement in the new "Lord Nelson" class. The arrangement provides a more uniform torque and more regular effect on the firebox draught than is customary and enables the engine to be worked more heavily without fear of 'breaking up" the fire.

The revolving and reciprocating parts were kept light by using high tensile steel, Vibrac, and the balance weight in the wheels was reduced in consequence: this produced a much lighter hammer blow and influenced the Civil Engineer in accepting an axle loading up to 21 tons.
The boiler was large, and a new feature for engines built at Eastleigh was the provision of a Belpaire firebox. The superheater was the Maunsell type with air relief valves. The boiler has probably the widest type Belpaire firebox that could be used within the limitation of the SR loading gauge, consistent with a clear view from the cab; also the longest firebox it is possible for a fireman to conveniently fire. The grate is virtually in two sections, the rear portion being horizontal as a landing and the forward portion sloped. This caused a definite break in the fire and on occasion led to indifferent steaming when inexperienced firemen are used to fire the locomotives.
The Lord Nelson boiler was originally fitted with steel and copper water stays in the firebox. Copper stays were used on the firebox side for the top six rows and the outer end rows only, the remaining stays being of steel. The steel stays were afterwards replaced by Monel stays and this was probably the first application of Monel stays as standard practice to locomotive fireboxes in Britain. The stays are fitted with steel nuts on the inside of the inner firebox.
To enable the engine to be built to the weight allowed by the Civil Engineer, great care was exercised, both in design and actual building. Certain parts normally left as forged or cast were machined to keep within the weight. So much care was exercised that the engine was actually well within the weight when completed, so the remainder of the class did not receive similar treatment. After the balancing of the engine had been calculated at Eastleigh, the figures were submitted to Professor Dalby, who agreed that the engine balance as shown would be very satisfactory in running.
Cocks notes that the Southern was unlike the other members of the Big Four: in its intensive passenger services, its electrification and its quest for punctuality. On the rebuilt E and D classes he noted that these shared the large N class piston valves.
Page 781: The rigidity af the conjugated valve gears was not mechanically quite satisfactary, for it was found that, awing to. back lash in the joints and springiness in the levers, and movement af the reversing shaft due to. slackness af the clutch, the central valve over-ran its travel when the engine was running at bigh speed in full gear. The clearance between the valve head and the valve cover was as much as 5/8in ; yet it was found, under certain conditions, that the valve head had actually touched the valve chest cover. Tests were carried out and it was found that this phenomenon only accurred when the engine was coasting with the valve gear in the full .over position and the engine working at speed.
A Caledonian engine, fitted with a similar gear, exhibited the same defect and prompted the question being asked as to. whether it were not better to. have a third independent valve gear than to. seek far simplificatian by conjugation.
If the problem had been tackled with sufficient enterprise and experiment a method af controlling this over-travel could possibly have been discovered, and I would suggest that a gear designed to. give as good a diagram as is always found with this type of gear would, even at this time, enable further economies to. be made in locomotive operation. The back pressure line, it will be noted, is almost line and line with the atmospheric pressure.
The 3-cylinder locomotive showed reduced wear, and coal consumption af 10. per cent. less than that of the 2-cylinder engines, but the water consumption was 11 per cent. more for the same amount af work. It was stated that the reduced coal consumption was due to. the better blast produced by six beats as compared with four beats.
page 796 Lord Nelson boiler was never quite so free steaming as King Arthur type. page 798: experimental boiler fitted to 857: "results of tests with locomotive have not carried any conviction. Locomotive with 135º cranks showed no appreciable difference. Page 799: 4-cylinder compound authorized by General Manager in 1932, but not proceeded with. It would have had poppet valves and 15½ x 26 hp cylinders and 22 x 26 lp cylinders.

802: Z class 0-8-0T: the three-cylinder layout gave an even turning moment and rapid acceleration. Originally it had been intened to use conjugated gear, but the Civil Engineer objected to the distance between the buffers and leading wheels. The cylinders were common to the N1/K1 type and gave a quiet exhaust.

804: Schools: originally considered as a small Lord Nelson, but not possible to use its boiler. Hence, King Arthur boiler was used to meet Hastings line loading gauge. Hammer-blow was reuced through the use of three cylinders. "In accordance with usual practice, standard parts are used, the cylinder, motion, bogie, etc are virtually duplicates with the Lord Nelson class".

812: proposed 4-8-0 freight locomotive (Lord Nelson cylinders and boiler); 815 proposed 4-6-2 and 2-6-2 designs: latter used U1 valve gear with 40 ft² grate. Noted conversion of LBSCR 4-6-4Ts to 4-6-0 type and Sentinel railcar (with standard steam wagon boiler) and intended for Aldershot branches

Precis published in Locomotive Mag., 1949, 55, 33

Discussion:
E.C. Poulney (823) contributed to the discussion giving some further information on the design of his poppet valve gear which appears to have been very seriously considered for the compound version of the Lord Nelson design. He also noted that a Midland compound version had also been considered. The poppet valve gear had also been considered for an LMS Claughton.

H. Holcroft (823-6) made a long contribution which considered how Maunsell had used letters or circulars to communicate all important decisions to his staff. The Maunsell era could be divided into three periods - one of development mainly prior to the grouping, a brief period of intense activity in the 1920s which culminated in the introduction of the Schools class, and one of stagnation. This last was dominated by improvements in electricity distribution - the Grid and the mercury arc rectifier when electrification of the entire system other than that west of Basingstoke, which might have been handed over to the GWR was contemplated.
Holcroft considered that Cocks had "passed over the E.1 and D.1 classes in a few sentences": yet these light 4-4-0 rebuilds were perhaps the most skilfully designed of any of Maunsell's engines. Their duty was to work boat trains of 300 tons tare between Victoria and Dover or Folkstone, and also the Kent Coast trains, over heavy roads with numerous speed restrictions. In order to run on the former LCDR section of the road they were severely restricted in weight, and their adhesion was no more than about 35 tons. Every pound of unnecessary material was stripped from them, so that they had a high horse power output in relation to their weight. Their 19 in. cylinders could develop approximately 1,000 i.h.p. at round about 60 m.p.h. On occasion, they touched 70 m.p.h. on the level with a 300 ton train. Their high mean effective pressure resulted from a steam-port opening of about ½ in. at 25 cent. cut-off and a very free exhaust, due to the use of Stephenson gear with long travel. This compared with about 850 i.h.p. as given by the Walschaert gear at 25 per cent, on the N class with 19 in. cylinders having the same 10 in. piston valves, which had .the advantage of 20 lb. higher boiler pressure and a 2 in. longer stroke as compared with the E.1 and D.1 classes. Great skill was called for on the part of the enginemen to abstract so much power from a comparatively light machine, requiring delicate handling, whether in high-speed running over the Kentish Weald or on the long 1 in 80 banks on the Maidstone East arid Chatham lines.

In riding, these engines gave one the impression of buoyancy and liveliness, of spirited horses which had to be held on a tight rein. For five years they worked the Continental traffic without fail, conveying notabilities of all nations, and it was important that the motive power should not let the railway down. These little engines carried on until the civil engineer brought his road bridges up to full strength, so permitting the running of the King Arthur class in 1925. It was a great tribute to Maunsell that such a high standard was maintained by what was little more than a well- thought-out and cheaply-effected improvisation.
On pages 793 and 794, the Author gave convincing figures of increased efficiency for engine No. 449 with cranks altered to 135⁰, but made no reference to improved performance. In fact the driver scarcely knew his engine after the alteration; he was able to lift heavy trains out of Salisbury yard in a way not possible before. The boiler had a shallow firebox, and only a comparatively light fire could be carried, and therefore it would not stand up to heavy exhaust beats. With eight beats per revolution it was possible to work the engine much harder, especially as the more even turning moment improved the adhesion.

It should be noted that the preliminaries towards the shaping of the " Lord Nelson " class, as evidenced by the experiment with No. 449 began some time before the King Arthur " class was visualised. The " King Arthur " was notable in not being designed in the normal way but evolved through force of circumstances. The factors in a critical situation arising from shortage of suitable locomotive power were weighed up by Mr. Maunsell, and the logical solution was the amended N.15 design, embodying S.E. & C.R. practices. The performance was so successful that altogether 54 engines were built, and it was difficult to understand why the original 20 Urie engines were not rebuilt in line with them at the time. Some question of accountancy might have come in there.
Another anomaly had been the "charmed life " of the Drummond T. 14 class 4-6-0 4-cylinder engines. They had shallow fireboxes, and therefore could not be worked hard, the valve gear was poor, the bearing surfaces were inadequate, and they were condeoined to drag about enormous tenders holding nearly 6,000 gallons of water. Their operational efficiency was therefore very low, and their continued existence a locomotive " miracle." He thought that the running people must hide these relics away in the darkest recesses of their gloomy sheds as a reserve, bringing them out at peak holiday periods when anything that would turn a wheel was welcome.

R.C. Bond (826) remarked that as had been the case with the two previous Papers of the same kind, the information disclosed concerning locomotives which, though contemplated, had not been built, was the most interesting part of the Paper. One could indulge in much speculation as to what the position would have been on the Southern Railway had 4-6-2 locomotives been developed directly from the Lord Nelson class and if a 2-6-2 mixed traffic locomotive, of which a diagram was given in the Paper, had been put into service. The Paper brought out very clearly the manner in which Southern locomotive policy had been influenced by two over-riding considerations; firstly, the preponderance of passenger traffic and secondly the rapid growth of electrification. Fortunately Mr. Holcroft's depressing prophesy was not entirely fulfilled and there was still a very buoyant and original outlook on the steam locomotive side of the Southern Region. Concerning the indicator diagrams shown in Fig. 17 obtained from the conjugated valve gear, and the conclusions which the Author had drawn from them, it was perfectly true that the the indicator diagrams gave a better shape for the inside cyclinedr, but this was indicative of excess work. Noting the 225,000 mileages achieved between general repairs by the Urie 4-6-0s Bond queried the amount of work done in shops between general repairs.

O.S. Nock (827-8) noted the excellent performance of the Schools class on the Bournemouth line, and that they (unlike Lord Nelsons) responded well to a variety of driving techniques.

J.P. Maitland (828-9) stated that he felt that "no railway company had ever had such a good investment as the "King Arthur" class." Regarding the "Lord Nelson" class he had the whole of those engines under his juridiction for about two years and they were a most remarkable fleet. "Hot bearings were unknown. Some trouble was experienced from tubes leaking owing to the design of the tube-plate, but that detail was corrected in subsequent design." The locomotive fitted with 6 ft 3 in driving wheels cost about 7½% more to maintain, and No. 865, with altered crank settings, consumed about 7½% more coal, but this was due to poor driving. No. 857 fitted with a boiler with a combustion chamber was a problem in that incadescent ash collected in the chamber. Up until 1939 the Lord Nelson ran 200,000 miles per engine failure.
F.J.R. Watts (829-30) commented on the Holcroft derived gear and considered that there appeared to be a need for at least four more pin joints.
W.G. Thorley (830-1) was highly impressed by the way in which Southern drivers were confident that their water resources were sufficient despite the lack of water troughs. He wondered why the Southern companies had been averse to compounding and the jointing arrangements for the three-cylinder locomotives.
E.S. Cox (831-2) requested information on the ARLE 2-8-0 type and this was provided in the form of an outline diagram (s. el.) on p. 860. He also asked about the top feed with a neat of trays, and of the relative advantages of the three cylinder layout, and gave information on the 3-cylinder 2-6-4Ts versus the two-cylinder type: coal consumption 54 lb/mile as against 51.5 (1936-8); and mileage between repairs (46,465 vs 53,390 over the period 1945-7).
T.E. Chrimes (832-3) spoke about the difficulties experienced with early superheated locomotives (the I3 class): notably in keeping the tubes tight and in lubrication, but that superheating soon had a marked influence. He also spoke in admiration of Maunsell's work on the E1 and D1 classes and on the Schools class.
W.A. Tuplin (833) queried the layout of the valve gear on the Ashford 2-6-0 locomotives, and was informed that it had not been possible to discover the reasons but, contrary to the general opinion, this did not influence the valve events adversely and in the case under review it was done to simplify the general layout of the valve gear and the reverse. Concerning Tuplin's further question as to the sluggishness of the Urie class compared with the "King Arthur" class, and was told that it was not possible to find out whether it was ever considered to fit larger valves during the era of the Southern Railway. He thought it would be fair to assume, in view of the considerable amount of alteration which would be required, it was deemed preferable to leave this class of engine alone and concentrate on the later designs of the" King Arthur" class, except that some of the Urie engines were fitted with double ported exhaust valves.
J.I. Scott (833) referring to conjugated valve gears on pp. 781-2 cited Shields' review of locomotive valve gears (Paper No. 443) and its mention of Pickering's adoption of dashpots on his brief attempt to exploit conjugated gears and observed that the aircraft industry had enhanced the design of dashpots.
H. Holcroft wrote at length (834-44): Looking back to 1914 Maunsell became in time something more than a chief: he could be likened to a chieftain with his clansmen around him. Each one was assigned his particular duty; everyone knew what was expected of him in a clearly defined sphere. Under Maunsell's sway were keen and enterprising works managers, first-rate foremen and willing workers. On the business side of the department the clerical, costing, estimating and other such matters were conducted by men of ability: many of Maunsell's decisions in locomotive design were governed by this question of costs. Holcroft stressed that it was in the Ashford drawing office that the foundation of future design was laid before grouping took place. Afterwards a C.M.E. headquarters was established at Waterloo with a small personal staff attached. It was here that all subsequent design was initiated under Maunsell's personal supervision. The outlined schemes were passed on for development to the Eastleigh drawing office, which had been reinforced by the transfer of men from Ashford conversant with design so far standardised. All important drawings were submitted to Waterloo, where they were thoroughly vetted and amended if necessary, before presentation to the C.M.E. for his final approval and signature. Once that was done no departure was countenanced without formal permission. If, for instance, the shops wanted alterations in design to suit methods of manufacture, the matter was discussed on the spot and this was followed up by an exchange of letters defining and approving the amendment.

The organisation of material inspection at 'makers' works, establishment of chemical and physical laboratories for check testing and research all contributed towards reliability of engines in service. All engine failures were classified, analysed and their causes traced, weakness in detail design eliminated, with the result that mileage run per failure was ultimately raised to 200,000.

Relations with the locomotive running department were excellent, not only as regards the officers, but with all grades. At the sheds nothing was ever too much trouble to those in charge in the way of facilitating enquiries and furnishing information to the C.M.E.'s representatives.

Works managers have a weakness for regarding the locomotive running department as existing to supply them with a steady flow of general and classified repairs throughout the year. The running people on the other hand want a large instalment of repaired engines by Easter and every available passenger engine when the summer service comes into operation. Once that is over their ambition seems to be to evacuate all engines possible to shops. Maunsell had experience in both camps and could sympathise with their respective views. He appointed a liaison officer who toured the shops, sheds and offices and arranged current and long term shopping programmes which reconciled as far as possible any conflicting claims and so produced smooth working by his tact and practical knowledge of the job to be done.

Although standard locomotive types did not eventuate from the meetings of the A.R.L.E., they did have an important effect in breaking down the old isolationism that existed between railway companies. The C.M.E. departments of the four railway groups were mutually helpful and co-operative. Visits to the works at Swindon, Crewe, Derby, Doncaster, Darlington and elsewhere for the purpose of ascertaining latest practices in design and manufacture were frequent and were reciprocated at Eastleigh and Ashford. The G.W.R. amongst other things machined the first three-cylinder casting at Swindon, supplied such items as semi-plug piston valves and a jumper top for trial in N class, and they adjusted a set of N class coupled wheels on their wheel-balancing machine. It, was with Derby, however, that relations were closest and most informal. Before the " Royal Scot " was designed preliminary discussions took place at Waterloo in which the circumstances leading to the necessity for the new locomotive were revealed, the proposed design was outlined and information was supplied regarding the design, construction and performance of the Lord Nelson." Later on, the scheme for the Royal Scot " boiler was brought along for review before the final drawing was made.

Such instances of goodwill and mutual assistance as these illustrate the spirit of the times in which Maunsell's engines were produced and operated.

Maunsell's first design, the N class 2-6-0, is a very versatile engine and is truly " mixed traffic." It can move 1,000 ton freight trains on easily graded sections and 600 ton trains over gradients of 1 in 120. It is much used as a passenger engine in the West of England and has on occasions taken up the working of semi-fast main line trains, Salisbury to Exeter and Exeter to Plymouth. While it is not suitable for speeds much over 6o m.p.h., it maintains the requisite average speed by its good acceleration and hill climbing.

The one weak spot in an otherwise sound design is the steam and exhaust pipe arrangement connecting the cylinders with the interior of the smokebox. The vertical and horizontal plates staying the frames at the cylinders form a chamber in which the pipes are set up in sections. To give access the smokebox has a false bottom through which the pipes pass, and it is difficult to make the base plate really air and water tight. The cavity above the plate is bricked up to facilitate cleaning of the smokebox, but the brickwork is liable to crack in service. The consequence is that small air leaks occur and affect steaming, while when washing out the boiler water seeps through to the chamber carrying sulphur compounds from the ash. A humid and corrosive atmosphere is thereby created in which the pipes are liable to attack and so have to be renewed at intervals. To overcome the trouble some running sheds cement over the brick-work, and this leads to protests from the shops when they have to break up the equivalent of a road surface to get at the pipes.

The three-cylinder N.1 class is free from these troubles because the cylinders are bolted back to back and the steam and exhaust passages are integral with the castings. Better balancing enables these engines to be run at higher speeds, and N. 1 822 was officially timed at 79 m.p.h. on one occasion.

The question of two- versus three-cylinder engines of comparable construction has been raised in the discussion. The three-cylinder engine is more expensive to construct, uses rather more lubricating oil while as compared with two cylinders steam leakages past pistons and piston valves is increased, as also losses due to radiation of heat from a larger surface. To compensate for this, some special advantage obtained from the use of three cylinders must be specially marked to justify the construction.

The benefit obtained by the third cylinder can take many forms, depending on the working conditions. For instance, the Author has made references to the Tonhridge-Hastings line in connection with three-cylinder engines. To those not familiar with this section of the Southern lines, it should be explained that this road traverses difficult, hilly country and has numerous sharp curves, speed restrictions and heavy gradients. The tunnels as constructed by the old S.E.R. had brick linings which proved to be too' thin. To have re-bored the cross section for thicker brickwork would have been a great expense and the short-sighted decision was made to thicken up the lining internally, thus reducing the structure gauge. The matter was not of much consequence with the narrow rolling stock of the period, but at the present day special 8 ft. 6 in. wide corridor stock has to be provided, some types of freight vehicles are barred and so are two-cylinder engines with outside cylinders. By constructing a three-cylinder modification of the standard engine, dimension over cylinders was reduced sufficiently to permit of its running. Thus in this special case use of a third cylinder enabled more powerful engines to be put to work over this heavy road.

In the case of heavy shunting engines, as exemplified by the Z class, three cylinders confer a number of desirable features which the Author has touched upon. These engines are specially adapted to hump shunting where a very steady push is needed. In all-out conditions they will continue to move almost imperceptibly under circumstances where an equivalent two-cylinder engine would have stalled. It should be noted that no superheater is fitted on the Z class. On an engine used intermittently in shunting duties no appreciable benefit results from the normal fire tube superheater; in fact it is a drawback in operation unless the regulator is on the cylinder sfde of the header.

The K class passenger tanks were very comfortable in their cab arrangements and smooth riding, but the three-cylinder No. 890 was outstanding by reason of its lighter balancing, apart from its performance. It was remarkable in that it could be operated with wide open regulator and a very short cut-off. The two-cylinder engines were more suited to a partially opened regulator witn a cut-off of not less than 20 per cent. I made several trips on No. 890 with an evening train very smartly timed to do the 26.1 miles from Tonbridge to Ashford in 28 minutes start to stop. The load was about 270 tons tare and the running was against the gradient shown in Fig. 17 (page 781). The engine performed its task with ease under a full regulator and with the reverser only half a notch from and- gear. Steaming left nothing to be desired and the trip was as near perfection in locomotive work as could be. Whether the superiority of this engine over the two-cylinder variety was sufficiently marked to justify the additional expense is a debatable point because the Sevenoaks accident occurred before any systematic tests were put in hand, and the matter was not, followed up after conversion.
Turning to the 4-4-0 type, even G. J. Churchward could not make a succes of his outside cylinder design, the original " County " class. Add a third cylinder to such a design and the situation is transformed to such an extent that the type can then compete in performance with a larger and nominally more powerful 4-6-0. Instances have been cited in the discussion of the fine work done by the " Schools " class in this respect. All the difficulties with balancing of the reciprocating parts disappear, and the effect on bridges can be observed in the almost perfect parabolas recorded by the instruments used in investigating bridge stresses.

Coming to the more difficult cases where two- and three-cylinder engines directly compete in normal traffic, more particularly as regards freight service, the only ways in which the three-cylinder can obtain advantage lie in improved combustion efficiency, higher mechanical efficiency and in reduced wear and tear.

Without careful dynamometer car tests which include indicating, it is not possible to detect improvement in mechanical efficiency, though it is to be expected with reduced nosing, more even turning moment, steadier drawbar pull, reduced thrust on bearing surfaces and smaller reaction on the riding of the tender.

As regards wear and tear, a reduction is to be expected owing to better distribution of loading on bearings; smaller flange and tread wear should result from less nosing and lighter balance weights. If three-cylinder engines fail to show improved mileages over corresponding t\vo-cylinder between general repairs it may be that it is because there is a tendency to put them to work on the heavier duties. A ton-mile basis is required to assess the matter.

Turning to efficiency of combustion, this is influenced by the modified smokebox action through the same volume of steam being metered by the exhaust beats at shorter intervals per revolution with three cylinders instead of two.

In drawing out with a heavy freight train observation of the fire in the case of the two-cylinder shows a white flash with each exhaust beat followed by a rather longer interval in which the firebox is filled by the dull orange glow of incompletely burned hydrocarbon vapour. This alternation goes on until some speed is gained, when the light gradually becomes white but flickering. This dies out with further increase in speed a steady white glow fills the firebox as - combustion conditions improve.
In the case of the three-cylinder, the stages in combustion conditions are far less marked and occur earlier, and moreover the flickering dies out at a much lower speed. If the fire bed is watched through a dark blue glass, it will be seen that the agitation of the particles of fuel is less than occurs with a two-cylinder engine at the same speed. When the r.p.m. reach 100 or so the difference between the two types is small, and combustion conditions tend to become equal. Fuel saving in a three-cylinder engine is therefore to be looked for in slow and heavy haulage in freight service or passenger service over steep gradients
Although the Author had not been able to produce any figures, Holcroft able by reference to his notebook gave particulars of coal tests carried out between two- and three-cylinder engines on freight service over the same road and using the same grade of Kent coal (Chislet), the

The coal returns supplied by the outdoor locomotive department for N class engines stationed at Ashford and working in the same link, prior to the first tests with No. 822 were as follows

The result in favour of the three-cylinder engine was a saving of 11.8 per cent, of the coal consumed by the other engines in the link.
The economy in fuel by No. 822 is not so apparent in' the overall results of the coal tests, but if the results of the individual test runs are plotted as a graph (see Fig. X) they are very striking.
All the trials were conducted during the summer months so that weather had little effect on them. Disregarding the abnormally large consumption of Engine 'No. 8i8, lines are drawn through the scattered points plotted, one for No. 822 and the other br INOS. 817, 821 and 825. These lines show, under the conditions of the test, that with a load of 47 wagons the coal consumption of two- and three-cylinder engines are equal.
For a smaller number of wagons, advantage lies with the two- cylinder, but as the load increases above 47, so does the relative economy of the three-cylinder engine. The deduction is obvious: give three-cylinder engines the maximum of hard work and they will repay their cost.

During the course of the trials there was little difference in superheat between the two types but the average of 8~ readings of the water gauge for smokebox vacuum was 1.9 in. with ~ wagons for No. 822, and for 6~ readings with No. 825 pulling 48 wagons it was 2.6 in. As the conditions in each case were strictly comparable and the ashpan dampers well open on both engines, it is apparent that the lower smokebox vacuum was due to a smaller resistance at the fuel bed, that is to say the three-cylinder could be worked with a thinner fire so that combustion conditions were more favourable.

When it is considered that the coal consumed in the trials included that for lighting up, standing in steam and light running, the saving by the three-cylinder is all the more remarkable because the opportunity for fuel saving only occurs when the engine is pulling hard in traffic.

The conjugated valve gear used in three-cylinder engines has been criticised, mainly on account of over-travel of the middle valve which has been known to occur on infrequent occasions. The critics are, however, "barking up the wrong tree.." The conjugated faithfully reproduces the combined movement of the two primary gears, and if they have any faults it reveals those faults in magnified form. If there is over-travel of the middle valve it arises from over-travel of the primary gears.

In the shops it is customary to take readings of the valve motion while the locomotive wheels are slowly revolved. It is too often Assumed that precisely the same results are produced at high speed in service.

If the matter is considered, it will be found that two Walschaert gears are anchored at five points, at the reverser and at each cross- head and return crank. The structure includes a none-too-rigid reversing rod, a reversing shaft subject to torsion and deflection and off-sets at overhung return cranks and other non-alignment. Thus we have an elastic set-up for the purpose of reciprocating a pair of valves with a harmonic motion, the principal effort required being to overcome the inertia of the valve at each end of their travel. Such a system will have a natural period of oscillation and at some particular speed over-travel will reach a maximum, being accentuated if the motion. is put into full gear, as when steam is shut off at high speed.

In the case of a conjugation being added for operating the valve of the middle cylinder in a three-cylinder engine, there are three light valves to swing about instead of two heavy ones, so the work expended is not much affected. The phase is altered, because the impulses occur at intervals of 60º instead of 90º. The remedy for over-travel or other irregularities of the third valve is to make the primary gears as rigid as possible, consistent with their weight. The deflection of the levers of the conjugation is small but could be further reduced by using a deep I section of aluminium alloy. It does not seem to be grasped that corrections can be applied by departing from the rigid 2 :1 ratio of the levers; the valve events can be advanced or retarded and also the travel can be adjusted by slight alterations in the proportions of the levers. In other words, the possibilities of the conjugated gears have not been fully explored: they can take many forms, there is no need to adhere to those forms so far employed.

The early Maunsell engines had top feed arranged in the dome in the form of shallow trays constructed as a helix and known to shops as " the helter-skelter lighthouse." Clack boxes fixed on each side of the dome discharged feed into the trays through short internal pipes. It is doubtful whether the feed obligingly took the course anticipated but at least the overflow from the trays would produce a shower of water through the steam space. When the dome cover was removed for periodical examination of the trays, they would usually be found to he heavily encrusted with brittle scale, necessitating their removal. Immediately below the trays was the nest of tubes and great care was necessary to prevent large pieces of scale from falling and lodging there. The removal of the trays in sections for de-scaling was a tricky and lengthy operation. After a time doubts arose as to whether the trays effected any improvement in the condition of boilers so the D rebuilds had no top feed and reverted to clack boxes located on sides of barrel. These boilers were identical with those of the E rebuilds which had top feed, and a direct comparison was possible as they used the same supplies of feed water. When D class first came in for general repairs the internal condition of the boilers was carefully compared with the condition of the E boilers and as top feed appeared to confer no appreciable benefit it was removed from all classes and short internal pipes extending to the water line were fixed to the delivery to divert feed water towards the sides of the barrel. The location of clack boxes on the dome was continued in new construction as it was thought that this method of introducing the feed was as good as any, if the short internal pipes were removed periodically for de-scaling.

The 22 engines of the L class, No. Sio (N class) and No. 790 (K class) and the 11 E class engines had mechanical lubricators for the cylinders. The remainder of the N and K classes and the D class were fitted with sight feed lubricators having a separate condensing chamber situated outside the weather board of the cab. Direct comparison was again possible and after lengthy experience with both types it was decided that the sight feed gave rather better results all round. It was cheaper to manufacture and maintain, consumption of oil was no greater and it was considered that surfaces were better and more evenly lubricated, and that there was, on the whole, less carbon deposit. A gradual change over therefore occurred until the mechanical lubricators were finally eliminated.

D and E classes compared: Observation on the running in service with the same train loads and timing over the same road showed that the E class with the smaller wheels was almost invariably a minute or so faster on the long uphill stretches but that the reverse was the case on the level. This was, however, because the D class had to make up time on the level to compensate for slower speed up-hill. Deposits taken from the blast pipes of the two classes when examined in the chemical laboratory were found to be entirely different as a result of the method of lubrication and point from which steam was drawn for the cylinders. "My notes on this have been mislaid" but as far as I can remember the deposit in the D blast pipe was mainly composed of boiler solids, while that in the E was mostly charred oil and smokebox dust.

Page 841 The Author did not mention the pulverised coal trials carried out on No. A629 U class in 1930. The success of the German Railways with pulverised brown coal prompted Maunsell to try the use of pulverised black coal. Accordingly, AEG of Berlin installed apparatus on this U class engine, and a pulverising plant was installed at Easthourne shed. After many vicissitudes, regular services were run between Eastbourne and London with passenger trains. Although pulverising put up the cost of the fuel delivered on the tender, the system had some of the advantages of liquid fuel firing, and it was hoped that enough fuel would be saved in reduced stand-by losses and lighting-up and by improved combustion to justify the expense. Careful costing, however, showed no overall gain over grate fired U class engines on the same working and the apparatus was removed. Subsequent investigation in Germany carried out on stationary locomotive boilers demonstrated that black coal would have to be pulverised to a much finer degree than brown coal to burn completely in conventional locomotive fire boxes, and the conclusion was reached that the degree of fineness needed rendered the use of this fuel uneconomic.

This contact with German engineers had an important sequel. They were full of enthusiasm for the solid-headed piston valve with plain rings which had come into use in Germany, and they brought over drawings. This led to their general introduction on the Southern Railway. A snag was struck, however, when these valves were applied to " Schools " class engines working on the Eastern Section.. Certain trains to Cannon Street or Charing Cross have to make a stop at No. 7 platform, London Bridge, which is on a curve and has an up gradient of 1 in 100. To make matters worse, there are catchpoints immediately in the rear of long trains, so that setting back more than a few yards is prohibited. Great difficulty began to be found with the Schools " class in starting their trains and an investigation was made. These engines, in common with the Nelson " class had a lead of ¼ in., but as the solid heads of the valves had a small clearance in the liners, some pre-admission in excess of lead steam started as far back as the first ring, when it passed the port edge. The amount of steam leak was insignificant when on the move, but on starting a heavy load enough leakage occurred to cause a negative turning moment and so seriously affect the tractive effort.

The remedy adopted for this state of affairs was to reduce the lead and transfer the point of cut-off from the edge of the head to the first ring, turning down the diameter of the head in advance of the ring to expose the side of the ring to steam.. This reduction in diameter of the head can only be small in amount, otherwise the reduced bearing surface of the ring in its groove leads to excessive groove wear.

While this alteration ameliorated conditions at starting in the case of the " Schools " class, it was applied also to the " Nelson " class in accordance with the policy of standardisation of parts. In my opinion this was most unfortunate and quite uncalled for from the performance aspect, and I attribute to this the blight which seemed to descend on the " Nelsons " after their earlier brilliance. The alteration did not matter so much to the " Schools," which are customarily worked at a 25 per cent, cut-off and part regulator, but the Western working of the " Nelsons " with full regulator and short cut-off was another matter altogether. Not only was the area of opening to steam restricted by the projecting edge of the head beyond, the first ring but lead steam was reduced as well, so that port opening was much smaller than before, with the same travel. This state of affairs remained until the front end was modified by Mr. Bulleid in recent years.

The solid-head piston valve has been commended for its cheapness in first cost and for maintaining steam tightness for a long period in service. There is another side to this picture, however, and I know something of the difficulty which occasionally arises in withdrawing these valves in the running shed if badly carbonised or if the rings are seized up. In the worst cases it may take days of patient struggle to force the back head through' the front liner, and in the end perhaps a hydraulic jack or improvised battering ram has to be used on the spindle, with the result that the front liner is sometimes forced out with the head stuck in it.

The Schmidt type valve head with wide ring as once made by Ashford works gave excellent results, although more expensive in first cost. After grouping it was adopted as standard and replaced the South-Western type head of composite construction. The Eastleigh-made Schmidt ring never gave the results anticipated, although Ashford men were sent to compare notes with the Eastleigh fitting bench. It seemed to .me that something valuable was lost when Ashford iron foundry was closed down under rationalisation after grouping and the old craftsmen retired. Their mixture and methods produced metal for rings which acquired a glass-like hardness and. resisted wear. All casting was concentrated on the mechanised foundry at Eastleigh after this event.

With the failure of the Schmidt ring to come up to expectations in extended use, the introduction of the' solid-headed valve with plain rings was hailed by the shops as a cheap and easy solution of their difficulties: but it may well be that the running department hold other views.

The Author shows a proposed. four-cylinder 4-6-2 locomotive in Fig. 32 and a three-cylinder 2-6-2 in Fig. 33; Intermediate between these designs was the proposal for a three-cylinder 4-6-2 with 6 ft. 7 in. wheels and 20 in. x 28 in. cylinders. After the success of the " Schools " class the view was definitely held at Waterloo that three cylinders were sdperior to four and that as long as three cylinders would suffice in a design there was no point in adopting four. The Author has apparently not located this scheme and it may be that an outline in pencil was discussed informally with the civil engineer in an exploratory way to see how far he could. stretch any concessions on weight restriction. The scheme was not proceeded with but was amended to the 2-6-2 type which therefore represents the final attempt to find a solution in Maunsell's day. It will be noted that the coupled wheel base was increased and a bogie tender reverted to in order to spread the axle loads and so decrease loading per foot run.

The use of smaller driving wheels in an engine of the " Nelson " class was the sequel to an earlier trial. It was thought that a 6 ft. wheel engine would be more suited to the working of the Dover boat trains and an engine of the H.15 class with taper boiler was brought over for trial; it was identical with the Urie N.15 class except for smaller wheels. This engine failed to come up to the " King Arthur " class at any section of the road, and' it was clear that the valve gear limited its performance, as the boiler power was equal in each case.

It should be explained that an increase of 5 m.p.h. to the slower average speed over the uphill sections will yield more saving in time on the journey than the same addition to the faster running on the more level portions. For the Eastern and Central sections 70 mph maximum is sufficient, and the K tanks had shown that a 6 ft. wheel would do the job and it had advantages in acceleration and hill climbing. On the other hand 80 m.p.h. is frequently attained and exceeded on the Western section, calling for a larger diameter, hence the adhesion to a 6 ft. 7 in. wheel. The Author dismisses the trials with Engine No. 816 (page 804) in a few sentences, and yet it was perhaps the most remarkable experiment ever carried out on a locomotive. The theoretically impossible task of returning the whole of the exhaust steam to the boiler as feed at about 225⁰F, was accomplished by means of apparatus no more complicated than some feed heaters which return less than io per cent, of the exhaust. Utilisation of the whole of the exhaust necessitated the fitting of a fan for creating induced draught, and it was the inadequacy of this and mechanical troubles in driving it which mainly led to the shelving of the experiment. No information regarding the trial was allowed to leak out at the time in case of a resumption, but an account has since been published giving the main facts (see " The Engineer," 6th, 13th and 20th September 1946).

AUTHOR'S REPLY

In reply the Author said he is in agreement with the President's remarks that the Southern Railway had been fortunate in having men of ability. With regard to his mentioning the present chief mechanical engineer and the " Merchant Navy " class locomotives, this was outside the scope of the' Paper and he was not prepared to discuss these locomotives at the present time.

He agreed it would have been a most interesting experiment if a " Lord Nelson " class locomotive had been fitted with the poppet valves, and shared Mr. E. C. Poultney's regret that the experiment had never been carried out.

With regard to the question of saving coal, when the cranks were set at 1350 instead of 8o⁰, this was probably due to the more even type of exhaust by the placing of the cranks and was borne out in the experimental engine No. ~ described in the Paper.

Mr. R. C. Bond's remarks with regard to the progress of the - Southern Railway were noted and the Author agreed that the position was never allowed to become so depressing as had been outlined due in a large measure to the enterprise of the chief mechanical engineer and his staff. Indeed they were so successful as to satisfy the enthusiasts of the traffic department who were clamouring for wholesale electrification.

In connection with the indicator diagram shown in Fig. 17, it was not intended to give the impression that the conjugated valve gear as fitted was all that could be desired, but rather that the shape of the diagram obtained from the middle cylinder was much nearer to the theoretical diagram and therefore should be the aim when designing valve gears.

In connection with the Urie engines which had averaged 225,000 miles between general repairs, it was difficult to ascertain at the present time the extent of the repairs which had been carried out in the workshop when it was necessary to give attention to wheels, etc., but a sample of a few engines and the details which were repaired have been abstracted from the papers still available and are given in the Appendix A.