Chapelon and related Matters

John Bellwood late of NRMY held the smaller wheeled A2 as best british.
Another source somewhere stated that the fifty square feet grate LNER boilers lost the advantage of better efficiency when working hard to loosing it again waiting for action.
Bigger wheels and grates are not best always.
Mr Witte of DB proposed to blank 20% of grate area of in the 8000 leftover boilers from baureihe 50/52 after WW2 when lignite burning was not longer required.

 

I believe it's to do with weight distribution and balance on lightweight track. The Italians made essentially three designs like that - Gr 625 (as above), Gr 640 (broadly the same but with larger driving wheels) and Gr 745, a 2-8-0. Sadly none of the latter survive, but plenty of the others do.





They even rebuilt 625s with Franco-Crosti boilers and Caprotti valve gear, but still with the inside connecting rods - not the prettiest class!

 

"Look Mr Picasso, we gave you a fair trial in the drawing office, but we really don't think you are cut out for railway work. Perhaps it is time you left our service and spent more time trying to sell the paintings I believe you dabble with in the evenings?"

Tom

 

Given that the French Atlantics had right angle bends in their steam pipes and slide valves for their LP cylinders but were pretty the equal of Churchward's Atlantics, it is surprising, to say the least, that Churchward failed to further investigate compounding.

 

Arguably the weirdest feature of that beast is the extra coupling rod to drive the Caprotti gear.

 

Thank you for that. They say a picture is worth a thousand words and I reckon that applies here, even if my understanding is still incomplete.

 

More an eccentric rod than a coupling rod. The furthest end is attached to a crank. The 623 Class had a Franco-Crosti preheater on both sides.

 

One feature of the GW arrangement is the relatively short link connecting the rocker to the outside valve. As the rocker moves to the back of its stroke the angularity of this link increases, as shown in Don Ashton's diagram but without further mention, thus further reducing the travel of the outside valve. Thus in general terms, the outside valve moves faster than the inside at the front of its stroke because of of the greater length of the outside arm of the rocker, and more slowly at the back of its stroke because of the angularity of this link and also because the rocker is moving away from the direct line.

 

Thank you for posting this. I assume the numbers are the amounts in inches for full travel, although I guess it does not really matter. This is a similar analysis to working out the backset for the expansion link as that is also informed by the fact that as you move away from the "vertical", the "horizontal" displacement reduces for the same angular swing. I wonder why the movement is asymmetric for the inside valve? It could simply be due to the distortion which usually shows up at maximum travel (see below re 44767).

I think Don Ashton did some tremendous work but he could be somewhat cryptic in his writing and his zeal for perfect valve tables while commendable and indeed the logical target for his analytic work, seems to have little relevance to the real world, witness the issues already discussed and the poor indicator diagrams for Gresley gear which did not seem to hamper the locos much. The best valve event tables I have seen are for 44767 (to compare with the King diagram, there is a difference in front and back port openings of 3/16" / 0.1875" at 80% cut-off). There are rumours that Ivatt wanted to show the GWR how to "do" Stephensons but the LMS valve events were otherwise nothing special although the two cylinder locos were usually fairly symmetrical in the normal running cut offs. It took them several attempts to sort out the 3 cylinder 4-6-0s which ended up as not very long lap locos, but with 5/16" lead ensuring they could get plenty of steam into the cylinders.

 

? Imho this would be true if there was no short link on the inside valve rod. Since there is a short link the above argument is also valid for the inside cylinder which has a reverse movement, slowing at one end and a faster movement at the other end of
the rocker seems contradictory.

 

No short link on the inside valves, the inside valve is attached to a tail rod that is driven directly off the radius rod/combination lever, the rocker arm is driven off the tail rod, so no reduction in travel at either end on the inside valves.

 

Yes, correct! On a second look it appeared to be connected with a sliding block!

 

Finally got a copy of Rogers’ Chapelon book. Now starting my read. Lots to do already!

 

After a distinguished career in an artillery officer during WW1, during which a new system of fire control was devised by Chapelon and embodied in Artillery Regulations, Chapelon graduated from the Ecole Centrale in 1921. He then joined the PLM Rolling Stock and Motive Power section of the PLM as a probationer. His instructor, a M. Tribolet at the Lyon-Mouche depot, arranged that Chapelon should receive practical experience of most of the classes of locomotives in use. One aspect that left Chapelon most unimpressed during his time at the PLM was the way that locomotives were driven. This particularly applied to the 6100 Class four cylinder simple Pacifics; they were normally driven at 50% cut-off with a partly opened regulator. This meant that a 12 bar boiler pressure was reduced to about 4 bar at the cylinders. This was in contravention of all that Chapelon had learnt during his time at the Ecole Centrale and the principles proposed by the founder of the science of thermodynamics, the eminent Frenchman Nicholas Sadi Carnot (1796-1832).

Carnot spent nearly all of his working life as a military engineer. His life in the French Army became increasingly difficult because his father had been Napolean’s Minister of the Interior. Eventually he took up other pursuits, attending lectures on physics and chemistry, but remained on-call at two-thirds pay. He became interested in understanding the limitation to improving the performance of steam engines. Although there was some intuitive understanding of the workings of a steam engine, scientific understanding of their operation was almost non-existent. Carnot set out to determine how the greatest amount of work could be obtained from a given amount of heat. In so doing, he invented the idea of the Carnot Cycle, being an idealized concept from which the maximum theoretical efficiency of any heat engine can be determined.

M. Tribolet was greatly impressed with Chapelon’s abilities but the latter became disenchanted, realising that there was little opportunity for promotion and little scope for a disciple of Carnot. In 1924 Chapelon left locomotive work altogether and joined the Société Industrielle Telephones and soon became Assistant Manager. However his interest in locomotives had not diminished. His professor of thermodynamics at the Ecole Centrale was M. Louis Lacoin, who had formed a high opinion of Chapelon’s technical ability and was well acquainted with his distinguished war service as an officer of the Heavy Artillery. Louis Lacoin was a cousin of M. Maurice Lacoin, Engineer-in-Chief of Rolling Stoch and Motive Power of the Paris-Orleans Railway; thanks to the former’s help, Chapelon was appointed in January 1925 to the Research and Development Section of the PO.

It was an appointment to which Chapelon was particularly well suited. He had studied steam locomotives intensively for years and had read widely on the subject. He had, as a result, formed his own opinion on the improvements that were required before he had even joined the PLM. Under the direction of M. Paul Billet, Chapelon was put to work in connection with trials and charged particularly with the improvement of the exhaust system of the Pacifics and other locomotives. The PO was not satisfied with the performance of its compound Pacifics. Indicator diagrams of cylinder performance revealed throttling of the steam at admission to the cylinders, high back pressure at the exhaust and a considerable drop in pressure in the intermediate receiver between the high and low pressure cylinders. The normal solution would have been to build new locomotives of modern design but this would have been expensive and because the decision had already been taken to electrify the PO main lines, the new locomotives would have been expected to have a short life. It was therefore decided to see what could be done to improve the existing stock.

Chapelon’s first task at the PO was to optimise the draughting of the company’s locomotives. This involves the mixing the exhausts from the fire and cylinders in such a way as to maximise the former and minimise the back pressure of the latter. Chapelon’s Kylchap exhaust was derived from an exhaust invented by the Finnish engineer Kylälä in 1919. This was designed to produce an adequate draught over the whole smokebox and to mix the gases so well with the exhaust steam that the mixture might be expelled with the minimum of effort. Trials in 1926 on the PO compound Pacifics of the 3500 and 4500 Classes, the simple expansion Pacifics of the 3591 Class and other locomotives showed a significant improvement in steaming and reduction in back pressure. On Pacific 4597 Carnot principles were verified by the locomotive’s performance, for on trial with the regulator fully open there was a pressure of 224psig, instead of the previous pressure of 112psig and this was with cut-offs of 40% HP/60% LP, instead of the previous 60% HP/70% LP. In addition the locomotive showed an economy in coal consumption of 16%.

It was apparent to Chapelon that even better results could be achieved by increasing the cross- sectional area of the steam passages and, to be on the safe side, he considered it advisable to double it. At the same time he would eliminate the sharp bends in the steam pipes and increase the volume of the steam chests. By these measures he would reduce the losses from throttling which were occurring in the steam circuit from the regulator to the HP and LP cylinders.

Thomas Crampton was the first locomotive engineer to appreciate the importance of large steam passages. His first two locomotives were built in 1846. In 1897 the Nord Company carried out experiments with Atlantic express locomotive No. 2158, as a result of which steam circuit proportions were established which included large steam chests to regulate the flow of steam between the boiler and cylinders and so lessen the losses from fluctuations in pressure due to throttling.

Chapelon’s proposed improvement to the exhaust and steam circuits should, with a 40% cut-off in the HP cylinders and at a speed of 60mph, result in an increase in the indicated horse power of 1,936 (unmodified locomotive) to 2,496, that is, a gain of 25% in both power and efficiency. However, further improvements were to come. Chapelon had measured the temperature in the intermediate receiver of one of the superheated locomotives and had found that the degree of superheat in the LP cylinders was practically nil. If condensation could be eliminated in the LP cylinders, efficiency could be increased. The existing superheat was 300degC but any gain was almost entirely in the HP cylinders. Making the LP cylinders contribute their share would contribute another 10%. Chapelon estimated that to obtain such a result, it would be necessary to increase the superheat by an additional 100degC.

From an unmodified 3500 superheated Pacific running at 60mph with an average ihp of 1,850, the improvements in the steam circuit would raise the ihp by 20% to 2,200. The increase in superheat would gain another 10% to 2400. On top of that the improvement in draughting with the Kylchap exhaust should increase the boiler output by a further 25% above its previous limit and so raise the ihp to 3,000.

As a result of Chapelon’s argument and in the light of the results obtained in the tests, it was decided in November 1926 to rebuild one of the 3500 Pacifics completely. The first locomotive of the class to be rebuilt was 3566 (about the worst of the class, so bad that it was nicknamed Cholera). It emerged from Tours works in November 1929. Completion was delayed by the death of M. Billet and the retirement of M. Lacoin, who was succeeded by M. de Boysson. Nobody believed in Chapelon’s locomotive, which was considered to lie in the realms of fantasy rather than to merit consideration as a practical project. Ultimately Chapelon had to write a note to M. do Boysson, drawing his attention to the interest shown when the rebuilding plan had been initially approved. As a result, M. de Boysson requested a full explanation of the whole project and at length agreed to its continuation.

In addition to the improvement mentioned above, the locomotive also incorporated an AGFI feed water heater and a Nicholson thermic syphon, the latter to improve water circulation. A double Kylchap exhaust was fitted with two chimneys. No. 3566 ran her first trials on the 19th November 1929. The results obtained were remarkable. Chapelon’s calculations were exactly fulfilled, for No. 3566 developed about 3,000ihp at a speed between 75 and 80mph hauling a heavy train with an economy over unrebuilt locomotives of 25% at normal power outputs.

 

COPIED FROM "STEAM SPEED RECORDS" THREAD:

The improved performance and power of the PO 3500 class were so great that the first reports of Chapelon's achievement reportedly generated widespread disbelief. According to Gerard Vuillet, even the PO's own engine crew were sceptical until they got their hands on the rebuilt machines and could confirm abilities to their own satisfaction. An analogous situation would be that of LMS and LNER engineers in the 1920s, when they first heard of the low coal consumption of the GWR Castles. It is arguable that Chapelon's first great success was also his greatest - with all 89 of the original 3500-class engines rebuilt and 28 more built new for the Nord.

Thanks for posting these figures - I don't think that I have seen them before. I assume that they re meant as generalized approximations of what may be possible, but based on the actual results for the 3500s? Actual results will of course vary, depending on how good or bad were the original design features of a loco type being considered for improvement. Another factor will be whether a rebuilt engine will be able to cope with greatly increased power output. The 3500-class were of robust construction, but that will not always be the case. In the wake of the revolution on the PO, the other French railways moved to improve their express engine types, but to varying degrees. In particular, the PLM was considered to be half-hearted in applying Chapelon, but one of its issues may have been the lighter frames of PLM types, designed to suit main lines following sinuous river valleys, and which might not have supported an overdose of power.

In the case of the Kylchap exhaust, adoption may have been limited by royalty payment concerns. As far as I know, Kylchap use was limited to the PO, Midi and Etat railways in France, and elsewhere to LNER, Spain and Czechoslovakia. Some other railways of course adopted the Lemaitre or Giesl systems, or more basic forms of multi-jet exhaust. But the German 05-002 topped 200 km/hr with just a plain single exhaust, so a sophisticated exhaust was clearly not essential for high-speed.

Although France had large numbers of simple-expansion 2-8-0s and 2-10-0s, it also had substantial numbers of 4-cylinder compounds. For example, the 280-strong Nord class 4.061 or the 200-plus (exact numbers vary between sources) of the PLM 140E class. At the end of WW2, the SNCF ordered a further batch of 150P compounds to bring the class total to 115.

It is, however, true that much heavy freight in the post WW2 era was entrusted to the American-built 141R class. It is also true that there was a greater tendency to use simple expansion on locos for short-distance passenger traffic, for example the large Nord 2-8-2Ts of 1932.

 

Indeed. There are a more than a few on this forum who share this disbelief! Perhaps a case of the usual English parochialism?

They would have been approximations whose totals were confirmed by subsequent testing. I believe it was the Etat that also built locomotives with thinner frames. The 241P also suffered from too thin frames; Chapelon's intervention, at the invitation of Schneider, was too late to alter the frames, which caused continual problems with hot axleboxes.

Don't forget that nationalisation of the railways in France was in 1938, so that might have limited application by some companies, at least statistically.

The 150Ps, because of their large section steam passages, could run freely despite their relatively small wheels. They were occasionally used for express trains, which they could haul at up to 65mph. Of the PLM 140 classes, some were simple expansion, some were compound but unsuperheated (which probably removed any benefit of compounding) and some had larger wheels for fast freight.

 

An updated list of locomotives built, rebuilt or heavily influenced by Chapelon.

Built/Rebuilt/Modified:

231 Est 051 to 073 23
231 PO 3701 to 3731 31
231 Nord 3.1111 to 3.1130 and 3.1171 to 3.1198 48 (Second batch new)
231 PO 3801 to 3806 and 3821 to 3829 15
140 PO 5001 to 5067 67
231 PO 3601 to 3650 50
141P SNCF 318 (New)
240A/240P 37
160A1 1
242A1 1
142 GELSA 66
242 GELSA 24
141R 165

Influenced:

150 Nord 5.1201 to 5.1230 30
150P 115
État 231-501 to 231-783 215
231 PLM 2 to 86 84
231G entre 2 to 285 284
140J 170
Est 230K 177
Est 231B 40
241 Est 241002 to 241041 and 241 État 241-001 to 241-049 90
232R 3
232S 4
232U 1
241P 35
Czech Series 556 510

Making a total of 2,604. This is a conservative (and most likely incomplete) total. For example, the official total for 141R modified is 165 but the French believe that many more than this were modified but as I don't have a total number, I have used the official number. Stalin even had a locomotive built by the Czechs to Chapelon methods; it's not clear whether he ever travelled behind it.

 

You've clearly been doing a lot of work on your list. I must confess that I find it very difficult to get my head around the large variety of French Pacific classes, including the considerable number of additional variants introduced by the rebuilding programmes of the 1930s and 1940s. We could do with a meaty RCTS volume on the subject.

I assume that your entry "231 PLM 2 to 86" refers to the class that ended up as 231K (having been rebuilt from the type designated as 231C on the PLM)? There were also 30 of the 231H, a variant of the 231G fitted with a higher pressure (20 bar instead of 16 bar) boiler. We also need to look at the PLM's large fleet of 2-8-2s, at least some of which were rebuilt with Chapelon features. "Steamlocomotive.com" quotes a figure of 444 rebuilds (to 141E) from a class total of 680 for the 141C class.

(As an aside, was the 680-strong PLM 141C the world's most numerous class of 4-cylinder steam locomotive?)

There were certainly further "odds and ends" for application of Chapelon features, such as the fitting of Kylchap exhausts to the 20 Midi 3101-class 2-cylinder simple Pacifics. And of course, there is the Chapelon influence on certain LNER types, including those given Kylchap exhausts. Moving to Spain, about half of the 242-strong RENFE 141F class appear from photos to have been built in the 1950s with Kylchap exhausts. Curious that the Wikipedia page on "Kylchap" has a photo of the fitment on a Spanish loco, but no mention of Spain in the text!

Outside of France, Czechoslovakia was possibly the place where Chapelon's precepts were most enthusiastically adopted, applied across their range of post-war classes. They even built a trio of 3-cylinder compound 4-8-2s, following the pattern of Chapelon's 242A1. But even Chapelon was unable to persuade railways outside France to re-adopt compounding for general use.

I don't think we can ever expect to precisely quantify the extent of Chapelon's influence. Many engineers would have been encouraged to adopt multi-jet exhausts, but not always of the Kylchap variety. Even more will have been encouraged to adopt more generously proportioned steam passages and "internal streamlining", but this will only be known where an engineer went on record to acknowledge such influence.

 

Yes, the 231 PLM 2 to 86 was the future 231K. The 5-141C's were built too early to be considered to be influenced by Chapelon, but the 141D/E/F (all 141Cs modified after nationalisation) could be.

141 D 124
141 E 254
141 F 195

The PLM 231H class, rebuilt from 231A and 231B after nationalisation, would certainly have been influenced by Chapelon.

231 H 30

For the RENFE 141Fs, 240 had Kylchap exhausts and two had Giesl.

RENFE 141 F 240

I have also added the LNER locomotives fitted with Kylchap exhausts.

LNER A4 35
LNER A3 78
LNER V2 8

I have no information on the Midi 3101 Class.

That gives a total of 3,598.

 

Don’t forget the P2, a number of Thompson/Peppercorn A2 and the Peppercorn A1 classes.