George Jackson Churchward locos of the GWR: 103 President & 104 Alliance GWR/De Glehn 4-4-2 Compound

103 President, and 104 Alliance, were locomotives of the Great Western Railway. George Jackson Churchward, Chief Mechanical Engineer of the Great Western Railway, was given authority to purchase three French de Glehn-du Bousquet four-cylinder compound locomotives, in order to evaluate the benefits of compounding. The first locomotive, no.102 La France, was delivered in 1903. Two further locomotives, nos. 103 and 104, were purchased in 1905. These were similar to the Paris-Orleans Railway's 3001 class, and slightly larger than 102. As with no. 102, these were built by Société Alsacienne de Constructions Mécaniques. They had two high-pressure cylinders fitted between the frames, and two low-pressure cylinders outside. The high-pressure cylinders drove the front driving wheels while the low-pressure cylinders drove the rear driving wheels. An external steam pipe was mounted just in front of the dome, looking rather similar in appearance to a top feed. In 1907 No. 104 was fitted with an unsuperheated Swindon No. 1 boiler, President herself being similarly reboilered in February 1910 and receiving a superheated boiler in January 1914. In 1926, the three locomotives were based at Oxford shed. In practice, they did not provide any significant improvement in either performance or economy compared to No 171 Albion, Churchward's prototype 4-6-0, which was converted to a 4-4-2 specifically for comparison with the French locomotives.




https://railway-photography.smugmug.com/GWRSteam-1/Churchward-Locomotives/Churchward-Tender-Locomotives/De-Glehn-Compound-442-locomotives/103-President-104-Alliance/i-BC6CcJd


103 President & 104 Alliance - Railway-Photography
railway-photography.smugmug.com




https://railway-photography.smugmug.com/GWRSteam-1/Churchward-Locomotives/Churchward-Tender-Locomotives/De-Glehn-Compound-442-locomotives/103-President-104-Alliance/i-M3ghSmN


104 Alliance at the end of the parcels Platform Paddington, with the "Great western & Metropolitan" (Hammersmith & City Railway) station sign for Bishops Road station visible through the bridge (Bishops Bridge Road).
railway-photography.smugmug.com

Neil

 

The position of that external steam pipe and the size of the outside cylinders suggest that it was the other way round.

 

RCTS says the high pressure cylinders were outside.

 

It's a de Glehn. Pay close attention to the size of the outside steam pipe. It has to feed not only the high pressure cylinder but also the low pressure. In other words the pipe should be sized to be able to feed a single cylinder equivalent in volume to the combined volume of the high pressure cylinder plus the volume of the associated low pressure cylinder.
Look at the original layout, the first picture. True the engine is saturated but the directness of the feed to the cylinders should have indicated something to someone. The steam pipes on the rebuild are nowhere near as direct. The rebuild is also superheated which would have been good if only some consideration had been given to trying to get the steam to flow adequately. I hope that no one ever expected this engine to perform well.
If you want to know what the h.p. steam pipes should have looked like there are pictures of those fitted to 231E41 (and these could have been of larger bore). This is an extreme later though readily accessible example so look for the designs of du Bousquet and Edward Sauvage. Or von Borries and there are many others.

 

Surely steam is supplied directly from the boiler to the LP cylinders only when starting, when the loco's speed and therefore the flow rate of the steam are low.

 

None of which was PRESERVED

 

When dealing with a compound locomotive you need to remember that the low pressure cylinder is fed through the associated high pressure cylinder. Ignore starting, steam demand is low here, as you have stated, and there has been much discussion about the best starting arrangements. When the engine is working in compound mode the h.p. cylinder should be working at a late cut-off, say 70%. The steam exhausted from the h.p. cylinder enters a receiver, and is preferably resuperheated in order to minimise surface effect losses (condensation), before being expanded in the l.p. cylinder which will have a volume of twice (or more) of the h.p cylinder feeding it. In order to achieve an adequate supply of steam to the 2 cylinder unit (h.p. + l.p.) you size your steam pipes as though it were feeding a single cylinder of volume equal to that of the total volume of the 2 cylinder unit. Hope this helps.

 

Sorry, does not help me! In a compound system the steam is used sequentially, while in a 4-cylinder simple it is used in parallel. The steam pipes could be of the same size. Please note that not only the h.p. cylinder works at a high cut-off but also the l.p cylinder, so with wide open ports. As a consequence the wire-drawing losses in the ports at 2 times 70% are less than in a simple wich works at say a comparable 50% cut-off. That is almost the only advantage of a compound.
Kind regards
Jos Koopmans

 

Where did the 4-cylinder simple come into this? When GJC rebuilt the French compounds he was not happy with the performance. The 4-cylinder alternative trialed was found to be more satisfactory. The h.p. steam pipes fitted to the compound were subsequently criticised. They might have been adequate to feed a cylinder identical in size but only used as simple expansion, though later developments in the understanding of the steam circuit would not support this view.
Steam chest pressure oscillates, in spite of many years spent trying to eliminate this problem it still continues. If the steam pipe feeding the steam chest lacked the capacity to feed the steam chest against the demands placed on it by the cylinder what would be the result? The compound was hamstrung in that the main steam pipes did not have the capacity to meet the demands that were placed upon them.
How do you avoid losses in a simple expansion machine? You want to minimise unexpanded losses. Equally you have a concern about wall effect. You are wanting to produce a certain range of power so how large are your cylinders going to be? Some variables you can juggle, others you cannot. The French tried making two cylinder simple expansion machines that were the equivalent of some of their compounds. These were the reverse of successful. In theory the steam locomotive is such a simple machine.

 

I think your enthusiasm is misleading you. 102 didn't receive any changes until 1909, when the Serve tubes were removed, and didn't get the std 1 boiler until 1916. 104 got a std 1 boiler in 1909, and 103 in 1910. By 1909 there were already 20 production Stars, so all the key comparisons has been made long before the boilers were changed to make maintenance easier. It was the 'as built' Frenchmen that didn't show sufficient advantage against the Churchward locomotives, not the modified form. Remember performance would be by no means the only factor under consideration, build cost, maintenance costs, all the things we would now call cost of ownership were vital factors.