As steam engine pressures increased, it was realized the exhaust steam still had useful pressure. The compound engine was invented to take advantage of this. The exhaust from a small high pressure cylinder fed into a larger low pressure cylinder to extract further work. The most famous example is probably the marine Triple Expansion Engine which greatly improved the efficiency of marine steam engines.

Compound engines were popular in marine and static applications, but their usage in railway locomotives was patchy at best. Although popular in a few locations (France comes to mind), the vast majority of 20th century steam locomotives were built with simple expansion engines.


  • $\begingroup$ This question is intended to push the envelope further on engineering questions, and is questionably off-topic when compared to my other engineering questions. Possible answers could involve scientific and engineering practicalities, but could also be cultural and/or economic. $\endgroup$
    – winwaed
    Commented Oct 30, 2014 at 18:39
  • $\begingroup$ I do not know the answer, but I conjecture that for locomotives fuel economy was a secondary matter (unlike for ships for which the ability to travel long distance without refueling is important). More important for locomotives was probably the size limitation of the engine. $\endgroup$ Commented Nov 8, 2014 at 1:27

4 Answers 4


This question was originally posted to test the limits of what is on and off topic. I was going to eventually post an answer, but I see we have two answers. In my opinion, both answers dance around the subject a little, so I am adding my own answer.

This kind of question is always going to be partially subjective. Locomotive development tended to occur in independent strands (companies, countries) with would often confer with each other. Hence the importance of each reason will vary between different countries and companies.

I will try to illustrate my reasons with examples, but they are only examples and there will be a bias towards the UK where I have greater knowledge.

  • Costs. Greater efficiency means you can get more power for a set amount of coal or require less coal for a set amount of power. This is important, even in a coal-rich country like the UK. As such, a working compound system should be an easy 'choice'. This would explain the choice in some later (1920s-30s) locomotive developments such as Chapelon in France (Gresley Observer vol. 156), and US Mallets.

    However there are multiple costs. In the case of the UK, maintenance costs often discouraged otherwise 'obvious' money saving issues - eg. feed water heaters (eg. LNER B12) and boosters (cf. the LNER's experience with the S1, P1). These developments often had a history of being fitted and then removed a few years later.

    Other answers have hinted at complexity. The reality is that it is possible to squeeze four cylinders into the restricted loading gauge of the UK (US and Continental Europe have more space). However compromises have to be made with steam passages and layouts. This adds inefficiencies and maintenance costs. These can be avoided with the three cylinder layout that was popular in the UK (eg. Stanier Duchess, and Gresley A3 & A4). Gresley's A4 has already been mentioned (one holds the world speed record). The internal streamlining is considered almost as important as the external (as determined by Chapelon and communicated to Gresley), and a complex layout such as four cylinders or compounding would not have allowed for this streamlining.

  • Effectiveness. This basically comes down to good design and train crews who know how to use compounding effectively. Early attempts at compounding in the UK at least, were imperfectly designed. UK companies tended to give limited training to the loco crews, whilst France (especially under Chapelon) implemented a high degree of education - hence they had much more success in the day-to-day use of compounding (Gresley Observer vol. 156)

  • Culture. This is harder to quantify, but it does seem that UK companies were largely poisoned to the idea after the initial failures. Although the North Eastern Railway finally managed to get a system developed by Walter Smith to work effectively, it never developed them further after his death. This might have been due to patent costs or management decision to move away from compounding. The Midland Railway did make extensive use of this, but only for a mid-power 4-4-0 (of which No. 1000 survives in the National Collection).

  • Requirements. The later use of compounding tended to be only when other factors made it more effective. For example, the US usage of compounds in Mallet locomotives. These were usually large freight-hauling locomotives with articulation and higher pressure boilers. Each power truck would have two cylinders. The pressure and space allowed for effective compounding. Similarly, Gresley used compounding on his W1 'Hush-Hush' experimental locomotive. This used an experimental 'water tube boiler' operating at the very high pressure of 400psi. To make effective use of this high pressure, Gresley implemented a compound system with two high pressure and two low pressure cylinders. The W1 never entered production, and it was the only example of a water tube boiler locomotive in the UK.


After bit research I got following things:

There were many efforts done in order to make compound steam engines by U.S., European countries, Russia, etc in 19th and 20th century. This engines were known as Mallets. Last try was done by Russia in 1954-55.

But By the mid-1950s, with economic recovery from the Second World War, production of diesel locomotives had begun in many countries and the diesel locomotive was on its way to becoming the dominant type of locomotive.

Also fastest steam engine was not having compound engine which was made in year 1938, thats proves compound engines are not as fast as normal steam engines.

So after reading a bit I come to following conclusion:

  • Due to invention of diesel engines, development of compound steam engines did not take place.

  • Normal steam engines with aerodynamics model were faster than compound engine.

  • Due to complexity of compound engine it was difficult and expensive in maintenance point of view.

Sources: Wikipedia - link1, link2, link3

  • $\begingroup$ Mallets may have used compounding but most compound locomotives were not mallets. Eg. Chapelon's compounds, the Smith system (used by the UK's Midland), and the van Borries system (used with mixed results by the NER). $\endgroup$
    – winwaed
    Commented Nov 16, 2014 at 18:42
  • $\begingroup$ "Normal steam engines with aerodynamics model were faster than compound engine." How about a compound engine with good aerodynamics? You've said that some single-expansion locos were faster than most or all compound ones, but that doesn't mean that it's because they were single-expansion. $\endgroup$
    – Flyto
    Commented Nov 16, 2014 at 23:09

Following the (mostly) very interesting links posted in other answers, I noticed that according to the Wikipedia page on compound locomotives, the use of compound engines on road locomotives was actually very common. This view seems to be borne out by other sources, such as this example and these examples.

It was apparently also very common for road locomotives to use one crank per cylinder, connected to a crankshaft. I believe that was also the case for triple-expansion marine engines. On the other hand I believe crankshafts were very rare on steam-powered rail locomotives; almost all pictures clearly show cranks attached to the outside of the driving wheels. The mechanisms to transfer power and to control the flow of steam would have had to be quite different for such engines than for engines with crankshafts. (For starters, it seems to have generally involved either multiple cylinders driving a single crank, or cylinders placed very far apart, presenting difficulty in delivering good quality steam from one cylinder to the other.)

Some builders succeeded in producing a number of compound rail locomotives nevertheless (an interesting account of this is here, continued here, including a panel discussion featuring Samuel M. Vauclain), and there were even some compound road locomotives with single cranks, but when you consider the investment in new technology that would be required, perhaps it is not too surprising that few builders tried to develop it and even fewer succeeded before diesel took over.


Each of the following references indicates that the "Compound steam engines" were fast because of the increased thermal efficiency.

But, the high maintenance costs brought on by the more complex equipment tended to outweigh efficiency advantages.

  1. Link1

enter image description here

  1. Link2

    There's a big difference between a marine VTE engine, which operates at relatively low speeds, and a Vauclain or de Glehn compound, which are high speed designs. Also, most of the early compound designs were withdrawn with the advent of superheating.The various D&H experimentals in the 1400 series were complex high-maintenance designs. It was said that they could pull well but you had to send half of the shop force out with them.

    1. Link3

Ty4 was the most powerful locomotive in the PKP service and these engines operated mainly on the Coal Trunk Line between Upper Silesia and Gdynia on the Baltic coast. Superseded byTy246s and later also Ty51s, they disappeared from this line, but soldiered on elsewhere. Later many were used as heavy switchers, often with smaller tenders. Repairs of these engines were complex and time-consuming, as their three-cylinder layout made them untypical and suitable maintenance facilities were scarce. Availability rate was thus, at least initially, quite low.

I know this doesn't seems like a complete answer but this is the gist of my research.

  • 1
    $\begingroup$ Could you include the references in the link text. This will make it a better answer. At the moment there's no context. thanks. $\endgroup$
    – winwaed
    Commented Nov 15, 2014 at 20:05
  • $\begingroup$ Its not complex :P $\endgroup$
    – tox123
    Commented Nov 16, 2014 at 0:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.