American steam locomotive efficiency- the effect of blastpipe size and superheat levels.

142 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    June 2008
  • 121 posts
American steam locomotive efficiency- the effect of blastpipe size and superheat levels.
Posted by Dreyfusshudson on Thursday, March 24, 2011 7:15 AM

It is generally accepted that American steam locomotives achieved thermal efficiencies of around 6% at the drawbar, and there are claims that with not too radical technology, this could have been significantly improved. Porta said American draughting systems (including blastpipe dimensions) ‘could not have been worse’. I am lacking some key bits of data I need for this, and am wondering if anyone in these forums might have access to the data.

The calculation of drawbar efficiencies requires knowledge of three things- engine efficiency (how much steam needed to produce 1 horsepower for an hour), boiler efficiency (how much coal to produce a given amount of steam) and the amount of power that needs to be expended to move the locomotive as a vehicle, which must be subtracted from the cylinder power to get drawbar power.

With regard to engine efficiency, I have a computer programme developed by a very distinguished UK Professor, which calculates engine efficiency from first principles. I have validated this against UK and Altoona test data, and with some caveats that I can share with those interested, it works just fine. What it says is engine efficiency is dominated by the beneficial effect of higher superheat, which generally gets higher at higher steam rates; the next most important factors are blastpipe pressure, which increases detrimentally at high exhaust rates, and cut off. When locomotives are operating in less than 30% cut off, cut off is not such a big deal, but higher values, as required at low speeds and when climbing mountains become quite detrimental. A final component on which I am still working is leakage of steam e.g. within the engine past valves and piston rings, but basically, if you know superheat, and how it varies with steam rate, and blastpipe area, which determines back pressure, you can very easily get a pretty good estimate of engine efficiency at any cut off or speed, hence power developed.

Boiler efficiency is similar for all hand fired locomotives at the same coal rate, or evaporation rate/sqft of grate; for hand fired locomotives it is about 70% at an evaporation rate of 600lbs/sqft/hr, falling to about 60% at the higher 800lbs/sqft/hr- not far from a probable practical upper sustained limit. For stoker fired locomotives these figures are about 59% and 50%. (UK and US data agree on this). If you think about the underlying theory, the reasons for this similarity are not surprising. There are two basic factors, the efficiency of heat transfer from flue gases to steam, which is pretty good, around 80%, not changing much with steam rate, and unburned coal loss, which basically depends on the combustion rate/sqft of grate, increasingly very dramatically at high rates; it is this component that is worse for mechanically fired coal. These figures can be improved by feedwater heaters and exhaust injectors.

Locomotive Resistance was not, in the UK at least, experimentally well defined in steam days. New analyses of old data have given a formula which is consistent with the best of that data. US locomotives are much larger, but the since the overall formula is based on realistic formulae for the basic components- aerodynamic resistance, the resistance of the machinery between the pistons and the wheel rims, and resistance of the locomotive as a vehicle, it should extrapolate reasonably well and at least get into the right ball park for US locomotives.

So, to estimate US locomotive efficiencies all one really needs to know superheat and blastpipe dimensions. The US data I have is basically for the period from the 1890s to 1920s from Altoona. The PRR K4 had a maximum superheat of about 610 degrees at high steam rates, and a blastpipe of 7” in diameter, but things did move on, I suppose, after the K4! Tests on the 1926 Baldwin High pressure compound 60000 show its superheat was 680 degrees, but had something like a 6.5” blastpipe- the backpressure was about 20 psi at an exhaust rate of 60000lbs/hr- 730lbs/sqft/hr- . What I lack is any data for the final essays in US steam, apart from 1984 tests on C&O 614 which showed superheat of around 800oF, as high as anything I am aware of.

Some illustrative sample calculations on a UP Challenger, and UP 4-8-4 are shown below at a constant cylinder power of 5000HP. Please note that these use entirely made up estimates of superheat (680oF) and blastpipe dimensions (two times 5 5/8” diameter)!


UP Challenger

Cut off %

Speed Mph

Cyl steam, lbs/hr

Lb Evap/ sqft grate*


Back press psi

Cyl eff %

Loco Res HP

Equiv DHP

Cyl eff ref to drawbar %

Boil Eff, %

Overall  DB eff, %

















































UP 4-8-4































































*Allowing 10% benefit for feedwater heater


Despite the uncertainties in the inputs, the basic messages are clear. Note that as speed increases, cylinder efficiency increases, as the cut off required for 5000IHP decreases. (For each locomotive the minimum speed at which 5000IHP can be developed is shown, assuming 75% maximum cut off). As the amount of steam needed to produce 5000IHP decreases, so the boiler efficiency increases; counteracting these two positive benefits however, is the increase in locomotive resistance, which means drawbar efficiency peaks at about 50-60mph, and then declines. Given US freight locomotives spent a lot of time working at low speed, the quoted 5-6% thermal efficiency at the drawbar is not surprising. Under the (unsupported) assumption I have made about blastpipe diameters back pressures are high, especially at low speeds. Note also that the Challenger is more efficient than the 4-8-4 up to 80mph, because of its larger grate and lower cut off- the comparison is unfair to the smaller 4-8-4 which, whilst it clearly can develop 5000IHP, would be more efficient working at 4000-4500IHP.

What I would like to be able to do is produce comparative calculations on the last and best coal fired US steam designs NYC, PRR, C&O, NW, UP etc. As a basic minimum, I would need superheat vs steam rate and blastpipe dimensions (free nozzle area, i.e. basic diameter, less any protuberances into steam flow) for these types (I have basic engine and boiler dimensions).  If anyone has access to these data (or even better full test reports), and would like to assist in this venture, I would be more than interested to receive them. 

I realise there are a number of sweeping assertions in what is written here - some of this is due to the need to keep this input simple, and would be happy to discuss any detailed objections people might have.

  • Member since
    August 2003
  • From: Northern VA
  • 480 posts
Posted by feltonhill on Thursday, March 24, 2011 8:26 AM

I believe we need to clarify some terms first.  The steam temperatures you mention are not superheat but total steam temperature.  Big difference.  250 degrees superheat was considered conservative for modern steam locomotives.  Tests of the NYC Niagara and PRR T1 showed superheat temps of well over 300 degrees at their respective boiler pressures (275 and 300 psi).

Another item is the constant IHP as speed increases.  Test data indicates that this is not how things work. Most estimating methods compute theoretical IHP as being asymptotic as speed increases.  However, actual IHP tends to increase up to about  driver diameter plus 10% then starts to decrease.

I'm also suspicious of the very high steaming rates at low speeds for both examples above.  Whereas both boilers could probably produce over 100,000 lbs of steam an hour, neither would be particularly efficient at such high rates.

Be very cautious about simplistic ratios when trying to estimate locomotive performance.  I've seen a bunch of them and almost always they produce only a meaningless average.  Worse, they tend to overlook various components that are important to consider in the overall scheme of things.  Steam locos are not easy to predict.  I've used generally accepted industry methods and tweaked them a bit for modern construction and  some other items.  There are abut 20 equations involved, none of which I would call simplistic.  Careful with this....  I'm interested but skeptical.

  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Thursday, March 24, 2011 10:58 AM

Thanks very much for your response. You are correct to pick me up on the loose use of the word 'Superheat'; I do of course mean inlet steam temperature not the level of superheat in the steam.

The use of constant IHP in my calculations was just for illustrative purposes. If you use constant steam rate as the basis, then you do indeed get IHP values that asymptote as speed increases. There is a possibility that IHP will decrease if you go above a certain speed at constant steam rate, i.e. efficiency decreases despite the fact the cut off is shortening and expansion ratio improving. UK test plant work was normally not above 80mph, so with your diameter +10% rule, you would not expect to see this, and you don't except for some hints thereof on older GW railway designs with low clearance volume, where there is serious overcompression at high speeds.

You can use the computer programme to establish when this would occur; it depends on the operating condition. With late UK designs, e.g. the streamlined A4, (80" drivers) it says that at constant steam rate in the normal working range, efficiency begins to plateau around 100mph and shows positive decreases about 120mph. The point is that where this turnover point is can be now  be calculated, if you have the detail of the engine dimensions (Lap, lead, clearance volume, port width etc). This is the next layer of detail that I would need to understand about US types, but tens of thousands of calculations all say that from the point of view of gross efficiency, the point in question, these things are second order.  

Now in practice locomotives were worked neither at constant steam rate nor constant power of course, and my argument justifying the 5-6% overall thermal efficiency is indeed a bit simplistic. However, I have done lots of work simulating logs of reported steam performance on various main lines (the only US one on which I have sufficent data is the Milwaukee from Chicago to La Crosse), and you can use equations derived from the outputs above to work out fuel consumption on a trip, and average DHP. It is really this experience of translating predicted performance into practical estimates of drawbar efficiency that leads me to believe that my conclusion about drawbar efficencies is sound- but that's another huge discussion!  Incidentally, the MILW calculations surprised me in showing how little power relative to the 96.5 sqft grate was needed to do what the F7s did when cruising at speed; the secrets of fast schedules were barnstorming starts and super slick braking. O to have witnessed them.

I would be very surprised, like you, if a 4-8-4 with a 100sqft grate was ever steamed at 100000lbs/hr, even a Challenger, so I do understand this point; in the UK outputs above about 750lbs/sqft/hr were rare, even though they are surely possible.  Bear in mind again I was only trying to illustrate the kind of calculations that can be done; what are the most relevant calculations is entirely discussable.

The calculations  are very easy to do- as long as it takes you to input 3-4 numbers and press. In the 1950s, UK test plant results were reduced to a 'power map' which had IHP on the ordinate,  speed on the abscissa, and lines of IHP at constant cut off, and constant steam rate, together with the coaling rate at each steam rate- a one stop shop that tells you everything you need to know about power and efficiency of both boiler and engine. If equivalents for US designs exist, I haven't seen them, but it is these power maps that I would like to create.

The calculation of engine efficiency is by no means simplistic, as you seem to fear; it is standard fluid mechanics, with cylinder heat transfer built in, plus steam tables, and requires simultaneous integration of 5 separate partial differentials. The big surprise is the simplicity of the headline conclusions that it leads you to, perhaps the most important of which is that many factors thought to be important in the steam era aren't.  And once you start to think about it, it all makes sense. In the UK at least, the data available in steam days was simply not good enough to work out all the subtleties; whilst people knew what might be important, they coudn't be precise about how important; this is what the computational model helps you establish, and it is quite revealing. So I believe there are indeed simple conclusions, but this does not mean the process used to achieve them is simplistic, far from it. You will need to forgive my enthusiasm! Having said this, having compared the computer predictions to about 1500  fully instrumented test plant runs, it is clear there are things that are not so clear, and you do need to be aware of this, but these uncertainties are factored into my enthusiasm.

If only we'd had high speed computing in 1910, things might have been different. Unfortunately, we needed the steam locomotive to do its job, and pass on, to put things in place that allowed us to invent computers!


  • Member since
    July 2006
  • 3,312 posts
Posted by locoi1sa on Thursday, March 24, 2011 6:48 PM

Google books has tests of locomotives from the Pennsylvania RR done on the test plant. These are very interesting reads. The PRR testing department tested everything. Even the coal is broken down to its BTU and ash content.

 Here is a test on an E2s Atlantic loco from 1910.


[Edited to shorten URL]

 I pray every day I break even, Cause I can really use the money!

 I started with nothing and still have most of it left!

  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Friday, March 25, 2011 7:00 AM

Dear Pete,

Thanks for the tip- unfortunately the Altoona reports on Google are the ones I have access to! Reports 1-32 were presented to the British Institute of Mechanical Engineers in two bound volumes ca 1925, personally signed by (I think) the PRR CME. These are now in the Science Museum Library in London, disintegrating rapidly. They are wonderful documents, comprehensive data, crystal clear write ups, confident without being overbearing- a model. In terms of testing the PRR was way ahead of the game at the time. I think  the 1914 Bulletin 24 on the effects of superheat is one of the most influential ever written; I'm sure Chapelon digested it and in this country, Gresley immediately copied the variations in superheater set up on his own Pacifics in 1926.

Still hoping someone has access to details of the final flowering of US steam from ca 1926-1949

  • Member since
    January 2002
  • 4,072 posts
Posted by M636C on Friday, March 25, 2011 7:17 AM

Didn't the late UP locomotives have Le Maitre exhausts?

Some photos of these show multiple jet nozzles arranged in a circle and the large diameter stack suggests that arrangement. That is at least part of the way towards Porta's ideas.


  • Member since
    February 2005
  • From: Vancouver Island, BC
  • 22,231 posts
Posted by selector on Friday, March 25, 2011 11:02 AM

In videos I have seen of UP 844, there are actually two stacks.  Or, so the twin plumes emitting from the 'stack' suggest.


  • Member since
    July 2008
  • 648 posts
new formulae for calculating power and efficiency of steam ?
Posted by Juniatha on Saturday, March 26, 2011 3:31 AM


Er-hm .. if you may pardon my joining your discussion, may I just add a few words?


The UP Jabelmann double chimney arrangement included quadruple blast nozzles each. With that draughting both Challenger and FEF-II could and likely did exceed 100,000 lbs/h steaming rates. However may I ask how backpressure, draught vacuum and efficiency of this or any other draughting arrangement should be calculated indiscriminatingly of design specs?

In paragraph one of your original posting you correctly quote Porta’s saying – yet this must be seen in context. First, as from what I’ve had an opportunity to lend and read of his papers it appears Porta had a liking for strong expressions and exaggerations to make his point clear (he enlightened in using the word ‘exaggeration’ for some of his own improvements, sometimes meaning to provide a design margin against malign influences, at other points meaning an emphasise on going for sth full stride – very best efficiency, mostly). From what I have read of his, as much as he tended to elevate importance and effect of his own developments, he also tended to degrade earlier technical standards.

There is no denying, draughting arrangements were simple enough in most of US steam loco designs right to the end of the development line. However, they sure could have been a lot worse. In fact, within their very plain and simple layout, more often than not they came tolerably near to best proportions and performed reliably, if at rather high rates of back pressure. In those times this was not considered too harmful in view of relatively long cut offs and consequently high exhaust release pressures with engines run flat out. Of course this consideration was short sighted as high back pressure did in fact reduce mean cylinder pressure – and that did of course adversely affect gross cylinder output and efficiency. Interestingly, the relatively frequent exchange of worn blast nozzles necessary as a consequence of highly abrasive effects of extreme flow rates of steam at rather high pressures was considered forbidding to more sophisticated designs that would have been more costly to build and to install. What was lost however: such designs should have worked on much lower steam pressures, thus at much less aggressive flow rates that again would have made these designs last much longer. Clearly, lack of introduction of more advanced draughting arrangements was due to the early and precipitous end of development of the American steam locomotive.

Some years ago I had written an abstract on thoughts of my own as how indicated power output of the Niagaras could have surpassed 8000 ihp at speeds around 100 mph just by more sophisticated draughting and valve gear, including improved cylinder tribology and accordingly raised live steam temperature, all that on the same amount of net steam heat, i.e. even at slightly eased steaming rate, or in other words exclusively by improved thermo-dynamic cylinder efficiency. Since these improvements would not be reflected in data you consider for calculations by your general formulae, how then could it get correct results for different steam loco types of differing qualities of design? For sure, it should be interesting to read your explaining about >> lap, lead, clearance volume, port width etc << and >> many factors thought to be important in the steam era << .. >> tens of thousands of calculations all say that [in] view of gross efficiency .. these things are of second order << to quote from your first reply paragraph three, lower, and paragraph seven, upper. So far I had believed on the contrary these factors do make a difference – and have further believed Porta, Wardale have believed so, too, when they remarked upon how these things had been unwarrantably slovenly developed?

In paragraph three of your original posting you mention 30 % cut off  c/o as something of a dividing point above which length of c/o has a decisive influence on specific steam consumption s.s.c. while below it doesn’t. I can see no theoretical basis for that assumption. The effect of improving s.s.c. by ever shorter cut off levelling out around a certain point depends entirely on imperfections of actual engine characteristics against the ideal steam cycle at the same live steam values. The more these imperfections bear on the actual working characteristics, the sooner a minimum point is reached beyond which further shortening of cut off does no more improve s.s.c.

This happened to be around 30 – 25 % c/o in German standard type two cylinder steam locomotives such as the 01 and 03 class Pacifics and their Decapod counterparts 50 and 52 classes and to an extend also the 42 class. 30 % cut off was generally used for constant speed running, often with laminated throttle and I think drivers were even told not to go below 30 % in view of mechanically hard running – which to me seems arguable at least and was proven wrong by Wardale for his improved SAR 19-D and 26 engines. Clearly it largely depends on mechanical design specs and actual engine condition as much as on valve gear characteristics. Certain poppet valve equipped simple expansion engines could advantageously run on c/o as low as some 5 %!

So, with engines of advanced cylinder tribology and valve gear short c/o does pay while with the more indifferent kind of designs it didn’t. Clearly these contrasting behaviours cannot be banged together in one common formula. Power output graph over speed is not just higher in a more advanced engine per same steaming rate but its line also evolves into a different shape: while the more indifferent type of engine shows an apex of indicated output at a certain speed above which the line starts to fall off like the flight path of a ballistic missile, the advanced engine shows incessantly increasing power output over the entire practical range of speeds, i.e. as a Niagara improved along the lines mentioned above would have shown power output increasing over the entire then legal speed range of 0 – 120 mph.

Back pressure on piston is the one that counts as for cylinder output and cylinder efficiency – not blastpipe pressure (measured at which point?); pressure in blastpipe, or actually: at blast nozzle does neither depend on speed nor cut off but on steaming rate only – logical if you come to think of fluid dynamics and the variable mass / time rate of steam passing nozzle(s) of fixed cross section and flow coefficient. Sorry, I believe your relative data on hourly steam to cylinders for your given 5000 ihp looks somewhat inconsistent, it suggests too low values for s.s.c. Also, the Challenger having a slightly lower s.s.c. at the same speed and ihp relative to the FEF-II would surprise me indeed, taking into account that it is a simple expansion Mallet type while the FEF-II is a straight two cylinder engine.

What c/o did you calculate for the FEF-II at 100 mph / 5000 ihp? In data in column four you allow >> 10 % benefit << for feed water heater – 10 % of exactly what? why don’t you use the same net steam rate to cylinders as in column three? what evaporation data does your boiler efficiency relate to? if all your data so far should be based on net steam to cylinders, then overall drawbar efficiency lacks to account for steam to auxiliaries / if not all of your previous data is based on net steam to cylinders then what data is based on what steaming?

In calculation of locomotive resistance, mechanical resistance (drive), mechanical resistance (rolling) and aerodynamical resistances should be kept separate or you get a blurred picture; I’d like to see the formula that allows for correct data of the total of all of them to be established over speed and then based on data extrapolated from British steam locomotives of, as it was, rather special mechanical characteristics with their composite plate frames, very light rodding and extremely small diameter pins, design prone to flexure under high stress, and their smooth external lines of small loading gauge. Maybe this has lead to rather high locomotive resistance hp losses in the upper speed range and somewhat low values at low speed.

To quote paragraph 5, last sentence of your first reply: >> I was only trying to illustrate the kind of calculations that can be done; what are the most relevant calculations is entirely discussable. << Well, sure there are many ways to put up calculations. Ok, likewise I’m only throwing in a few words of mine to an interesting discussion – no insult intended.

So, to quote once again from your last paragraph in your reply: >> You will need to forgive my enthusiasm! << Well, I guess there will be no problem here, we are all enthusiasts of railroading power one way or another – steam, diesel or eclectic electric. 





  • Member since
    August 2003
  • From: Northern VA
  • 480 posts
Posted by feltonhill on Saturday, March 26, 2011 8:01 AM

Regarding modern US locomotive data:

Test reports for both the PRR T1 and NYC Niagara survive, but they are not published for sale.  I know they exist because I've used them extensively over the years.  I believe the Q2 test report also survives, but I've not seen it.

A book called Santa Fe's Big 3, by S. Kip Farrington, has considerable operating data on the last classes of  ATSF 4-6-4, 4-8-4 and 2-10-4.  I have it and it's proved to be a good source of information.  This book is usually available on line from several used book vendors.

There's a very good book available in the UK called Dropping the Fire by Phil Atkins.  Although this is not technically rigorous, it has exceptionally good write-ups on several US locomotive classes (Niagara, PRR T1, N&W Y6's...).

Back to the original post.  I notice that total evaporative surface is used.  I believe this is shortsighted, particularly with US locomotives.  After 1930, more locomotives had large combustion chambers, and this altered the proportion of evaporation produced between the direct heating surface and indirect heating surface.

Two examples -

(1) The PRR M1 4-8-2 of the mid 1920's had PRR's "standard" grate area of 70 sq. ft. ( used for the K4 4-6-2, L1 2-8-2, I1 2-10-0 and later the K5 4-6-2).  However, it had a very large combustion chamber which allowed higher firing and evaporative rates.  This increase in furnace volume enabled this improvement

(2) A much later example can be found in N&W's modification of the interior proportions of its famous Y6 2-8-8-2 compounds in the early 1950s.  Initially they had 24 ft flues and a short combustion chamber.  This was later modified to 20 ft flues and a 48" extension to the combustion chamber all within the same boiler shell.  This produced in increase in the direct HS and a reduction of the indirect HS.  The total evap heating surface was reduced, but the overall steaming capacity was increased because most of the evaporation takes place over the direct HS rather than the indirect HS.  N&W stated that this modification improved the combustion characteristics of the firebox and the evaporative capacity of the boiler  substantially.

Juniatha has also brought up valid points regarding the basic construction and configuration of UK locomotives vs those in the US, not to mention differences in operating philosophy.  The frame design differences alone have an effect on locomotive resistance.

Based on appearances, the UP 4-8-4's, 4-6-6-4's and 4-8-8-4's all had similar exhaust designs: double stacks with double blast pipes separated into four exhaust nozzles each (a total of eight).  It bears some resemblance to  later European designs.  This was done primarily because UP used low-quality on-line coal, which had a fair quantity of fines.  This required an easier exhaust blast, so that the fire wouldn’t be lifted off the grates.

While we're tossing numbers around, my estimate for a UP FEF-3 (844) would be about 5,500 IHP at 80 mph and 86,000 lbs/hr total evaporation.  For the 4-6-6-4 (3985), it would be about 6,000 IHP at 70 mph and 94,000 lbs/hr total evaporation.  These are relatively moderate figures, not full-flog readings.

 I’m not sure where this discussion will go, but I’m still reading and trying to understand.  Exhaust design is not my best subject by a long shot!

  • Member since
    April 2003
  • 305,206 posts
Posted by Anonymous on Saturday, March 26, 2011 11:20 AM



While we're tossing numbers around, my estimate for a UP FEF-3 (844) would be about 5,500 IHP at 80 mph and 86,000 lbs/hr total evaporation.  For the 4-6-6-4 (3985), it would be about 6,000 IHP at 70 mph and 94,000 lbs/hr total evaporation.  These are relatively moderate figures, not full-flog readings.

talking about efficiency, how much power could modern "all axle roller-bearings based" steam-engines

produce at the tender's draw-bar, including mechanicary and air-resistance, actually?

I have never seen a "conversion table", but starting with your data, maybe ~20%?

So, 4500 DBHP for  a FEF-2/3, 4800-5000 DBHP for the Chally?  Just as a rule of thumb...



Those figures in Mr. Dreyfusshudson's table talking ~10% ? Is not it awesome, from that point of view that modern diesel-electrics do not do better, regarding shaft power to drawbar.




  • Member since
    April 2003
  • 305,206 posts
Posted by Anonymous on Saturday, March 26, 2011 11:34 AM

Er-hm .. if you may pardon my joining your discussion, may I just add a few words?
The UP Jabelmann double chimney arrangement included quadruple blast nozzles each. With that draughting both Challenger and FEF-II could and likely did exceed 100,000 lbs/h steaming rates.




Do you have the source of  "100,000 lbs/" ? Sometimes, I have those "weirdo" thinking, that UP-Engines were pretty underestimated, concerning that...

Looking at the BB-tests in 1943, the engines produced all less than that rate (but have a bigger boiler than a Chally or Northern). All evap.-rates were beyond 100.000 lbs/h. Though, they were not running fast, what is is the missing link here?





  • Member since
    April 2003
  • 305,206 posts
Posted by Anonymous on Saturday, March 26, 2011 12:26 PM



Thank you all for this interesting post + data, can you tell more about this program? Is it shareable? 

Kind Greetings


  • Member since
    July 2004
  • 2,404 posts
Posted by Paul Milenkovic on Saturday, March 26, 2011 6:53 PM

I commend you on your work on this topic, and I offer you encouragement in the face of some of the naysayers.

What you offer is an engineering model, that is, a somewhat simplified version of a collection of much more complicated formulas and equations.  The purpose of your model is 1) to predict the performance of known steam locomotives, with the view of understanding the rough contribution of various system components -- the grate, the evaporative surface area, aerodynamic resistance in the steam circuit, exhaust back pressure, and so on, and 2) armed with the ability of the model to predict known steam locomotives, to make educated guesses regarding what enhanced or improved steam locomotive designs could have done.

The fact that you get into "the ballpark" of the 5-6 percent thermal efficiency often quoted for 20th century U.S. steam, and that you get into that ballpark based on thermodynamic, fluid dynamic, and empirical relationships tells me that leaving aside the naysayers, that the quirks of individual steam engines are unknowable apart from much more complicated formulas, that you indeed have a good model of steam locomotive thermal efficiency and are to be commended for your efforts.

I am sorry I don't have data or measurements on steam locomotives to contribute.  My only suggestions is that the David Wardale 5AT steam locomotive project has a Web site along with links to technical information you could check out (does his prediction of efficiency on the low to mid teens for this proposed design square with your methods?).  David Wardale also has this out-of-print book The Red Devil and Other Tales from the Age of Steam -- if one could get a copy somehow, I am told it too has considerable technical detail on steam locomotive efficiency.

I commend you on your efforts and look forward to more results from your model.

If GM "killed the electric car", what am I doing standing next to an EV-1, a half a block from the WSOR tracks?
  • Member since
    August 2010
  • From: Henrico, VA
  • 8,955 posts
Posted by Firelock76 on Saturday, March 26, 2011 7:22 PM

To Lars Loco:  How much power could an all roller bearing axle locmotive have at the drawbar?  Well, have you ever ridden on a jetliner?  I rode behind Norfolk and Western's  Class "J" 611 several times before the excursion program was stopped, and got that same set-back in the seat when the engineer opened the throttle!  Not as much, mind you, but it was certainly there!   A   "J" could produce  73,300 pounds of tractive effort, can't tell you the horsepower, none of my books have the answer,  but  my dear God, what a magnificent machine!   If General Lee had one available to him in the 1860's, along with  "Seven-Elevens" and pick-up trucks, the South would have won The War!

  • Member since
    August 2010
  • From: Henrico, VA
  • 8,955 posts
Posted by Firelock76 on Saturday, March 26, 2011 7:25 PM

Well, all right!  Juniatha's back!  You go, girl!

  • Member since
    April 2003
  • 305,206 posts
Posted by Anonymous on Sunday, March 27, 2011 3:40 AM


To Lars Loco:  How much power could an all roller bearing axle locmotive have at the drawbar?  Well, have you ever ridden on a jetliner?  I rode behind Norfolk and Western's  Class "J" 611 several times before the excursion program was stopped, and got that same set-back in the seat when the engineer opened the throttle!  Not as much, mind you, but it was certainly there!   A   "J" could produce  73,300 pounds of tractive effort, can't tell you the horsepower, none of my books have the answer,  but  my dear God, what a magnificent machine!   If General Lee had one available to him in the 1860's, along with  "Seven-Elevens" and pick-up trucks, the South would have won The War!

while reading your last lines...

do you know Charlie Chaplin´s "The General" ? It shows the most dramatic stunts ever made in movie´s history, now imagine he would have had a 611-class engine available ... phew... too though even for Charlie...

  • Member since
    August 2010
  • From: Henrico, VA
  • 8,955 posts
Posted by Firelock76 on Sunday, March 27, 2011 12:19 PM

To Lars Loco:  Oh yeah, I've seen "The General"!  A very funny movie, but seeing that old 4-4-0 get wrecked on the burning bridge makes my skin crawl!  Oh well, at the time it was OK.  and the locomotive was due to be scrapped anyway, but I can't look at it now without thinking of it as the loss of a fine railroad artifact.  And doing it to a Class  "J"  doesn't even bear thinking about!  By the way, I misspoke on the  "J"s tractive effort.  Post- war modifications brought them up to 80,000  pounds of tractive force, making them the most powerful  "Northerns"  ever built.  Oh, and the Norfolk and Western always called them  "J"s, there was no way a Virginia road was going to call them  "Northerns"!   "Old times there are not forgotten...."

  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Sunday, March 27, 2011 6:50 PM

First of all, let me express my appreciation for all the comments you have all made- a very considerable efforts in some cases.  Thank you very much indeed. I will try to produce a single response to all these points, starting with some generalities, then getting to specific individual comments and queries later.

Let me start by referring to Feltonhill’s comment ‘I’m not sure where this discussion is going’ and reiterate my hope as to where it might go. What I have done, as Paul Milenkovic observes it to link together a series of engineering models of the key components of drawbar efficiency; engine efficiency, boiler efficiency, and Locomotive resistance (LR); the first is a fully worked through first principles model, the latter based on detailed estimates of the likely magnitude of the components of LR; the boiler efficiency model is simply an empirical model based on test results.

Now such theories are wonderful things, and the question is rightly posed, but what of the experimental data?  As I reported, I have explored the value of these models by comparing their predictions to a wide range of UK data collected at the test plant, about 1500 2 hour long test plant runs, about a dozen different locomotives from different builders, with different design philosophies, plus road testing. The models predict the outcome of these tests remarkably well. In this one ought to bear in mind that there are differences of 5% or more in measured efficiencies of different members of the same class, both ostensibly in ‘as new’ condition (a fact not reported), for reasons I think I understand, and tests also showed that between visits to the shops, efficiency could deteriorate by 10% or more.  This needs to be borne I mind when trying to validate the models against data. Also, it was not unknown in our country for enginemen to waste steam through safety valves, draincocks whistles etc., use it for auxiliaries such as braking, worst of all actually using precious steam to keep passengers warm.

I’m interested in finding out how well these models apply to US types. The stimulus was the Porta comment I quoted, which implies that there might have been major scope for improvement; I’m a bit sceptical about this. My hunch is that latter day US designs would all have efficiencies within 10% of each other, and at most the best could have been improved by 20% by the kinds of technologies advocated by Wardale and Porta, half of this from the GPCS firing system,  significant, but not game changing I think.

What I therefore need is US data of two kinds.

Firstly, input data for the models. The prime need is for information on blastpipe dimensions and superheat levels, for the work I have done shows these are the things which will most impact engine efficiency. The engine efficiency models also require detailed knowledge of 19 separate dimensions of the motion, to calculate accurately the openings of the ports during the engine cycle; I also need to know lap, lead, exhaust lap, port width, the perimeter of the ports (i.e. % of valve circumference with open ports), clearance volume, operating pressure, cylinder diameter and length. Hoping to keep things simple, I asked just for superheat and blastpipe dimensions. I have access to some of the other data, and can work my way round what I don’t have. This would allow first pass estimates of engine efficiency to be made.

The second need is then for US test data to see how well, or otherwise, these predictions fit with experimental reality; in this it should be noted that as well as taking into account the variability in different class members, in the UK there were in initial plant testing some very basic problems in measurement- fortunately there’s enough raw data available to be able to pinpoint what these problems were. So measurements are not always as reliable as they seem. There is evidence of this in the Altoona test results too.

Even if such test data were not available, I think these first pass calculations would allow one to say how much scope for upward improvement existed, and give an idea of how much the engine efficiency of latter day US designs in fact varied. I would be happy to publish the results of these calculations in this forum on a ‘for what it’s worth’ basis. The quality of the output will depend on the quality and amount of input data I can get- garbage in, garbage out. To improve estimates of boiler efficiency and LR would require more data still, though for first pass estimates, I think what I have is good enough. I am acutely aware that there are aspects of US practice that my models may not incorporate.

One final generality. My interest is in how much efficiency can be improved, as Wardale, Porta and the 5AT group, with whom I am in regular correspondence. However, locomotive efficiency is but one element in the overall financial mix. My understanding is that even in diesel days, US philosophy has been to send out   freight trains with just enough power to get over the road without stalling (I exaggerate). Working heavy trains at low speeds leads to low drawbar efficiency, as my illustrative calculations show. Drawbar efficiency could be improved simply by running shorter, faster trains;  even then this is not the full story, because if you run faster you need more dhp-hr to cover a given distance, so putting coal consumption up. How this all pans out is a question for accountants; all engineers can do is provide an accurate assessment of efficiency over a range of operating conditions, to allow you to estimate operating costs. Overall it appears to have made sense to run longer, slower trains, thus depressing drawbar efficiency.

Now to specifics.

Firstly, Juniatha, March 26th 09:31. Thanks again for such a detailed set of comments. I’d obviously be very interested in knowing the dimensions of the of the Jabelmann arrangement. It was watching the soft exhaust of 844 and 3985 on You Tube that led to me wonder about Porta’s assertion. With respect to your question about draughting, estimating this is not part of the process. I have done some work, based on ideas of the author of the engine efficiency programme, to attempt to work out what the entrainment ratio (ratio of flue gas to exhaust steam) of a front end is, hence how much coal can be burned. This foundered on the inability to estimate how much the chimney is able to flatten the ‘peaked’ velocity profile of the exhaust leaving the blastpipe. Others are working in detailed computational fluid dynamic modelling to address this. The engine efficiency programme only needs to know the blastpipe pressure, which is determined by the flow and temperature of the steam. Since the programme does not know this at the outset, it does a series of iterations until the flow and temperature match the flow into the cylinders, and efficiency of expansion.

On the steam rates, I’m in no position to challenge your assertion that the UP types can and did exceed 100000lbs/hr, but it would have been very uneconomic (though see comments above about economy). However, I think a distinction needs to be drawn between what could be done, and what was regularly done. If I may illustrate; in this country, we had a class of 10 wheelers, the Scots, that had small grates (31 ¼ sqft) relative to the amount of work they were expected to do. Consequently, the specific evaporations they achieved were higher than any other type. On their best runs they sustained about 1650IHP and produced shorter term bursts of 1750 IHP. This translates to about 700-750lbs/sqft evaporation. Crews claimed they ‘wouldn’t steam’ (I think they meant they burned too much coal!), and one was sent to the test plant where they got 1020lbs/sqft/hr on second grade coal, the highest recorded there. One of the class is still working one of the many steam specials we have. Without wishing to suggest anything other than top class professionalism, crews love to thrash their steeds. In 2009, it delivered a short term effort of 2350IHP needing nearly 1150lbs/sqft/hr, 50% more than they habitually delivered, higher than anything else I can find in UK history, but hardly representative!

Understand your comments about Porta.

Comments on US draughting systems; understood and agreed; it is just how much more might have been possible with better draughting that interests me.

According to my programme, if you had a blastpipe area equivalent to a 10” diameter single pipe you could get 8000ihp from a Niagara, 45% cut off at 100mph, cylinder rate about 117000lbs/hr  if superheat were 750oF, quite possible from a 100 sqft grate + feedwater heater, but not I think a regular practicable proposition. Because of the need to have sufficient cylinder volume to develop high power at low speeds, all engines can deliver far more power than their boiler could hope to produce at high speed: the issue is boiler capacity, not engine power.   

I’m not saying that lap, lead etc. make no difference, just that the effects are not that large, this on the basis of analysing UK style dimensions; it might pan out differently with US dimensions, though I doubt it. There is a well told story in this country about how changes to these settings led to a massive improvement to the efficiency of the original Doncaster Pacifics. One of the ‘high priests’ of valve gear design in this country, and a Doncaster man too, saw the results of the programme I am referring to just before he died, and, understanding the basis of the calculations, had the grace to admit that the explanation put about  for the previous 75 years  had to be wrong; the benefit was all about the redesigned gear being able to operate in shorter cut off, compared to the long cut off/ low steam chest pressures allowed by the original design. I have corresponded with Wardale, but not on this subject.  

There is indeed no magic in 30% cut off; it is just that if you plot efficiency vs cut off, you find that the decrease is relatively small increasing from 15 and 25%, and then progressively begins to get bigger and bigger for each additional % cut off- 30% is an arbitrary breakpoint between ‘not much difference’ and ‘quite a lot’ of difference, and beyond 40%, ‘very big’ differences. As in Germany, drivers in this country often, for the practical reasons you describe, drove on part throttle, and longer cut off than they needed to, offending the purists who correctly said that the most efficient way was full throttle, short cut off; however, providing cut off was not much above 25%, any efficiency losses were marginal; test plant results confirmed this.

Poppet valves are another big discussion point; maybe this should be shelved.

On the point about it being cylinder back pressure, not blastpipe pressure being important for efficiency, I agree.  However, what the programme I am describing demonstrates is that to all intents and purposes the two are the same thing, or at very least, highly correlated.

With respect to your comments on the SSCs of the UP types at 5000 IHP being too optimistic, this is entirely possible! As I indicated in my comments, the input values I chose for superheat and blastpipe area were completely made up; I am looking forward to repeating the calculations with the correct values. The first column of my Table shows that develop 5000IHP at 100 mph is about 30%; the exact figure will depend not only my guesses about superheat and blastpipe dimensions, but also other guesses that I have made with respect to engine and valve dimensions. The allowance for feedwater heater on boiler efficiency is a bit of a botch; I am in essence assuming that the amount of heat from coal needed is about 10% less than that would be needed to produce the flow of cylinder steam in no feedwater heater were present; this is just another assumption that needs tightening up if one were to do more accurate estimates. No account is made of the steam required by auxiliaries. This may sound like a damning series of admissions, but remember my purpose was to illustrate what could be done, if the relevant data were available, not to produce accurate estimates, these will come later if the necessary data is available. Having said that, I do believe the drawbar efficiency estimates are in the right ball park, and the changes I alluded to with respect to speed realistic.

The same goes for you questions about resistance; the underlying formula for resistance I use does separate mechanical and rolling resistance as you suggest.  I think the approach used to define mechanical resistance, developed by a friend, is sound, and with data about the mass of parts of the motion, pistons, valves, journal sizes, spring forces etc, could be applied to US designs. I am not confident that my extrapolation from UK locomotives is sound; this is another big project to put right, but I think we are only talking of 10-20% deviations from the figures I quote, so  I am I the right ball park at least.

I am wondering, having written this, if it would not be better rather than to try to do a broad brush survey of many types with partial data as inputs, to take one example for which all the details are known, and for which there is test plant data- e.g. the Niagara as mentioned, just to see how applicable the models are to US practice. It all depends on the data which is available.

Feltonhill, March 26th 2.01pm Good to know the T1 and Niagara reports are still around- how can I get hold of them? This would go a long way to answering my query. I might be able to source the Santa Fe book over here. I got to know Phil Atkins well when he was librarian at the NRM, and if he still has the same e:mail address, I’ll see what he can add. (He was the one who hired a pick up truck and went to the Rugby test plant just before it was bulldozed, and saved all the raw data and correspondence associated with UK testing- probably the only surviving raw data we have). I’m not sure what gives the impression that I have used total evaporative heating surface- the surface area referred to is that of the grate.

With respect to combustion chambers, there were some attempts to incorporate these into UK fireboxes, but not I imagine of the size possible in the US; I’m not familiar with the theory but suppose the idea was that the more space you have the greater the time for combustion to complete. Now, Orsat analyses of smokebox gases showed that it was rare on UK designs for combustion to be incomplete, except with respect to unburned particulate coal, so if they did have an effect on combustion it would be to burn some of these small particles; however, given that the extra lifetime of these particles in the firebox would be small, I am not sure there would be much benefit- that’s where the US data would be most helpful.

The Professor who produced the first principles model of the engine began by trying to do a first principles model of the boiler. He abandoned the attempt because he felt he couldn’t do a proper treatment of radiation vs convection- if it wasn’t robust, he didn’t want to know. I think there is an even more fundamental problem that he didn’t address, namely the loss of unburned coal from a lump coal firebed. You can measure it, but as far as I’m aware, no research was ever done to try establish the principles involved (there is some work quoted by Wardale, but I think this is the wrong concept).

Your description of the changes to the Y6 fireboxes is of great interest to me, because there is something about US boiler design I don’t understand. In this country, 19th century engineers believed that a long distance between tubeplates was a good idea to maximise heat capture. When superheaters were added, this meant that whilst the steam was hot near the firebox tubeplate, as the flue gases cooled along the length of the flues, they actually began to cool the steam in the superheater! Bad news. Boilers tubes reached a maximum of 22’ here; it took about 30 years to figure out that if you wanted top superheat, it was inadvisable to go above 17’.  And, in fact if you work out how much residual heat there is in the flue gases, the extra amount you extract between 17 and 22’ is pretty small anyway. So, looking at the massive size of some US boilers, how, I ask, did they manage to get top superheat levels?

Your SSC figures for the UP types give about 15.6lbs steam/hp-hr; with the original assumptions I used I get 6000IHP at 70 mph in ca 33% cut off, 89600lbs/hr to the cylinders: at this stage this is remarkable agreement! So we’re not too far apart, though I’m a bit better at 14.9 lbs steam/ihp-hr; from what I read, my blastpipe pressure of 19.7 is way too high, which seems to suggest the superheat is not as good as I suppose.

Lars Loco 26th March 5.34pm

Interested in how the discussion on 100000lbs/hr pans out- see my comments above on ‘normal’ vs  ‘maximum possible steam rates’

Lars Loco 26th March 6.26 pm

I’m more than happy to share the engine efficiency programme with you, if you can suggest how I can get it to you; as far as I know Microsoft won’t even let you send zipped .exe files now. It’s only 500KB!

I can also send two published papers from its author, describing its fundamentals; there is also one from me describing how I went about applying it, though I have to say this is rather long! I should also say that I’ll send you a 1 sider on what’s wrong with it- its author would have wished for nothing less than complete honesty, but this is not to undermine its value. It is in my view a complete game changer in our understanding of steam engine efficiency, provided it is handled with care.

Paul Milenkovic March 27th 12.53 am

Thanks for your supportive comments; as outlined above, you are correct about the gist of what I am trying to do. I have Wardale’s book and it is indeed a treasure trove. The point about Wardale  however is that he is very proud of the fact that he designed his 5AT using only techniques which were available in the steam age! He has used the programme referred to above to check out his calculations on the 5AT, and found good agreement, though I have the suspicion he prefers the old ways!

Thanks again for all your responses; 9 in 24hours; that’ll teach me to go off enjoying myself doing other things!

  • Member since
    April 2003
  • 305,206 posts
Posted by Anonymous on Monday, March 28, 2011 1:57 PM
However, maybe he could have had a chance to fly over the bridge with a "J", if available that time.
Faux Pax : How I could dare to mess up C. Chaplin with a Buster Keaton Movie. Sorry about that...It was Buster, of course.
you have started one of the most interesting threads on the web, I have long time looking for.
Thank you for sharing and presenting all your interesting data.
Will send you a private mes. for solving your .zip problem, but I do not see any reason for limiting data files to 500kb.
Just try to send it some way, any package, program or file-extension is fine to me.
If you have correspondence with Mr. Wardale, ask him about his dream steam-engine from the past, to "Wardale" it and improving it to next stage- 
he will answer you "The Allegheny". 
( source Sorry link is obsolete and unavailable now, from German magazine )
Hope to help you for input-data, I can scan you the 1943 Big Boy tests results from Kratville, if you do not have the book.
They show in a some nice way, how efficiency goes down, when the engine is hard working with a heavy a train.
Their output looked OK when they were handled as originally suppossed, to take a 3600tons train. 
However, long time seeking for an answer, how they did actually handle 20% heavier trains, as a 1948 showsTT shows.
  • Member since
    September 2010
  • 17 posts
Posted by NM_Coot on Monday, March 28, 2011 8:36 PM

In addition to Farrington, The Big  Three, you might want to check Brasher, Santa Fe Locomotive Development.  There is quite a bit in it but not in easy to use format.  Another series that might be useful is the "Loco Profile" series published by Profile Publications.  There are issues on NYC Hudsons, Nord pacifics, LNER A4, the American 4-8-4, and many others.  The ultimate of course is La Locomotive a Vapeur, by Andre Chapelon.  There is an Engilsh edition translated by G W Carpenter.  If anyone knows high efficiency steam, it is Chapelon.

I think it was Churchward that said he could improve the efficiency of any engine by 10% simply by painting the funnel blue.  His point was that day to day operation depended much more on correct operating procedues than pedantic design detail. If a crew believe they are being closely watched, they'll do things right.  And of course what's best depends on whether you want speed, drawbar effort, water economy, simple maintenance, etc.

The objective of the boiler is to produce steam faster than any combination of engine parameters can consume it.  The objective of the engine is to produce an effective output using all the steam the boiler can produce.  To pass great quantities requires large cross-sections through-out the exhaust path.  Higher superheat gives more fluid steam.  Numerous designers got the passage part right.  Fiddling with just the exhaust outlet is only a part of the answer.  High superheat in daily practice really is more a matter of cost and reliability than it is pure engine efficiency.

Having said this, few steam locomotives operated near their maximum for any length of time.  The NYC Niagaras were great but for day after day long, high speed runs, you have to give the blue ribbon to the ATSF 4-8-4s.

  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Tuesday, March 29, 2011 11:54 AM

Dear Lars,

Thanks for these comments. Where should I look out for the private message?- this is new to me. I would be very intertested to see the Big Boy data when we have established direct contact

  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Tuesday, March 29, 2011 12:44 PM

Dear NM Coot,

Thanks- understand and largely agree with what you write. Will check out the reference if I can get it. One of the first books I bought on US railroading was Duke and Kistler's photographic treatise on the Santa Fe in California, and I remember being very struck by the fact the 4-8-4s ran the 1765 miles from Kansas City to Los Angeles, over as many obstacles as you would care to mention. Inspired by the photograph of 3777 climbing Cajon Pass with 11 vehicles unassisted, I had a go at producing the power map for a 29xx class, again using guesses for many key parameters (Seems I can’t post this Excel graph)

I then got a crew with a miraculous ability to work at a constant 65000lbs/hr to the cylinders to drive 2926 all the way from Seventh Street, San Bernadino to the Summit, using this map. This is what I got:



Tare wt


gross wt


Engine wt


total wt



instant ihp

speed mph

total time

av speed

av ihp

av edhp

av dhp

height ft

av grad


San Bernadino






Seventh Street












































Old Keenebrook






















































It’s so long ago I can’t even remember where I got the heights and distances from- probably off Google Earth, so it’s all very approximate, but this is another illustration of where I’d like to get to.


  • Member since
    February 2005
  • From: Vancouver Island, BC
  • 22,231 posts
Posted by selector on Tuesday, March 29, 2011 5:06 PM


Dear Lars,

Thanks for these comments. Where should I look out for the private message?- this is new to me. I would be very intertested to see the Big Boy data when we have established direct contact

On the index page, where all the forums are listed, nearer the top you'll see a clickable post from anyone who wishes to start a conversation with you.  Similarly, if you click on the bold username in any text message posted here, a page will open and one of the options will be to start a conversation.


  • Member since
    June 2008
  • 121 posts
Posted by Dreyfusshudson on Thursday, April 7, 2011 5:14 AM

Thanks for all your replies. I’m now getting some data, and have reports on the PRR T1, Q2, NYC S1a S1b and S2a, and C&O H8 to hand, and have ATSF and UP references on order. It will take me some time to digest and analyse everything, several weeks, if not months I would have thought.  I will report back then for those who are interested.

An initial look see at the data says that on Porta’s critique ‘US exhaust design could not have been worse’ there is a case to answer; at top steam rates of 125000lbs/hr the PRR Q2 had an exhaust pressure of 31 psi, blastpipe area about 58 sqins; the Niagara already a back pressure of 10 psi at 52500lbs/hr, 44 sqins blastpipe, which means that at the quoted rating of 6000IHP the blastpipe pressure would be approaching 40psi!  If it were possible to reduce that figure to 5 psi at the same steam rate (and I’m not claiming it is), you would get about another 1200IHP for ‘free’. The T1 had a back pressure of 25-29 psi at 100000lbs/hr which implies a blastpipe area of less than 50 sqins; star of the show so far is the C&O H8 which showed around 13psi at the same steam rate.

Now these backpressures look horrendous to European eyes, but there are mitigating factors. Firstly the % loss of efficiency you get from high back pressure decreases as MEP rises, and US MEPs it seems were higher than elsewhere. Secondly, these figures, apart from the H8, are probably at steam rates, which whilst achievable on test may not have been called for in daily service; remember to maintain 100mph with 1000 US short tons on the level needs only around 5500 IHP from locomotives of this size, exhaust rates less than 80000lbs/hr because of feedwater heater recycling. So an important piece of contextual evidence is what the steam rates required by the most demanding schedules were. If logs of actual running are available one can work back to required evaporation rates. (I’ve just piloted this approach by driving a brand new T1 + 1000tons from Crestline to Lima in 46 minutes, as one of its forebears did; all works well- the locomotive was not flat out. The residents Upper Sandusky were taken by fright as it passed at 111mph at the foot of the downgrade; most didn’t know that something approaching this used to happen several times as day. Lawsuits were threatened.)

If one were to set a design goal of a back pressure of no more than 5 psi at some maximum design evaporation rate- let’s say 800lbs/sqft/hr, i.e. before boiler efficiency really plummets, then a few calculations suggest that a blastpipe area sqins of 0.887* grate area in sqft- 2.14 would be required, so the target for a Niagara would be 86.5 sqins, for the Q2 about 106sq ins, about double what they had; relaxing the criterion to 10 psi gives about 60 and 75 sqins respectively. The Jabelmann exhausts on the UP 4-8-4 and Challenger, should offer a significant improvement on the Niagara and Q2 respectively, so getting these areas is a high priority to understand what US best practice was.

Exhaust design is historically a trade-off between steam raising ability and back pressure; in the UK, and I guess everywhere else, steam raising won out over backpressure i.e. efficiency every single time. This is the real constraint when aiming for lower back pressure.

  • Member since
    July 2008
  • 648 posts
Just some notes on draughting
Posted by Juniatha on Friday, April 8, 2011 12:42 AM


@ Dreyfusshudson


A few more notes

        quotes in >> French Croissants <<


– 1 –

>> Now these backpressures look horrendous to European eyes <<

Umpfhh - erh, do they, really? well, maybe to eyes in 'Old Europe' as Rumbling Rummy would have said ;-) From your previous postings I thought you knew these back pressures since originally you stated that the formulae you work with would allow to predict performance (basically) from just three parameters and back pressure was one of them – although not precisely defined where measured?


2 –

>> .. Niagara already a back pressure of 10 psi at 52500lbs/hr, 44 sqins blastpipe, which means that at the quoted rating of 6000IHP the blastpipe pressure would be approaching 40psi! <<

That compares ‘back pressure’ against evaporation on one hand with ‘back pressure’ against cylinder horse power on the other. Back pressure essentially depends on mass rate of steam flow and steam temperature for a given design of blast nozzle. It is not the same for different shapes and designs of draughting blast nozzle(s) of given numerical free cross section tip area at given steam flow! As I remarked on the Porta statement: draughting could have been a lot worse in that poor dimensioning would have produced much higher back pressures with same numerical cross section nozzle tip area or produced much less draught at same given back pressure. To make it plain and simple: Porta’s statement, as I had modestly let on before, was characterized by his eagerness and love of exaggerations – thumps down for anything FGS / thumps up for his SGS. In this context, mind that given parameters were severely constricting design of draughting arrangement with large engines and inevitably it got worse with larger diameter of boiler within a given loading gauge!

Further, an extrapolation of back pressure from a given lower steaming rate to back pressure at an upper steaming rate about double the former, first of all needs to take into account the specific flow coefficient and the pumping efficiency curve of that draughting arrangement – only to remain a vague enough approach with an exponentiated rate of inaccuracy as the quotient higher : lower steaming rate increases. Again, there is a decided difference between round nozzles of plain ordinary type and De Laval type as in the Niagara! Not to mention effects of a – properly designed – Kordina! Without knowledge of the design specs and behavior of the draughting this numerical extrapolation will lead astray. That is probably why your estimate back pressure (40 psi) for the upper steaming rate is way too high. The draughting arrangements of the NYC S-1b vs PRR T-1 (non-regarding the Q-2) were not directly comparable for several reasons and showed characteristic differences in performance. The UP Jabelmann double arrangement of the 4-8-4 (again the Challenger and BB should not be considered in this comparison) was as inferior in some aspects as it was superior in other ways.


3 –

>> .. C&O H8 which showed around 13psi at the same steam rate <<

Ooops ?? where did you get that figure from? Without considering very special interfering conditions this would appear an exceptionally low figure, actually unwarranted. In SE Mallet type locomotives there is a long exhaust line from the second drive unit cylinders to the draughting, therefore it would be all the more important to know exactly where this ‘back pressure’ has been measured. Principally, for a number of reasons inherent with basic arrangement of live and exhaust steam conduct in these engines as they were, back pressures on pistons were rather higher – not lower – than in ‘straight’ engines i e of but one drive unit with cylinders under the smoke box. In SE Mallets, what percentage of back pressure on piston would then have been left as pressure at blast nozzle(s) highly depended on the degree of free or not so free flow of exhaust steam to that nozzle(s) and could indeed vary widely.


4 –

>> Firstly the % loss of efficiency you get from high back pressure decreases as MEP rises << and

>> US MEPs it seems were higher than elsewhere <<

Why, actually it doesn’t. Again, it’s important to distinguish between back pressure on piston and ‘back pressure’ somewhere in the exhaust line or pressure at blast nozzle!

For an increment of exhaust pressure base line in the cylinder, for example from 10 psi to 20 psi, including consequential higher compression – provided the higher exhaust pressure base line does not lead to over-compression! – the absolute amount of lost energy conversion is directly proportional to the diagram area defined by the two exhaust pressure and compression lines from exhaust lead opening to admission lead opening. Never mix intake and expansion processes with exhaust and compression processes. In the Walschaert’s / Heusinger valve gear characteristics on both sides of the piston are mechanically interconnected but that does not say losses on the exhaust side could be compensated or downsized by beefing up on the intake side.

In fact, evaluating progression of steam pressure on the intake side during admission and expansion has to distinguish between depression by throttling or drawdown by expansion. As I understand, in paragraph three of your reply to me, you postulate valve gear characteristics while not irrelevant are of but small influence. Next, you offer the known example of the Gresley A1 valve gear redesigned for the A3, where, to quote >> well told story .. about how changes to these settings led to a massive improvement ..<< were proven by your program to have been >> wrong << and instead >> the benefit was all about the redesigned gear being able to operate in shorter cut off compared to .. the original design. << Well, now that exactly was the difference the redesign (basically from short lap to long lap) did make and that was what yielded higher thermo-dynamic cylinder efficiency in the A3 and in the A4 further improved along the same line of development, allowing higher outputs in the upper speed range, and higher rpm speeds the A1 did not reach. So, it did make quite a difference. After all, it is known the original A1 proved a slouch, frankly, when in the exchange trials it was compared with the Swindon Castle class four cylinder simple – which by itself certainly was not too inspiring in thermo-dynamic cylinder performance with its valve gear mounted between frames and derived piston valve actuation on outside cylinders (quite the opposite of what would have been mechanically preferable) and with Churchward’s cherished low degree superheating.

However, as you acknowledge, to quote from paragraph four in your reply to me, >> beyond 40 % ‘very big’ differences << we would agree there is a decided loss in cylinder efficiency at c/o ~ 50 % or over as often found in large American loco types at flat out performance. High >> US MEPs << were thus no protection against losses by elevated exhaust back pressure. On the contrary long c/o caused losses by truncating expansion rate. To claim with these high mean cylinder pressures, increased piston back pressure didn’t matter much would mean to suggest when driving a Big Block Chevy of the early years of unleaded gas and catalyser, a badly tuned 4 barrel carb doesn’t make much of a difference in mpg.


5 –

>> If one were to set a design goal of a back pressure of no more than 5 psi at some maximum design evaporation rate- let’s say 800lbs/sqft/hr, i.e. before boiler efficiency really plummets, then .. a blastpipe area sqins of 0.887* grate area in sqft- 2.14 would be required, so the target for a Niagara would be 86.5 sqins, for the Q2 about 106sq ins, about double what they had <<

You never seem to account for pumping efficiency of draughting and for boiler characteristics, incl grate and firebed resistances. Principally, with draughting arrangements as they were, already compromised in proportions by inevitable restrictions of loading gauge, just to increase the cross section of blast nozzle(s) would have lead nowhere. This sort of ‘re-draughting’ would simply have repeated faults that had been found in the German Reichsbahn standard types intendedly equipped with Wagner ‘wide’ blast nozzles in an effort to minimize piston back pressure. While with these comparatively light designs in contrast to the massive American engines here considered the desired design aim was reached tolerably well, effect on draughting left something to desire. A suggested ‘Super-Wagnerization’ of American engines would have taken out draughting to a much more serious degree and reliable steaming could no more have been sustained even at moderate steaming rates!

Although steam locomotives weren’t too complex as compared to more modern engines, they represented the type par excellence of inherently interweaving, interrelating design characteristics in most all of their components and functions. That is why no item could be modified without effects on several related elements and according changes needed to be made to the whole chain of functions if any real advantage was to be realized. That was one of the underlying basics of steam locomotive design obviously not fully appreciated, to put it mildly, by many designers, engineers and craftsmen. It may well be so your formula would predict an increased performance by super-enlarging blast nozzle(s) – yet in reality the contrary would have happened.


6 –

>> Exhaust design is historically a trade-off between steam raising ability and back pressure; in the UK, and I guess everywhere else, steam raising won out over backpressure i.e. efficiency every single time. <<

If the first part aims at balancing draughting arrangement between lowered piston back pressure and increased draughting over a wide range of steaming rates, then there is something in it, if considerations are strictly limited to single, round plain nozzle, single round chimney type of design – and then it still lacks to take into account boiler design in view of gas flow resistance.

As for the second part: on British Railways steam locomotive design in the 1950s had been decidedly advanced by Riddles in general – famed for his 9F Decapod were ‘F’ actually proved to stand for ‘fast’ as much as it was meant for ‘freight’ – and by S O Ell developing his theory on draughting, enabling a couple of engine classes formerly notorious for bad steaming to shape up with new draughting arrangements fitted. In France, SNCF equipped a large number of engines with more or less sophisticated draughting arrangements and attained – more or less – substantial improvements in reaching both aims at the same time: good steaming and low back pressure. In Germany, with introduction of standardized design of Reichsbahn series, ‘soft’ draughting intentionally preferred low back pressure at the expense of compromising steaming rate ceiling; the rather extreme original proportions were set right on DB by the 1955 re-draughting campaign while on DR improvements to grate, air conduct to grate and ashpan design were introduced with or without Reko type combustion chamber boilers which proved steaming well with Wagner type of draughting because their characteristics were especially designed to suit. This did not mean the same boilers wouldn’t have steamed even better with more advanced draughting arrangements but this never materialized as the Giesl engines were used to burn low grades or indifferent mixtures of coal.

What should have been expected of a reboilered 5035 or 5280 class light Decapod with high efficiency draughting can be concluded from CSD 556 series Decapod of sensibly similar leading dimensions if at 99 t compared to some 90 t of the DR classes. With a very similar if slightly larger boiler these engines equipped with double Kylchap attained 3000 ihp. The Czechoslovakian railways post war two and three cylinder S E engine series by Mares, were of both carefully evolved external lines and highly competent internal design, combining – if given suiting quality of coal, which was rare during most of their time in regular service – high specific power output with comparatively very good cylinder efficiency.

It would be interesting to try your formulae on the 556 versus 5035 or 5280 classes to see how improved features in the former's design numerically nearly identical in leading dimensions will show up.


7 –

>> Secondly, these figures, .. are probably at steam rates, which whilst achievable on test may not have been called for in daily service <<

Oh my, on exacting never-relaxing Union Pacific, in fast freight services these engines were usually run flat out at anything test runs had proven them capable of – and then some if conditions were favorable. After all, there were many small adjustments made over the years and incessantly things were being optimized, all in view of improving reliability and capacity of every day revenue service – so, maximum output was very much the daily challenge of performance asked!

I bet-cha!




 Edit: my usual odd words and formatting .. plus seven sentences clarified, one intro added - April 9th

  • Member since
    September 2010
  • 17 posts
Posted by NM_Coot on Friday, April 8, 2011 8:06 PM

In all the discussion on this topic I'm really surprised to so rarely see the name Chapelon mentioned.  His designs and performance standards really exceeded everything being discussed in this thread.  The ultimate steam locomotive, 242.A.1, was a rebuild of a poor performer and comfirmed Chapelon design principles and showed what could be possible in the future.  Both his original designs and his rebuilds of poorly performing engines gave dramatic increases in power with significantly reduced fuel and water consumption.  In North America, compounding had a horrific history and would never be accepted.  That left 1, free exhaust clearance, 2, soft but effective exhausting, and 3, relatively high super heat to increase steam fluidity.  I previously mentioned Chapelon's own book but there have been others that give considerable detail for both test bank and field trials of his locomotives.

Imagine a steam locomotive capable of continous output of 4000 hp at the drawbar at 50 mph on the test bank and 4000 to 4200 hp on service trains.  Total weight was 148 tons and adhesive weight was 84 tons.  Boiler pressure was 292 psi and a triple klychap exhaust was fitted.  The grate was 54 sq ft.  Because of the three cylinder configuration the locomotive showed no tendency to slip even at maximum outputs.

if you want dream steam imagine a Chapelon-designed 4-8-4 to the North American loading gauge.  (Or better yet, a Chapelon-designed Garratt 4-8-4+4-8-4...)  No North American locomotive would even come close either performance or economy.

  • Member since
    July 2008
  • 648 posts
more remarks on draughting, engines and engineers
Posted by Juniatha on Sunday, April 10, 2011 4:26 AM

Three men, two 4-8-4 and one valiant Mikado


Ok, NM Coot, you put attention on André Chapelon – and rightly so: no discussion on steam locomotive efficiency could be complete without due consideration of his work. So far, I for one had kept my comments strictly limited to the initial topic of this thread, more precisely: the question in mind if locomotive performance and efficiency can be ‘predicted’ (questionable wording in view of ol..erh, classic! engines) respectively, can it be assessed closely enough and with tolerably correct correlations between various types of engines? So far, I didn’t want to extend discussion onto ‘how to optimize classic steam loco efficiency?’ 

In my remarks on draughting I was already hinting it  would actually take more than just widening blast nozzle area if you wanted to lower back pressure at cylinders while to ensure gas pumping capacity always remains slightly above demand of complete combustion (i e  λ ≥ 1.0 up to nominal boiler steaming rate, wherein ‘nominal’ is somewhat relaxed from maximum (like nom ≈ 80 to 85 % of max) if ‘maximum’ is correctly defined by grate limit, not by premature front end limit or by secondary considerations).

The Kylchap, often used in double arrangement on larger European engines, for a long time featured the highest pumping efficiency with neither Bulleid, nor Ister, nor Kiesel, nor Jabelman multi nozzles types coming close in performance and extent of efficient working range. In post WW-II years, an at least equally high if not superior efficiency was claimed for the Giesl ejector with Adolf Giesl-Gieslingen himself pointing out simplicity of his device, its ease of installation and (claimed) self-adjusting and the advantage gained by slimline shape of improved access to smoke box tube plate for cleaning and maintenance. A direct comparison of performance between Giesl ejector and double Kylchap on two engines of one suitable class of locomotive should have been interesting but I have no knowledge of it ever having been done; testing a double Kylchap 556 against a Giesl 52-80 light Decapod would have come closest – likely however, no comparative tests have been made, or I haven’t heard of it, since the classes were of different, if neighbouring, state railways: DR (Deutsche Reichsbahn) and ČSD (Československé Státní Dráhy). Drawing conclusions from what I know of relative performances of these locomotives it would appear it’s a close cut with the Kylchap 556 almost as much heavier as superior in maximum power output at speeds above some 40 mph – yet in absence of established data it is left to conjecture how much of that must be attributed to somewhat larger if closely similar boiler (equipped with stoker firing) and more efficient cylinders of slightly smaller relative volume (556 class b p of 256 psi against 228 psi was only 95 % of pressure for same t e with 21.7” x 26” as compared to 23.6” x 26” cylinders and identical 55” wheel diameter, disregarding difference in adhesion mass).

Later, Livio Dante Porta evolved the Kylchap into the Kylpor and from that derived the Lempor type, involving a Kordina and vortex or swirl flow of steam from blast nozzles. As by diagrams put up by David Wardale with these types of draughting devices even higher pumping efficiency can be reached. However, implications of these techniques must be considered with some caution. These highly effective draughting devices may lead to heftier than ample air aspiration through grate and firebed with abortive rather than supportive effects to continuous up-keep of combustion. Basically, these devices were developed with regard to running a thick firebed as used with gas producer combustion, involving a high degree of secondary air induced above the firebed.

Wardale has described intricate teething problems with his rebuilt 19-D 2644 to keep the firebed from being partially lifted and tumbling towards the front of the grate. I could imagine the typical railroad steam testing department having given up after a number of runs to just dump that tricky matter including rebuilt engine(s). While a comparatively steep incline of grate in the 19-D class certainly was not supportive of thick firebed plus high draught operation, Wardale identified several further adversely influencing factors and by scrutinizing sequences of testing / analyzing as well as committing co-workers and crew he was able to sort things out in the end. By reading his account on the testing period I got the idea these late hour advancements while holding a promising potential to increase both performance and efficiency of coal fired steam locomotives have much remained in a prototype stage of development (in spite of a time of regular running on the quite special Rio Turbio) and are far from simple to realize, keeping in mind Wardale’s noting of continuous and close support by Porta. This is not the sort of technique to put into effect hands down by your average preserved engine group. In the end, goals very much were accomplished with 2644 ‘Spooky’ and 3450 ‘Red Devil’ – in view of time then running out fast for steam on the SAR during the very years of rebuilding and test running these engines, they should be considered about as much of a success as could be realized under ruling conditions. While his superiors choose to remain blind for reasons obvious, I can read between the lines how reassuring it must have been to Wardale when the driver of 3450, ascending a bank at exceptional speed yet on shorter than usual cut off, in plain words summed it up by stating “there is a difference!” The engine certainly performed inspiring.

Much could be commented on the admirable 242.A.1 – a rebuild prototype to a proposed new family of types, an engine that stood all road tests, shattering established marks of steam traction top performance in France on any mainline tried and even beating electric traction on two runs Paris – Le Mans, arranged at a time when plans for Chapelon’s proposed family of new steam had been shelved and construction of the 2-10-4 had been cancelled in pretense of an alleged shortage of coal which was more likely imagined by politics than real, transient at any rate and concerning grades of coal used by steel industries rather than the railways. The proposed line of three cylinder compound 2-8-4, 2-10-4, 4-6-4 and 4-8-4 of 6000 ihp nominal output were to have many components in common, design incorporated many modern features of American practice which Chapelon had seen and appreciated during his visit to the United States.

As things turned out, this was not to be. However in the years of high hopes during tune up at Vitry and trial running the 242.A.1 attaining 5500 ihp, the series of test runs on the Ligne Imperiale, the classic mainline of the old PLM (Chemin de Fer Paris Lyon Mediterranée) passed with flying flags with a formidable sustained speed on the long 1 in 100 ascent to Blaisy Bas by the first and only 4-8-4 in France – Europe’s most powerful steam locomotive – marked André Chapelon’s triumphant return to the railway of his beginnings, the railway he had initially started to work for because of his early admiration of trains passing by on the PLM line near his parents house in a town in the vicinity of Lyon. However, leading heads of the engine department had turned down his ideas of technical optimization of existing locomotives and so he had left the PLM and ‘went west’ so to speak, to join the PO (C d F Paris Orleans), later PO-Midi, in the western part of France and covering hilly lines over the Massif Central. There, he found his proposals well received and in 1929 first materializing in Pacific 3566 massively rebuilt to his design. The story of success written by this engine and the following, renumbered 3700 series (later 231.700) is well known. Astonishing as these engines performed, they were topped when Chapelon even more radically transformed a number of smaller 73” wheeled Pacifics into 4-8-0s (4700, later 240.700) followed up by more rebuilts in the early years of SNCF, séries 240.P, shortly before WW-II. It was one of these engines that turned out 4700 ihp at continuous rating (real time steaming at constant water level) marking the highest specific power output per unit of engine mass to that date. It was with this series Chapelon engines started intruding the Ligne Imperiale as the smaller 4-8-0s vastly outperformed the large boilered yet cumbersome PLM Mountain types of various configurations of cylinders and drive axles of 79” and 71” wheels (PLM séries 241.A, .B, .D with prototype 241.C.1 and rebuilt 241.E).

With SNCF André Chapelon having been appointed head of the DEL (Division d’Études de Locomotives – locomotive development division) a new Berkshire type (pardon my expression) four cylinder compound of 5000 ihp was being considered. However with the onset of war, the PLM league had fought off what seemed hard to accept with established thinking and in view of austere times managed to bend the project into a Mikado basically intended to be an SNCF follow-up series of the PLM 141.D. Chapelon was then trusted with the task of turning a tolerably mediocre design into a good one – hopefully without changing all too much. In 1942 this gave birth to the 141.P series, the first lot allocated to high ranking Lyon-Venissieux, mainly for passenger service on the Ligne Imperiale. These were soon followed by more engines until later on early Lemaitre engines were exchanged for Double Kylchap engines and a number of the former got shelved serviceable in nearby yards (while lesser ranking sheds continued with old 141.D). The petite slender yet athletic 285 psi, 65” drivered four cylinder compound Mikado featured a steam flow arrangement reversed from the old 228 psi PLM design with l p cylinders now between frames while below the deflectors compact h p cylinders with large steam passages for live and receiver steam flow were set slightly angled to better comply with the angled position of the inside drives. Mechanically, the well balanced engine unit was designed for 150 km/h (93 mph), in actual service speed was limited to 120 km/h (75 mph). Power output characteristic differed from the original PLM Mikado as a superheated steam engine differed from its saturated predecessor: while output curve of the old 141.D leveled off at some 2000 ihp around speeds below 60 mph starting to fall off towards higher speeds, the 141.P curve kept climbing throughout the tested speed range reaching some 4000 ihp at 75 mph without attaining a maximum level – i e what had been a rather dull mule had been re-designed to become an exceptionally free running sparkling universal performer.

When the PLM mainline was electrified in the early 1950s, many 141.P were sent on a tramp never really ending to the last of their miles. With 318 engines built between 1942 and 1952 by several manufacturers – or ateliers, as railway workshops are sometimes called in France – during the main period of transition from steam to electric and diesel traction on the SNCF, from mid fifties to mid sixties, 141.Ps could be found at the head end of about any sort of train on the SNCF from miscellaneous freight to heavy express as well as any odd jobs on rural country lines while with electrification expanding and first class assignments getting lost to steam, increasing numbers became shelved with a cloudy future on doubtful tracks in the quietness of weeds detached from the busy centers of the roundhouse yards.

American built two cylindered 141.R proved better withstanding reduced maintenance and neglect as in the twilight of steam traction the formerly prevailing system of fixed assignation of engines of major classes to permanent crews had been replaced by common user system. The 141.P’s Achilles’ heel was their PLM design ancestry, mechanically out of date in certain ways from composite plate frame structure to plain bearings throughout or boiler with small firebox. ‘Les R’ as they were tersely called were less demanding in maintenance, yet reliable and thus ‘destinée d’écrire la dernière page de la vapeur sur l’SNCF’ – destined to write steam’s last page on the SNCF.

In his 1952 proposal of future high performance types of steam, André Chapelon was to conjoin the best of American practice in mechanical design with advanced thermodynamic design features – or in presenting his paper at the Arts et Métiers in Paris he offered exactly what was perceived as worst case by those who were already planning for steam’s abdication …







 front elevation of planned Chapelon / DEL 152.Q   drawing: SNCF / bearings added plus coloring: Juniatha



  • Member since
    September 2010
  • 17 posts
Posted by NM_Coot on Sunday, April 10, 2011 11:16 AM

Hello All


It wasn't my intention to divert the thread.  It was my intention to indicate that you can't consider either predicted or actual performance without taking Chapelon's work into account.  North American practice has always differed significantly from the rest of the world.  The rest adopted some of the best NA practices, but it seems to me the converse is not true.  I did try to indicate that what Chapelon achieved with 242.A.1 was simply a starting point for what could be achieved.  Other engineers offered changes with considerable potential improvements.  For example, Bulleid's concept for oil-bath valve gear was a massive conceptual leap but was let down by very poor detail work.  Having said all this, the classic Stephenson steam locomotive could never compete with either electric or diesel-electric performance or operational benefits.  But it is interesting to speculate on what could have been...

  • Member since
    March 2016
  • From: Burbank IL (near Clearing)
  • 12,063 posts
Posted by CSSHEGEWISCH on Tuesday, April 12, 2011 10:19 AM

It's time to throw a bomb into this discussion.  Chapelon's 242.A.1 may have pointed the way to improved efficiencies and performance, but that won't mean too much if such locomotives spend an inordinate amount of time being maintained.  French steam was alloted a lot more shop time than would be tolerated by a North American CMO, so the fuel/maintenance trade-off needs to be considered in any discussion of efficiency.  Vernon L. Smith, longtime Superintendent of Motive Power of the Belt Railway of Chicago, opined that North American steam locomotives were the best at what a locomotive should do by producing more ton-miles and passenger-miles with less down time than anywhere else.

The daily commute is part of everyday life but I get two rides a day out of it. Paul
  • Member since
    July 2008
  • 648 posts
Progress in engineering means to advance thermo-dynamic efficiency and mechanical design combined for increased performance and lower maintenance costs
Posted by Juniatha on Wednesday, April 13, 2011 5:53 PM


Well , I see what you mean . I agree with you as regards virtues of American Super Power steam .   Yet , may I kindly invite you to read my paragraph on why the 141.R out-lasted the 141.P on the SNCF ?   In a nutshell :as far as decisions were taken upon technical reasoning – there were numerous other considerations playing a part in which steam on which lines of which sheds etc had to go first – it was not regarding matters thermal efficiency in any which way , it was because of lower maintenance of the newer and more numerous , standard engine against the older and more traditional mechanical design in view of lesser assignments of steam where highest specific outputs were no longer needed .

The Achilles’ heel of the 141.P was their out-of-date mechanical design inherited from PLM predecessor 141.D .  However , composite plate frames and plain bearings of traditional design for instance were neither stringently correlated with nor sequel , effect or outcome of high (or low) thermodynamic efficiency .  As a matter of fact , the mentioned Chapelon engines all were rebuilds of old engines inevitably inheriting old fashioned concepts of construction , mind it !  In contrast , the engines he had planned were to be of the same standards of mechanical design as was then practiced in the States – so where , then , was the difference upon which to state those engines would have been more costly to maintain ?  Or in simple words : higher thermodynamic efficiency in cylinder work did not as such necessarily demand any extra expenses or complications in mechanical construction and therefore didn't have to change maintenance costs either .  A steam loco of higher thermodynamic efficiency wasn't costlier to maintain just because of its high efficiency .  Or in other words : you can't save on maintenance costs by sending consumption soaring .  And further : costs of running coal traffic for locomotive coal were one of the issues that stood against steam !  The higher fuel consumption for a given work , the more fuel needed , the more traffic for loco fuel , the more costs for loco fuel and for bigger tenders (which added up to locomotive total mass and reduced edbhp for same ihp [2] ) and so on .  Efficiency did matter in the steam / diesel competition years .  [1]

Coming back on the 242.A.1 – this also was a rebuilt engine .  Still the 241.P based on the PLM 241.C.1 prototype was chosen for building a final series .  This design concept was inferior both thermodynamically and mechanically at the same time , suffering for instance from overheating crank axle bearings and crank axle crackings and thus clearly was second choice in comparison to the 242.A.1 having a much sturdier single crank axle and no problems with it .  [2] So , where was the advantage gained for maintenance by sacrifice of excellent efficiency and performance for second choice ?  In fact , in determining decisions as to which designs were chosen there were other than technical reasons as I had already hinted with the pathway of the Berkshire / Mikado project . 

The American Super Power types certainly were magnificent engines for their time – no doubt about that .  However , to propose they could not have been improved and improved upon would mean to deny technical progress .   Let's not forget more than half a century has passed since these things were an issue .  So , we don't need to get too deeply involved in all that , I don't think there is too much in discussing matters all over again .  We will never find an undisputed lasting truth that would apply in general – too widely differing were the circumstances and the influences that intervened and played a role .  Who was to decide what was right and what was wrong in individual decisions ?    

There is romance in steam's lost case and we like to ponder it and contemplate what could have been – that's all .  We should never forget about our hindsight perspective .   

I've offered my view – no insult intended .  That said , I step out of this thread .  Don't get at loggerheads about this matter , folks – the locomotives never come back anyways .  [1]





Addenda :

In view of the latest post commenting on comparison of SNCF Mikado types 141.P and 141.R I feel obliged to disclaim what seems to be a miss-interpretation of my comment written earlier on this matter .   Inviting to re-read my text , I pointed out reasons why SNCF concentrated dwindled remnants of steam loco park on mainly one loco type , the 141.R , were largely disconnected from engine-technical considerations . Quite certainly , the fact a small number of 141.R engines outlasted practically all other types of steam on the SNCF was caused by vastly eased demands of engine performance with steam traction at that stage during the transition having dramatically lost mainline assignments – all first class train running handed over to electric traction and much of the more important secondary train running done by diesel traction , 141.P high performance standards were no longer needed .   To secure remaining few secondary and lower ranging traction services , SNCF therefore preferred to keep the 141.R Mikado because of universal presence on the system , younger age , more sturdy construction and simplicity in design , parts being readily available .   

No conclusion can be drawn on that in trying to find an answer to the age-old steam design issue ‘two cylinder simple expansion against four cylinder compound’ .   There never was anything like the colloquially alleged ‘mass engine competition test’ carried out on the SNCF and thus , logically , there could be no winner. The question which of these two series of  Mikados , 141.P or 141.R , was ‘more efficient’ could not be answered in such a widely undefined , general way – simply because engine characteristics vastly differed .   Looking at heavy express run assignments formerly steam handled , the ‘P’ could clearly outperform the ‘R’ ;  the ‘P’ could handle demanding runs the ‘R’ lacked both speed and power for – that left the ‘P’ without alternative in this class of train running .   At the other end , in low ranking secondary train traction where performance was of little or no importance , clearly the sophistication of a 141.P was not needed .   Yet , steam depots in the western part of France continued with their 141.P engines in such services as long as they kept steam running , never asking any 141.R for replacement .   Both ‘P’ and ‘R’ stoker fired , the ‘R’s coal consumption was vastly higher at both ends of performance range – at the upper end because of degraded combustion efficiency and lesser thermo-dynamic efficiency and at the lower end because of grate too large for this sort of service .   Yet , the ‘R’ was better digesting indifferent coal qualities because of larger grate and sharp draught .   In average freight traffic , there was an advantage with the ‘R’ that grew with increasing load and decreasing speed (until getting at a disadvantage against the 150.X , German 44 class , in what was a good working range of that three cylinder Decapod) – yet this was not so much a question of two against four cylinders as it was a profit gained by sturdier construction with cast steel locomotive bed, Franklin automatic wedges , roller bearing on axles , boxpok drive wheels and a very good boiler design for high steaming rate .   Beyond the obvious difference in cylinders , ‘P’ and ‘R’ differed drastically in mechanical design and this blurred any practical comparisons of maintenance costs as to seeking an answer to the old ‘two cylinder SE against four cylinder C’ issue .    So , in actual service at different regions and in various qualifications of service there could be all sorts of answers to this issue – even on the same railway system !   I think this explains why there never was and never could be a definite answer found or agreed upon in the perennial ‘two cylinder SE against four cylinder C’ issue .


Edit (lighter shades of color) :

sentences added [1]

sentences clarified plus two missing words added [2]

Addenda inserted on issue two cylinder simple against four cylinder compound [3]



Join our Community!

Our community is FREE to join. To participate you must either login or register for an account.

Search the Community

Newsletter Sign-Up

By signing up you may also receive occasional reader surveys and special offers from Trains magazine.Please view our privacy policy