• Why didn’t Alco run with the 539T?

  • Discussion of products from the American Locomotive Company. A web site with current Alco 251 information can be found here: Fairbanks-Morse/Alco 251.
Discussion of products from the American Locomotive Company. A web site with current Alco 251 information can be found here: Fairbanks-Morse/Alco 251.

Moderator: Alcoman

  by mtuandrew
Also makes sense why Fairbanks-Morse had some success with the 38-8 1/8 platform: lower stress on each crankshaft while still maximizing output and fuel efficiency. They were willing to trade additional complexity for what they saw as a potentially lower failure rate. I suppose that until pretty recently, metallurgy hadn’t reached the point of allowing big, big multi-bank diesels - weren’t crankshaft failures one of the reasons the SD45 and 45-2 became unpopular and the SD80 never made as many orders as desired?

That tandem 538 sounds very interesting. On its face it sounds fairly straightforward, just two engines in one block. A little deeper, it sounds like a nightmare of differential heat expansion, vibration control, gear breakage and equalized supercharging. It may not have been compact enough for most locomotive applications anyway, not with 1930s technology as a limitation. The more I read about it, the heavier it sounds - the Navy-spec welded block 540 must have been lighter, but also must have been more expensive to build in an era where Alco had extensive casting facilities. (Also see: likely why Alco didn’t forge its crankshafts.)

It looks like you folks have discussed the twin-bank 539 before, most recently in a thread about the 244A block (referencing a Rolf Stumpf list including the DL202-1 and DL203-1.) Seems like you came to the same conclusion, that Alco never built such a thing because the 241 and later 244 were available sooner & cheaper.
  by Pneudyne
Allen Hazen wrote: Sun Jun 14, 2020 12:53 am Why British Rail insisted on an engine design originally adopted for 1920s crankshaft metallurgy I'm not sure...
The Sulzer LVA24 vee engine was not available for initial use until c.1962-63, so it post-dated BR’s major fleet purchases. Thus to the extent that it wanted to use Sulzer engines for its higher-powered locomotives, BR was stuck with the double bank 12LDA28, which was effectively a post-WWII development of the late 1930s 12LDA31. Evidently Sulzer had done a good selling job in particular to the folks at the London Midland Region (LMR) of BR, to extent that the latter wanted to dieselize using only home-built Sulzer-engined locomotives. The two concepts were realized as the class 24/25 and 44/45/46 series. Also, Sulzer was offering more power in a single-engine (2300 hp in 1955, 2500 hp soon after) than any other potential supplier with in-UK building capability at the time, and this made it attractive for the fast passenger locomotive. The inception of the class 47 was a complicated exercise that from some angles looked to have involved a certain amount of finagling. By that time, Harrison, ex LMR was BR CME (Chief Mechanical Engineer) and it appears that he was bound and determined to have a Sulzer-engined locomotive built by Brush. One argument used was that the Sulzer 12LDA28 engine was tried and tested, and required only a modest uprating to go from 2500 to 2750 hp. There the problems started. As well as uprating, the 12LDA28C was subject to some lightening of the structure in order to meet the BR weight limit. This, combined with an increase in rotational speed, turned out to be a bridge too far. Various problems ensued, and the whole fleet had to be derated. One issue was balancing. The two crankshafts were too close together to allow the use of a full set of counterweights, and so full dynamic balancing, on each. Thus there was a tradeoff made between crankshaft phasing and the accommodation of balance weights to get the best compromise. This flags a problem with the double bank arrangement, in that a choice has to be made between close spacing of the banks with resultant sub-optimum balancing, or wide spacing for optimum balancing but with a resultant very wide and probably heavier engine.

The 12LDA28C was quite heavy anyway, at around 45 000 lb dry. That made it slightly heavier than its obvious competitor in the UK context, the English Electric (EE) 16CSVT, which was 43 500 lb. By way of comparison, the Alco 16-251B was 42 000 lb, and the GE 7DL-16 was at the time around 39 000 lb. The Sulzer double-bank engine was also relatively expensive, at £45 000 c.1961-62, as compared with £26 000 for the EE 16CSVT. In part that was probably due to the fact that it had two crankshafts, and the crankshaft is usually reckoned to be the most expensive single part of a diesel engine. Whatever were the perceived advantages of the double bank diesel engine in the 1930s, it turned out to be something of a blind alley longer term. So it probably would not have been a good longer term choice for Alco had it chosen to pursue it further.

The Sulzer LVA24 vee engine was developed by CCM-Sulzer in France at SNCF’s request. It would appear that this engine was never fully developed before Sulzer abandoned the rail traction market. BR did fit 5 new-build “47” class with the 12LVA24 engine in 1965, as the class 48, but evidently these were not entirely satisfactory, and were eventually re-engined with the 12LDA28C.

  by Allen Hazen
Thanks, Pneudyne! I had a feeling that I was "trolling" British Rail there, and it's good to have a corrective response.
--The LVA24 wasn't very much later: the first SNCF locomotives with it were 1962 or 1963. But BR wads in a hurry to dieselize, and so waiting one or two years was probably not an acceptable option.
--The website I mentioned (I'll try to find it again and post a link) had "articles" with a very pro-Sulzer viewpoint, so maybe should be taken with a grain of salt. Which said, one of them attributed the problems BR experienced with the LVA24 to a maintenance-related fiasco. Certainly SNCF was willing to take delivery of the LVA24 engines pulled from the Class 48 locomotives when they were "regularized" to Class 47, and I have so far seen no indication that SNCF was unhappy with the LVA24.
--Engine weights... are hard to find. Published sources (paper and web) that I've found so far ... aren't always clear whether they are comparing apples or oranges (in particular, one weight I found apparently included the generator, others didn't....). So thank you for numbers! (i) 45,000 pounds seems very light for a 12-cylinder engine with 1300+ cubic inch cylinders! One response to vibration problems (see the history of GE's efforts with the HDL and GEVO: I forget the exact weight of a GEVO-12, but it is well over 40,000 pounds) seems to be to add tons of extra cast iron to the frame, but that would not have been possible with BR's weight limitations. (ii) Are you sure of the weights you have for the Alco and GE engines? My impression is that an FDL is usually somewhat heavier than a comparable 251.
Thanks again! (I was secretly hoping you would comment.)
  by Allen Hazen
https://derbysulzers.com/12LDA28C.html ... has a narrative of the history of this engine on BR
https://derbysulzers.com/1702.html ... is about the LVA24 in the Class 48. (When I said the problems with the Class 48 were a "maintenance-related fiasco" I was summarizing and doubtlessly oversimplifying from this.
  by Pneudyne
Allen, as you say, good information on engine weights is hard to find. And when you do find it, it is not always clear whether it is a dry weight or a wet weight.

In this case I used the following sources:

Alco 251: A 1961 May Diesel Railway Traction (DRT) article “The Alco Engine and Its Development”. The 16V-251B weight was shown as 39 000 lb on an engine-only, dry basis. The 16-244 was shown as 40 000 lb.

GE 7FDL-16: A 1956 November DRT article on the GE/Cooper-Bessemer FVB engine. The FVBL-16T was shown with a 39 200 lb dry weight, which I took as a reasonable proxy for the FDL7-16 in the late 1950s/early 1960s timeframe.

I have also just checked the Jane’s 1969-70 listings, which showed 41 750 lb for the Alco 16-251E, and 44 500 lb for the 7FDL-16, indicating that there had been some beefing-up of the latter during the 1960s.

The Sulzer and EE weights came from the respective manufacturers’ brochures.

The Derby Sulzers site is excellent, and I think is the leading repository for all things Sulzer in the railway sphere. As you have observed, the article on the LVA24 is rather pro-Sulzer. The same article also appears at: http://vdmw.ch/joomla/images/Geschichte ... Engine.pdf. For example: “Retrospect is a wonderful thing and if these LVA24 engines had not suffered such catastrophic failures, it is almost certain that they would have been specified for the later Class 50 locos. In this respect, the licence for the LVA24 engine was bought by English Electric who used many of the design features in their CSVT engine and promptly dropped the Sulzer licence even though at their insistence, a 16LVA24 engine with a cast cylinder block, had passed a full 840 hour test in Switzerland!” does not exactly accord with the conventional history. The EE CSVT engine dated from the late 1950s, and was a relatively simple charge air-cooled upgrade of the established RK series (SRKT/SVT), developed in the immediate post-WWII period from the pre-WWII K engine. This alternative account by Tufnell might be closer to the truth: “In 1966 English Electric took over Ruston and Hornsby (including Davey-Paxman) and as a result plans were initiated to phase out the RK engine and to concentrate on two engine sizes. These were intended to be the Paxman 'Ventura' for the lower powers and a version of the new Sulzer LVA24 to be produced by English Electric for powers up to 4000 hp. Unfortunately, for various reasons this last project did not work out at all well, and after spending a lot of time and money the RK had to be hastily reinstated and uprated. This time a thorough job was made, and the new design emerged as the present Mk III version with an output of 262 hp per cylinder at 900 rev/min.” One change that EE would have made to the LVA24 was to use a cast crankcase in place of Sulzer’s fabricated type. The RK Mk III might have included some detail from the LVA24, but for the most part it followed the established RK pattern, with the major changes being those that had been planned quite some time back but not implemented for cost control reasons.

  by Allen Hazen
Thank you for your replies!
-- I have a later "Jane's" from (1986/1987): it gives a slightly higher weight for the FDL than the 251, though not differing by as much as the ones in your edition: 42,500 pounds versus 43,500 for the two V-16 engines, both weights specified as "dry". (I have a disturbing vision of a bit of dialogue at the manufacturer's office: A says "That sounds about right, but we really should check the specifications for the latest version," and B replies "It's only for an ad in "Jane's".")
-- I have the impressions that cast frames are a bit more robust and preferred for very high output engines, and that cast frames tend to be a bit heavier than fabricated. Are either of these accurate?
  by Allen Hazen
George Elwood's "Fallen Flags" rail image site has a GE engine service manual for the U25B. Date... difficult to determine. (Not surprisingly: "Typewriters" at the "Railroadlocomotives.blogspot.com" site has an article on GE manuals: GE, particularly when reprinting revised editions, didn't make dating easy.) It says 3/62 on the cover, and a footnote about firing order says it was revised in 8/62.
Anyway, as to weights, it has a list of weights "approximate, for lifting purposes only."
"Engine and generator complete" is given as 57,000 pounds.
"Main generator with auxiliaries" is given as 17,500.
But then "Engine less generator" is given as 37,500.
I would have expected the first to be the sum of the other two, and this is an official GE document!
  by Allen Hazen
Maybe there is more of a relationship between the unbuilt Alco twin-bank 539 derivative and the Sulzer LDA28 than we thought!
Page 231 of Steinbrenner's Also book says that Alco built a small number of Sulzer-designed stationary engines under license in the 1930s. (Not in itself surprising: Alco's Diesel engine department was McIntosh & Seymour, a diesel (and gas) engine builder of several decades standing, that Alco had bought in the late 1920s. It would have continued its participation in the non-locomotive side of the diesel business.) And then, during WW II Alco, "through its continuing relation with Sulzer," was recruited by the US Navy for an (ultimately abortive) project to develop a submarine diesel based on Sulzer designs, incorporating Sulzer components which couldn't be obtained from Switzerland until the end of the war. So there was a fair bit of communication between Auburn (Alco's McIntosh & Seymour plant) and Winterthur (Sulzer's headquarters)...
So the "prior art" in two-bank engine design that Alco's engineers were trying to improve in the early 1940s (cf. patent mentioned above) would at least have included Sulzer's pre-war predecessor to the LDA28 design. (The numbers in the Sulzer model number represent cylinder bore in centimetres: this was a 31cm x 39cm engine: about 10% bigger in per cylinder displacement than an Alco 539.). And maybe, just maybe, the investigations by Alco's engineers into ways of alleviating problems with the twin bank design were helpful to Sulzer's when they designed the post-war replacement.
Buchi, of turbocharger fame, was born in Winterthur -- his father was a Sulzer employee -- and worked for Sulzer at times in his career. So the Winterthur mechanical engineering community had multiple links with the Auburn one.
  by Pneudyne
Allen Hazen wrote: Wed Jun 17, 2020 2:40 am
-- I have the impressions that cast frames are a bit more robust and preferred for very high output engines, and that cast frames tend to be a bit heavier than fabricated. Are either of these accurate?

Stronger and heavier for the cast case I think aligns with the conventional wisdom. Nonetheless, those in favour of the cast crankcase would argue that accurate casting allows the metal to be placed exactly where it is needed, and not where it is of little benefit, something that is more difficult to do with the fabricated approach.

Be that as it may, it would appear that satisfactory results and broadly similar weight-to-power ratios are achievable either way, given good design execution and subsequent development.

Whether the relatively small weight differences between the Alco 251 and GE FDL engines stem from the respective choices of crankcase construction, or from the other significant design differences is hard to say.

Although cylinder size is a good starting point for comparison purposes, sometimes the same or cylinder sizes are used for engines whose intended power outputs are quite different. An example that comes to mind is that of the Paxman Ventura and Valenta high-speed engines. Both shared a cylinder size of 7¾ x 8½ inches. The Ventura was designed to match the German high-speed engines of the era, such as the Maybach MD655. The Valenta was of much higher power output, so was much more robust, and as far as I know had a larger bore centres dimension. Basis the Jane’s 1973-74 numbers, the highest output 12-cylinder Ventura engine produced 1500 hp for a dry weight of 11 190 lb and a length of 81 inches. The corresponding Valenta produced 2250 hp for a dry weight of 14 930 lb and a length of 96.75 inches.

In the ocean-going marine heavy medium-speed engine field, it has been customary since the 1990s, and perhaps before then, to categorize engines on bore size alone. The rationale was that the latest engines at least operated with mean effective pressures (mep) at the practicable upper limit, then around 25 bar but higher now, I think, and also at the upper practicable mean piston speed limit, around 10 m/s for heavy fuel and 11 m/s for distillate fuel (gasoil). Thus stroke length and rotational speed were interrelated, and an increase in the former meant a decrease in the latter, more-or-less self-compensating in power output terms. Stroke-to-bore ratios were around 1.2:1, with some variation, with stroke length chosen to achieve maximum piston speed at close to a 50 or 60 Hz alternator synchronous speed, given that the same engines were also used in stationary powerplants. Anyway, the net effect that bore size (along with cylinder count) was a good predictor of the power output range of a given engine type.

Returning to the Alco and GE case, these engines appear to have been aimed at essentially the same target. The GE FDL might have somewhat more development potential than the Alco, but whether that could be attributed to its cast crankcase is I think uncertain. Its articulated connecting rod arrangement may have been another contributing factor.

Applicable I think is the commentary provided in a Diesel Railway Traction 1960 November article which was essentially a précis of a paper by J.C. Rhoads of GE, presented at a Pan-American Railroad Congress in Brasil.

“Some of the more important criteria applying to a locomotive diesel engine were considered by Mr. Rhoads to be: (1) It should be of the V configuration for best locomotive space utilization and maximum maintenance accessibility; (2) It should have maximum h.p. per cylinder to provide the minimum number of cylinders and other engine parts for a given power output; (3) it should have as high a crankshaft rotational speed as possible to minimise engine and transmission weight; (4) Ratings should be based on long-term field experience for optimum life; (5) It should have utmost reliability; (6) It should have minimum fuel and lubricating oil consumption rates; (7) It should be able to idle for long periods without undue distress; and (8) The basic cylinder design should be offered in a multiplicity of cylinder arrangements for maximum h.p. range and maximum interchangeability between engine sizes.

“There is a high percentage of V-engine applications for diesel locomotives in the world. Top cylinder accessibility is excellent and space utilization good. Generally, 8.5 in. or 9-in. bore can be accommodated in a 6 ft. overall width with a 45 deg. V angle. Larger bores can be accommodated only with difficulty; smaller bores do not fully utilize the space. In this connection it is of interest to note that all three of the U.S.A. engines used for locomotives have this 45 deg. V angle and bores in the range just stated.

“Maximum output per cylinder comes from the maximum bore that can be accommodated in a given locomotive cab width in the V configuration. Based on U.S. practice a 9-in. bore appears to be about the maximum obtainable in four-stroke engines, and an 8.5-in bore in two-stroke engines. Output per cylinder ranges from about 140 b.h.p for two-stroke engines to 165 b.h.p. for four-stroke engines.

“Transmissions can be made smaller as engine speed is increased. Some engine design practice outside the U.S. includes speeds up to 1.500 or 1,600 r.p.m., but even at this higher r.p.m. the h.p. per cylinder may not be greater than American practice, since the bore and stroke are less. Consequently, more cylinders may on occasion be required to produce maximum h.p. per locomotive. American practice is to run four-stroke engines in the neighbourhood of 1,000 r.p.m., and two-stroke cycle engines at 800 to 835 r.p.m. Piston speeds are not currently so high as those of modern quick-running engines.”

I think that there is an element of circular argument in the foregoing, but even so the basic premise is valid, and supported by other data, particularly if the upper bore size is moved slightly, say to 10 inches. For example, Lima chose a 9.5 inch bore vee engine, which became the Cockerill CO240. By 1969, the 16-cylinder version provided 4000 cv (3944 hp) at 1050 rev/min for a dry weight of 18 000 kg (39 690 lb). English Electric stayed with its 1930s-origin 10-inch bore for its post-WWII vee engines, widely used in CM-gauge applications. EMD went to a 9 3/16 inch bore with its 645 engine. Sulzer chose a 240 mm (9.45 inch) bore for its late 1950s vee engine.

Thus it could be said that of the options reasonably available to Alco, it made the right choice in doggedly (and perhaps dogleggedly) pursuing the development of the high-specific output, 9 x 10 ½ inch vee engine.

  by Pneudyne
Pneudyne wrote: Mon Jun 15, 2020 7:33 pm Thus to the extent that it wanted to use Sulzer engines for its higher-powered locomotives, BR was stuck with the double bank 12LDA28, which was effectively a post-WWII development of the late 1930s 12LDA31.

Actually, the 12LDA28 might have been mapped out in the late 1930s, along with the rest of the LDA range. The complete range included engines with bore sizes 190, 220, 250, 280 and 310 mm. (7.48, 8.66, 9.84, 11.02 and 12.20 inches.) All were to be available in in-line 6 and 8-cylinder forms, with the 12-cylinder double-bank version listed for the 250, 280 and 310 mm bore sizes only. But the only double-bank engine Sulzer built pre-WWII was the 12LDA31, in a very limited way for two prototype double locomotives, one for PLM/SNCF and one for CFM Roumania. Post-WWII, and in accordance with the progressive uprating of all models, Sulzer evidently saw that power needs would adequately be addressed more compact and lighter 12LDA28, and although the 12LDA31 remained in the list for a while, it was never built again. Improved balancing for the range dated back to c.1939, in part required for Sulzer’s improved control system that used multiple closely-spaced engine speeds rather than the 3 or 4 previously used with the BBC-type control. The first 12LDA28 build, by Sulzer’s French associate CCM-Sulzer, was for the SNCF 060DA class, ordered in 1953, with first deliveries in 1955. As far as I know, the 12LDA25 was never built.

The only other true double-bank engine used in a locomotive that I know of was the Lister-Blackstone ERS12T, one of which powered the solitary BTH “Explorer” metre-gauge prototype of 1959. It was trialled and later bought by East African Railways, but it was not repeated. The ERS12T was an 8.75 inch (222 mm) bore medium-speed engine, with widely spaced banks. Either crankshaft could be uncoupled from the phasing gears so that the engine could be operated on one bank if necessary. As is self-evident, it was an overly complex way of powering a light, 1100 hp locomotive when 6- and 8-cylinder conventional medium-speed and light 12-cylinder high-speed solutions were readily available. I suppose though that Lister-Blackstone needed to work with the engine sizes that it had, and double-banking its 8.75-inch bore engine was probably the quickest and lowest development cost option, albeit a dead-end.

MAN produced c.1938 an ersatz double-bank engine for a PLM/SNCF double-locomotive prototype that was counterpart to the Sulzer-engine example mentioned above. This was effectively a pair of 6-cylinder 30/38 (300 x 380 mm) engines mounted on a common base and sump, but facing in opposite directions, and each with its own main generator. The two engines operated independently, with no rotating connection between them. This arrangement did not set a precedent for future development, though.

Returning to the locomotive medium-speed vee engine and the previously mentioned 9 x 10½ inch (+/-) cylinder size cluster, there were some less significant others in this group. The Mirrlees JV was intended as a multipurpose engine, aimed in part at a British Admiralty specification which nominated a 9.75 x 10.5 inch (247.7 x 266.7 mm) cylinder size and the use of a fabricated crankcase. It was used in quantity by British Rail until it was doomed by intractable crankcase fatigue problems. The Paxman YL was aimed at the same specification and had the same cylinder size. Its only locomotive applications appear to have been in Italy, where it was licence-built by Breda, who also modified it to have underslung rather than bedplate form. Also in Italy, the Fiat 230 series was a 230 x 270 mm (9.06 x 10.63 inches), 1000 rev/min engine.

Whilst one should be circumspect in making a hard inference of “inherent correctness” from the modality (in a mathematical sense) of the (generalized) 9 x 10½ inch, 1000 rev/min engine type cluster, nonetheless it is reasonable to attribute a higher weighting to prudent engineering choices than simply to happenstance. It could be said that Alco would having been playing a “long shot” for any choice other than the 9 x 10½ inch, 1000 rev/min vee engine.

  by Allen Hazen
At least one engineering team in the early 1950s seems to have agreed with you on the "best" size for a locomotive diesel. When GE was looking at going seriously into the diesel locomotive business, including engines, in the run-up to the end of the Alco-GE partnership, they studied a large number of Diesel engine designs from several countries... and decided that the one that seemed most hopeful was the Cooper-Bessemer F series!

Which said... something a bit bigger might have been possible. The Ruston RK, with 10"x12" cylinders, was used successfully in locomotives for British Rail, whose restrictive loading gauge would have made it harder to stuff a big engine into a usable carbody than it would have been in the U.S., and in locomotives for narrow (metre or 3'6") gauge railways in the Empire. And GE itself, in the 21st Century, has moved to the comparably large GEVO.

Re: "Improved balancing for the range dated back to c.1939". --- So, before the Alco design studies represented by the patent application PCook pointed out.
  by Pneudyne
I imagine that anywhere in the cylinder size range from about 8½ x 10 inches through 10 x 12 inches would have been workable.

Mean piston speed was another variable to consider. Back in the 1940s, the upper limit for medium-speed four-stroke engines was probably in the range 1700-1800 ft/min. If your desired rated speed was 1000 rev/min, then a 10½ inch stroke would do that at a piston speed of 1750 ft/min, which would have been acceptable. The RK, with its 12 inch stroke, operated at lower speeds, 750 rev/min initially, 850 rev/min for the Mk II version, 900 rev/min for the Mk III version in 1969, and reaching 1000 rev/min with the RKC in the 1980s. By that time, the acceptable speed had moved upwards to 2000 ft/min and beyond. Whilst bigger cylinders and lower rotational speeds tended to be self-compensating for power output, the latter meant larger and heavier main generators. Apparently though generators became more difficult to design as rotational speeds became faster (somewhere I have seen both English Electric and Alsthom pontifications on the matter; whether they remain findable I’m not sure), and around 1000 rev/min give or take might have been something of a “sweet spot”.

Re Sulzer and improved engine balancing, whilst its early work predated that of Alco, it was probably a work in progress, and quite possibly Sulzer used some of the Alco ideas as the 12LDA28 was developed during the 1950s and early 1960s, during which time the balancing was changed.

  by Allen Hazen
Re: "Whilst bigger cylinders and lower rotational speeds tended to be self-compensating for power output, the latter meant larger and heavier main generators. "
---Within limits, increasing the rotational speed of the generator will allow use of a smaller generator (saving weight and cost). On the other hand, gears are ... expensive and a potential maintenance issue, so usual practice is to have the generator directly connected to the engine, and so rotating at the same speed as the engine. The designers of the Class 47, stuck with gears anyway because of the engine configuration, managed to use a smaller generator than they would otherwise have had to by choosing gear ratios so the generator rotated FASTER than the crankshafts.
And sometimes there are economic considerations going the other way! In recent years, a U.S. locomotive rebuilder (Knoxville Locomotive, I think) has rebuilt EMD roadswitchers with a smaller (so maybe cheaper in first cost; as a modern design it is also better in terms of pollution than the Roots blown 16-645 it replaces) Diesel engine of the same power. (It's a descendent of the German design used, in the 1960s, on some Diesel Hydraulic locomotives.) It has a speed twice that of the EMD 645. On the other hand, EMD AR-10 traction alternators (and, I suppose, spare parts for them) are readily available and familiar to American railroad mechanics. Solution: 2-for-1 reduction gearing, so the generator doesn't realize it's connected to an 1800 rpm engine instead of the 900 rpm engine it was originally coupled to!