Alco diesel hydraulic

The Alco DH643 was built for the Southern Pacific. Here's a link to the Wiki article on it. https://en.wikipedia.org/wiki/ALCO_DH643" onclick="window.open(this.href);return false;

I lived near Roseville, CA and saw them many times (mostly stored, but also in action). Odd looking ducks, like all Alco double diesels. I guess they were more successful than the KMs since they lasted longer. They typically ran up and down the Central Valley MU'ed with one small unit (like in the picture), I assume to get the train into a siding if the Alcohaulic broke down.

The Wikipedia article says
"Dissatisfaction over the poor performance of diesel-hydraulic locomotives"
was one factor leading to their early retirement. Which is pretty frustrating, since it doesn't tell us HOW their performance was poor. Was it a matter of poor reliability, or was there something wrong even when they were working right?
--
One of the claims made for diesel-hydraulic locomotives is that they are lighter in weight than diesel electrics. Comparing otherwise similar d.h. and d.e. types on British Rail (there was one diesel electric demonstrator that was built using the same engines as the d.h. "Western" class), I'd be tempted to conclude that perhaps the d.h. type tends to be a bit lighter, but not by much. Still, the American examples seem to suggest the weight saving might have been significant. The DH-643 is said (by Wikipedia) to have weighed 478,000 pounds. The most nearly comparable diesel electric model, also with two 12-cylinder engines, was GE's U50C, which weighed in at something like 514,000 pounds, despite (disastrous: aluminum wiring) attempts to keep the weight down. It would be interesting to know how much of the weight difference was actually due to the choice of transmission.

On the comparative weight issue, I suspect that the difference was exaggerated back in the 1950s. Back then, most diesel-hydraulic locomotives used high-speed engines, which were much lighter than their similar output medium-speed units, and unusual forms of construction, such as stressed skin. Thus it was difficult to separate out the weight decrement solely attributable to the use of hydraulic as opposed to electric transmission. The combined weight decrement from all of these design elements (just) allowed the building of approximately 2000 hp 4-axle diesel-hydraulic locomotives with a total weight of around 80 tonnes (roundly 180 000 lb), thus suitable for European main line services, but that was not a target that diesel-electric locomotives easily met, even when high-speed engines were used. Thus diesel-electric locomotives of that power level that were suitable for European conditions almost inevitably had six (or more) axles, and so a lot of extra weight; as a result they were compared unfavourably with diesel-hydraulics.

E.S. Cox (1) of British Rail (BR) is on record as saying that the diesel-hydraulic weight advantage disappeared once six axles were involved, and the available evidence from the beginning of the 1960s does tend to support that. For example, the BR 52 class “Western” C-C diesel-hydraulic of 2700 hp, with two high-speed engines, weighed 108 long tons. The DP2 prototype C-C diesel-electric, also of 2700 hp, but with a single medium-speed engine, weighed in at 105 long tons. The 52 class used the Krauss-Maffei form of monococque construction first seen with the DB V200. At the time DB had abandoned it for new construction, instead preferring to develop the single-engined, conventional construction V160 series as being a more economical approach. BR seemed to have missed that cue. DP2 used carefully executed conventional cab unit construction, derived from that used for the class 55 “Deltic” diesel-electric, which had 3300 hp from a pair of high-speed engines and weighed just 99 long tons. Of course at that time, at the 2700 hp level, the diesel-hydraulics required two engine-transmission units whereas diesel-electric could be built with one.

That said, for a given power capacity, I’d expect that a complete hydraulic transmission assembly would be lighter than a complete electric transmission assembly, but right now I cannot put my hands on numbers to support this. Furthermore, the extra weight required for hydrodynamic braking would be less than that required for an electric dynamic brake assembly. In the hydraulic case, the braking unit itself was typically a double fluid coupling mounted on the transmission block, and the existing cooling group was used to provide heat dissipation.

Re the Alco-SP case, Strapac (2) notes that Alco’s original proposal quoted a weight of 373 000 lb if two Alco 12-251 engines were used, but 336 000 lb if two Paxman high-speed engines were used (presumably the Ventura model, but that is a guess on my part). Alco and Paxman had a cross-licensing agreement that I don’t think was ever used in a material way. So that gives some measure of the high-speed vs. medium-speed engine differences.

Unfortunately, Strapac does not really give much insight into whatever ailments afflicted the SP diesel-hydraulic fleet, other than that the K-M engines and cooling groups were apparently troublesome. I suspect though two significant items were that they were very different, both Alco and K-M, and even more very different in the case of the K-Ms. I haven’t found much else, either, although this excerpt from Keller (3) (of Voith) provides some information, and suggests that the Alco engines were mostly trouble-free, as were the Voith transmission blocks.
Another excerpt from Keller though suggests that the DH643 was not completely free of secondary vibration problems.
I have also looked at the Alco DH643 vs. GE U50C weight comparison, but I am reluctant to infer too much from it. They may well have been designed with different weight targets in mind. It is conceivable that with the U50C, the UP was aiming for about the same (relatively high) per-truck weight as on its GTEL8500s. Perhaps with an apparently generous weight limit, GE thought it had plenty of room to manoeuvre, and was well into its design before it saw any problems that then required whatever lightening measures could be applied. Whether Alco’s fishbelly mainframe conferred any advantage unknown, but the main longitudinals look to be shallower than those of the GE U50C over the trucks.

Cheers,

I doubt if the lighter weight really made much difference to US railroads, although I assume it did allow the twin-engine C-C approach to be feasible for Alco and stay within SP's weight limits.

The advantages I have read about:
1. Since all axles in a truck are geared together, they achieved higher adhesion (33% vs 25%).
2. They worked as well as a diesel-electric locomotive by the same builder.

The problems I have read about:
1. The hydraulic transmissions cost about the same to maintain as an electric transmission, so no cost savings.
2. The heavy transmissions geared to the axles were unsprung weight, and thus were harder on the track.
3. The traction motor rapidly became more capable and thus one of the reasons for using hydraulics was lost.

Number 3 has other examples - EMD electric transmissions were somewhat unreliable in those days:
1. GP35/GP-40: WP bought GEs partly because of these problems.
2. SD-40: CP bought MLWs after being unhappy with their first SD-40s.

Hope this helps,

Thanks, Tgibson and Pneudyne!
Pneudyne said:
"I have also looked at the Alco DH643 vs. GE U50C weight comparison, but I am reluctant to infer too much from it. They may well have been designed with different weight targets in mind."
And, in the absence of further information, I suspect reluctance to infer too much is the wisest policy…

Further remarks.
Alco DH 643 supposedly 378,000 pounds. GE U50C perhaps 417,000: a difference of about 10%. Not really all that much more than the difference between the lightest and heaviest examples of a single locomotive model: customer options make it very hard to infer anything about the "intrinsic" weight of the design! (Note that about 5,000 of the 39,000 pound difference is due to the engines themselves: despite the similar specifications of the engines, the GE 7FDL-12 was significantly heavier than the Alco 12-251.)

I mentioned British examples as evidence relevant to claims about the comparative weights of DH and DE locomotives, and Pneudyne mentioned what seems to be a counterexample: The (English Electric prototype) DP2 (diesel electric) and the "Western" Class 52 (diesel hydraulic) were both 2,700 h.p. (nominal) locomotives, but DP2 weighed only 105 (long) tons to the Class 52's 108… despite using a very heavy diesel engine. (The English Electric engine had 10"x12" cylinders, but usually had slightly lower per-cylinder output than contemporary Alco or C-B/GE engines with 9"x10.5" cylinders…) But there are so many differences between the designs that it is hard to tell what the comparison amounts to. (Aside #1: English Electric's designers seem to have been extraordinary in their ability to come up with very powerful but light-weight locomotive designs! Aside #2: I'd love to witness the fireworks at a debate between historians of British railways on the topic "Resolved: British Rail would have done better by ordering duplicates of DP2 from E.E. rather than asking for the design modifications that led to the Class 50." (Grin!))

But it seems to me that British Rail ALSO performed something closer to "controlled experiments" to compare DE and DH designs: they acquired at least two pairs of DH/DE designs which, in other respects, were very, very similar.
(1) The Class 52 Diesel Hydraulic used the very same engines as the Class 53 (one unit built: the Brush prototype Falcon) Diesel Electric. Class 53 weighed 115 tons: between 6% and 7% heavier than the DH.
(2) Class 22 Diesel Hydraulic (68 tons) and Class 21 Diesel Electric (72.42 tons) also used the same engine (well, slight variants of the same engine design)… and were built in equal numbers (58 units each), suggesting that they may have been chosen deliberately to allow a comparison of the merits of the two designs. Weight difference here is just 6.5%.

So. I am reluctant to infer too much from all this. (But the weight differential between the DH 643 and the U50C seems larger than for other DH/DE pairs.)

(Information on BR locomotives from B.K. Cooper, "BR Motive Power Since 1948" (London: Ian Allen, 1985).)

Allen Hazen wrote:
Further remarks.
Alco DH 643 supposedly 378,000 pounds. GE U50C perhaps 417,000: a difference of about 10%. Not really all that much more than the difference between the lightest and heaviest examples of a single locomotive model: customer options make it very hard to infer anything about the "intrinsic" weight of the design! (Note that about 5,000 of the 39,000 pound difference is due to the engines themselves: despite the similar specifications of the engines, the GE 7FDL-12 was significantly heavier than the Alco 12-251.)

There’s the rub – just what was the actual weight of these two locomotive types?

For the DH643, Strapac quoted an “invoice weight” of 400 900 lb as well as Alco’s original design estimate of 373 000 lb. Railway Gazette (1964 October 16) said 174 long tons, say 390 000 lb.

What might be considered to be the “catalogue” weight of the GE U50C was indeed 417 000 lb, as reported in Railway Age for 1969 February 24. But Cockle, in “Giants of the West”, reported 442 660 lb for the UP fleet. One may wonder whether that significant increase was the joint result of UP-specified options and a major weight overshoot on GE’s part.

Allen Hazen wrote: (Aside #1: English Electric's designers seem to have been extraordinary in their ability to come up with very powerful but light-weight locomotive designs)

That was mostly close attention to detail, I think, particularly by its Vulcan Foundry subsidiary, and a capability honed by addressing the needs of the CM-gauge export market. But EE did miss the mark on the high side now and again, such as with the BR class 23.

Allen Hazen wrote: (Aside #2: I'd love to witness the fireworks at a debate between historians of British railways on the topic "Resolved: British Rail would have done better by ordering duplicates of DP2 from E.E. rather than asking for the design modifications that led to the Class 50." (Grin!))

That would take us well off topic, so I’ll stay mute – perhaps try it on one of the UK forums – as a “light blue touch paper and run for cover” exercise. (November 05 would be good timing, I think).

Allen Hazen wrote: But it seems to me that British Rail ALSO performed something closer to "controlled experiments" to compare DE and DH designs: they acquired at least two pairs of DH/DE designs which, in other respects, were very, very similar.
(1) The Class 52 Diesel Hydraulic used the very same engines as the Class 53 (one unit built: the Brush prototype Falcon) Diesel Electric. Class 53 weighed 115 tons: between 6% and 7% heavier than the DH.
(2) Class 22 Diesel Hydraulic (68 tons) and Class 21 Diesel Electric (72.42 tons) also used the same engine (well, slight variants of the same engine design)… and were built in equal numbers (58 units each), suggesting that they may have been chosen deliberately to allow a comparison of the merits of the two designs. Weight difference here is just 6.5%.)

I suspect that the first pair was a simply happenstance situation. Quite probably even when it was planning the class 53 prototype – which was not compliant with the BR specification - Brush was confident of completing the “back door” deal that bypassed the BR design competition and resulted in the class 47. So the 53 was something of a red herring. It had Maybach engines simply because Hawker Siddeley, who owned Brush, was the UK licensee.

On the other hand, the classes 21 and 22 were chosen in part to allow comparison between electric and hydraulic transmissions in otherwise similar locomotives. But they were deployed in different regions, and were probably beset by enough troubles generally that any performance differences between the transmissions were lost in the noise. They were built by NBL, who, it has been said, as a diesel locomotive builder was somewhere out on that lonely desert road between hopeless, helpless and hapless at one end, and clueless at the other. An oddity was that they had very different truck designs, cast Commonwealth inside equalized on the 21, good but probably quite heavy, and a fabricated design on the 22, which was essentially a two-axle derivative of the three-axle version originally designed by Ivatt and Fox of the LMS, also good and probably a bit lighter. So one could not say that they fully followed an objective of being the same except where their different transmissions dictated otherwise. Working them as MU pairs, one of each, might have been one way of comparing them, but then they couldn’t interwork. (BR’s laudable original objective in 1954-55 of having a common control/MU system throughout was essentially torpedoed by some members the locomotive building industry.)

Cheers,

As I suspected, Pneudyne, you know a lot more about British diesel locomotives than I do! When I looked up the weights, I also noticed that the Diesel Electric Class 21 was a few feet longer than the DH Class 22. Did the components of the diesel electric transmission (generator, control switchgear, etc) require a bigger "box" to house them than the pieces of the diesel hydraulic's machinery?

Since it was a one-off prototype (though apparently good enough that BR actually bought it and assigned it a Class number), I knew Falcon was hard to compare with anything, and it's over-all design is very different from that of the Class 52 diesel hydraulics. But I don't have many "data points" for dh/de comparison! And it was an example of what a d.e. design, for British service and with British designers of (I assume) similar feelings about weight saving, using the same engines as a fairly successful d.h. design, COULD be.

American railway enthusiasts don't get enough exposure to overseas practice. Maybe we should START a string about British diesels on the "General Discussion - Locomotives and Rolling Stock" forum of Railroad.net. (And I had a feeling that loyalties and emotions about the Class 50s etc might make for fireworks in a discussion. November 5th sounds about right!)

And of course, the actual weights of the DH 643 and the U50C (weighed under the same conditions: no fair comparing empty weight with the weight of one with a full fuel tank, for instance!) are hard to be sure of… though I didn't realize the situation was as bad as the numbers you quote suggest!

(Anyway, thank you for your comments. Informative, as always!)

Some other problems with the diesel-hydraulics were:

1. All of the wheels on a truck HAD to be kept almost exactly the same diameter. Because all three axles were mechanically connected together any variation in diameter would cause torque differences which could break the cardan shafts;

2. the KM units required greater preventative maintenance than the SP was used to doing;

3. the usual problem with minority builders, they were different. Parts were harder to come by, and in the case of the KM's, there was the problem of metric parts.

There is an excellent book about the diesel-hydraulics available, Southern Pacific and the KM Hydraulics by Robert Zenk. It covers the development of the diesel-hydraulic program with the D&RGW and SP, the original six units, the 21 "production" units, the 3 "Alco-haulics", and the one surviving unit. Also covered are the diesel-hydraulics sold to Brazil (meter gauge), and proposed versions.

https://www.amazon.com/Southern-Pacific ... 0984624791" onclick="window.open(this.href);return false;

Stuart

I knew there was one more drawback I was forgetting - equal wheel diameters. Thanks, Stuart.

I've read that the hydraulics took longer to reverse direction, than a diesel-electric. Something about the gear boxes needing to get properly meshed, before the throttle could be opened. (Perhaps not a big consideration with road units, but another small drawback to add to the pile).

JR

... And, one more that may have been a big one: I have read that d.h. transmission isn't as efficent as d.e., that a smaller proportion of engine horsepower gets translated into train-pulling horsepower. (I.e.: the fraction
horsepower-at-rail/horsepower-of-engine
is smaller for d.h. than d.e.)

Don't have much in the way of documentation, but one source I have suggests the peak effective (i.e.: at rail) horsepower of the British Rail d.h. Class 52 was unimpressive when compared to that of d.e. types of the same nominal engine horsepower. Perhaps explaining why B.R.'s later orders for big diesels to finish the elimination of steam were for Class 47 and Class 50, both d.e. types.

mp15ac wrote:There is an excellent book about the diesel-hydraulics available, Southern Pacific and the KM Hydraulics by Robert Zenk. It covers the development of the diesel-hydraulic program with the D&RGW and SP, the original six units, the 21 "production" units, the 3 "Alco-haulics", and the one surviving unit. Also covered are the diesel-hydraulics sold to Brazil (meter gauge), and proposed versions.

https://www.amazon.com/Southern-Pacific ... 0984624791" onclick="window.open(this.href);return false;

Thanks, Stuart. I wasn’t aware of that book; I have now ordered a copy.

Allen Hazen wrote: When I looked up the weights, I also noticed that the Diesel Electric Class 21 was a few feet longer than the DH Class 22. Did the components of the diesel electric transmission (generator, control switchgear, etc) require a bigger "box" to house them than the pieces of the diesel hydraulic's machinery?

Looking at the respective equipment layout diagrams, that appears to have been the case. Clearly, there is an irreducible minimum box length required, but often locomotives are longer than that for various other reasons, such as the minimum truck spacing required for bridge loading, accommodation of underslung fuel tanks, and for weight balance. The latter may be more difficult for diesel-hydraulic locomotives, particularly the single-engined type. The transmission drop-box needs to be more-or-less central, and that means that in turn that the engine must be somewhat off-centre, immediately creating a potential weight balance problem. The situation was eased by the use of lightweight high-speed engines, which is perhaps another reason why they were often preferred for diesel-hydraulic locomotives.

From the diagrams, one gets the impression that apart from the general body style and type ( both were “cowl” units rather than true cab units with truss-type body sideframes), the engine and the wheel arrangement, the 21 and 22 otherwise differed quite a bit, as dictated by their respective transmission types. Even the underframe depths followed different profiles. The cooling groups were quite different, but that I think was to be expected given that in the diesel-hydraulic case, it also served to cool the transmission fluid. The 21 was probably designed to meet an 18 long ton axle loading, and came out just a little heavy. What was the target for the 22 I don’t know, but conceivably it was 17 long tons.

I have attached the diagrams; probably off-topic but illustrative of the general theme, in that the any post-facto analysis of the SP diesel-hydraulic program in general and the Alco DH643 in particular probably needs to be done against a broader background of diesel hydraulic relative to diesel-electric traction generally.
In the Alco case, perhaps the closest comparison we can get to is between the DH643 and the C855; in an ersatz way, the latter was a three-quarters edition of the former.

Allen Hazen wrote:...I knew Falcon was hard to compare with anything, and it's over-all design is very different from that of the Class 52 diesel hydraulics. But I don't have many "data points" for dh/de comparison! And it was an example of what a d.e. design, for British service and with British designers of (I assume) similar feelings about weight saving, using the same engines as a fairly successful d.h. design, COULD be.!

That’s a good point. The 53/Falcon represented a “state-of-the-art” attempt to build a twin-engined, C-C, diesel-electric locomotive of around 2700 hp within a weight limit of 114 tons. It came out just above this target, at 114 tons. The class 47, powered by a very heavy medium-speed engine, and using a generally similar carbody, came in at about the same weight, suggesting that at this power level, there was not a material advantage to using two high-speed engines. The 52 on the other hand could be seen as the best effort, using the K-M form of advanced light-weight construction, at a broadly similar diesel-hydraulic locomotive.

jr wrote:I've read that the hydraulics took longer to reverse direction, than a diesel-electric. Something about the gear boxes needing to get properly meshed, before the throttle could be opened. (Perhaps not a big consideration with road units, but another small drawback to add to the pile).

The Voith transmission (as did the Mekydro) used air-operated dog clutches to engage the respective forward and reverse output gear trains. It was possible to have tooth abutment that prevented engagement, and some early diesel-hydraulic locomotives had operator’s desk warning lamps to show this condition, as well as an interlock that prevented converter filling until either forward or reverse was properly engaged. (I think that the usual remedy was “try again”.) In fact, early diesel-hydraulic locomotives seemed to have more warning circuits than diesel-electrics and so typically required more wires in the MU bundle. This may well have retarded the advent of mixed-MU operations, even though Baldwin had included that in its 1954 list of diesel-hydraulic locomotive desiderata.

What also may have been a complication for SP is that the Maybach engines almost certainly required a different lubricant. I very much doubt that they would have survived for very long at all on the regular US zinc-free railroad diesel engine oils of the time. And measures would have been required to ensure that the Maybach engine oil did not inadvertently find its way into any EMD diesel engine, for which zinc is quite toxic. The Voith transmissions also required their own oil type, which was also zinc-containing.

Cheers,

Pity the Class 21 and Class 22 differ inso MANY respects (though you suggest good reasons why they may have): it makes it harder to regard them as a "controlled experiment" to compare the two transmission types! (And thank you for the diagrams: the drawings in the Cooper book are exterior side views, not general arrangements diagrams!)
Cooper does give a small bit of information about the Class 22 truck ("bogie" in British English) design:
"Mechanical constructionwas on similar lines to the D600 class [= BR Class 41], the bogies being a two-axle version of the Class D600 three-axle bogie."
Given the remark on the Class 41,
"Mechanical construction of the locomotives was in the substantial style favoured by the BTC [= British Transport Commission, the gornment "parent" o the railways] and also in the tradition of a firm with a long and famous history of building steam locomotives,"
This doesn't inspire confidence!

Allen Hazen wrote:... And, one more that may have been a big one: I have read that d.h. transmission isn't as efficent as d.e., that a smaller proportion of engine horsepower gets translated into train-pulling horsepower. (I.e.: the fraction horsepower-at-rail/horsepower-of-engine is smaller for d.h. than d.e.)

Don't have much in the way of documentation, but one source I have suggests the peak effective (i.e.: at rail) horsepower of the British Rail d.h. Class 52 was unimpressive when compared to that of d.e. types of the same nominal engine horsepower. Perhaps explaining why B.R.'s later orders for big diesels to finish the elimination of steam were for Class 47 and Class 50, both d.e. types.

There is quite a bit of detail on the BR comparison in the book:

David N. Clough
“Hydraulic vs Electric: The Battle for the BR Diesel Fleet”
Ian Allan, 2011
ISBN 978 0 7110 3550 8

There is also a précis in the attached magazine article.
The apparent lower efficiency (or should that be lower apparent efficiency) did not seem to be a deal-breaker for BR.

In the case of the class 52, not-so-good transmission matching was a factor. This is never going to be perfect with torque converters, but Voith reckoned that ±2% loading variation over the speed range was achievable. But the 52 was apparently quite some way from this target. At some speeds, the 52 engines were overloaded, thus dragged down below their rated speeds, with resultant reduced power output. At other speeds, the engines were underloaded, so they ran up the governor droop curve with less than full rack, again reducing power output. The reported efficiency numbers were probably inclusive of the overloading and underloading effects. This was fair, because they represented loss of installed pulling power for whatever reason, even if some of those losses were not in the form of wastage through heat rejection.

I imagine that close attention was paid to transmission matching for the SP diesel-hydraulic fleet. Even so, the tractive effort-speed curves for the K-Ms and the Alcos have that characteristic Voith triple-converter transmission “look”, almost as if they consisted of three tangents of successively reduced slope. The Voith L830rU transmission had a two-stage first converter and a single-stage third converter. The second-converter was a bit odd; essentially a two-stage unit but lacking the second stator between the second turbine and the impellor.

Cheers,