"Best" 4-8-4

The title is, of course, meaningless: "best" can mean too many different things. Try to specify a bit more definitely by saying "best at meeting the owning railroad's needs" and you still probably leave open too many factors... in addition to complicating the calculation into infeasibility, since the -- 40+? -- different railroads around the world that had 4-8-4 locomotives all would have had different requirements.

So let's go with "best from the perspective of a railway enthusiast long after the end of steam". Still indefinite. But let's rule out the obviously subjective aspects like aesthetics. Power is going to be a key point here.

So. The best U.S. Norther was, of course, the New York Central Niagara. (Of course! People who want to argue for, say, the Norfolk & Wester J can write their own posts.). Horsepower... well, we'll get to that. Maybe around 6,000?

In the other corner, the non-U.S. contender: French National Railways (SNCF) 242-A-1, widely thought of as one of the most refined steam locomotive designs of all time. (This despite a handicap: Chapelon, the designer, had to create it by rebuilding an existing, not too great, 4-8-2, and would have done things a bit differently had he been able to build an all-new locomotive to his own design.) How powerful was it? In a couple of sources I've seen 4,000 and 4,200 horsepower quoted. This is probably metric horsepower (= .986 hp), so let's say somewhere in the 3940 to 4140 range.

But what do these measure? One of the sources helpfully adds that this is power "at the tender coupler," so this is going to be closer to what Americans call "drawbar" horsepower rather than "indicated" horsepower. (I guess they'll differ by the amount of power used to pull the tender itself: I'll ignore this complication because I don't know how to estimate it.) So the comparable figure for the Niagara is (from memory) 5100 or maybe 5300. The S2 (Niagara with poppet valves) prototype was never tested for maximum power (tests were run with valve cutoffs calculated for maximum efficiency), but the view has been expressed that it probably ***could*** have managed a dbhp equal to its road number: 5500.

But we may still be comparing apples to oranges. I think the quoted power for the 242-A-1 is a power it could deliver at any of a moderately broad range of speeds, not what its absolutely peak output at its best speed (which may have been higher than what it would be allowed in regular service), and for this I have seen suggestions well over 5000. There are figures for what this would be for a Niagara (spoiler alert: the 5300 is a peak output figure...), but going into this would get complicated... So let's just say 4000 vs 5500.

BUT. The Niagara is a much bigger locomotive. The S1a prototype supposedly weight 471,000 pounds, the S2 485,000 (did poppet valves and their gear really add 14,000 pounds to the locomotive weight? I suspect something else is going on here.). French track wouldn't have supported a locomotive of that weight: Chapelon was working within a 19 metric tonne per axle weight limit. The 242-A-1 weighed 328,000 pounds, or about 10% less than a New York Central Hudson! So, ***if*** we scale up power in proportion to weight (yes I know this is overly simplistic), the 4000 hp French Northern would be equivalent to a 5600 hp locomotive at the New York Central's axle loadings!

I'm sure there's lots more to be said. I hope that comments will be, well at lest minimal civil (though they can be as provocative as you like!).

Just to start further comparisons... The Niagara had 100 square feet of grate area, the 242-A-1 only abut 50. My guess is that a French fireman would be disciplined for wastefulness if he used as much coal per square foot of grate as was considered normal in main-line service in the U.S. So, if we want to go beyond power in our considerations, my ***guess*** is that the 242-A-1 was MUCH more fuel efficient than the Niagara.

Fuel efficiency did seem to be a particularly important parameter in France. Apparently coal was never very abundant there and relatively expensive. In part at least, that is why compounding remained mainstream, although largely abandoned elsewhere. In part it was probably also a self-sustaining development vector. Mechanical complexity, higher maintenance costs and reduced availability were all accepted as the trades-off for lower fuel consumption.

A perspective on this situation, as it obtained in the post-WWII era, may be found in the writings of E.S. Cox (LMS and BR locomotive design engineer), attached.

I think that this case well illustrates the difficulty that arises in detaching locomotives from their operating circumstances and environment for the purposes of comparative evaluation. One might wonder at the outcome if say the SNCF 242A1 propulsion technology were transferred, suitably scaled up, to the NYC Niagara, setting aside its likely mechanical unacceptability in the American operating environment. It might well have resulted in quite amazing peak power numbers, without much change in the curve at the lower speed end. This might have had marginal utility. At higher speeds, the Niagara had more power than its mission required. The main rationale for its existence – other than allowing NYC and Kiefer to join the elite “4-8-4 club” – seems to have been a desire for an 8-coupled locomotive without booster rather than a 6-coupled with booster as being a better way to reliably and regularly start and quickly accelerate heavy passenger trains. This I think may have developed from NYC’s experience with its L3 and L4 class 4-8-2s in passenger service, the next step being an 8-coupled locomotive that could also run as fast and as freely as its 4-6-4s. The natural consequence was a surplus of high-speed power.


Cheers,

Revisiting after 3 years...
One of the things that frustrates armchair rail historians is that the New York Central apparently never tried to see what the maximum horsepower output of the S2 (the Poppet-valved Niagara) was: the recorded tests used cut-off/speed combinations that were calculated to give high efficiency. (So, we don't know what an S2 could -- in principle -- do in a speed test: what we know is that it could do the same train-puling as an an S1 with significantly less coal and water.)
Pneudyne's comments on what the motivation for the Niagara was suggest an explanation: as far as high-speed power was concerned, any Niagara was more than the New York Central needed, so (assuming that there was still some question about whether to develop steam further instead of dieseling as fast as possible!) the relative efficiency of the two designs at as close to the same speeds and accelerations as possible was of much greater practical interest.

Brian Reed did provide summary comparative numbers in Loco Profile #20, transcribed as follows.

S-1b #6023, 75 inch drivers, 275 lbf/in² boiler pressure, dynamometer car:

Cylinder power maximum 6600 hp at 77 mile/h
Drawbar power maximum 5100 hp at 58 mile/h
Starting tractive effort 65 800 lbf


S-1b #6023, 79 inch drivers, 275 lbf/in² boiler pressure, dynamometer car

Cylinder power maximum 6610 hp at 85 mile/h
Drawbar power maximum 5050 hp at 62 mile/h
Starting tractive effort 62 500 lbf


S-2a #5500, 79 inch drivers, 275 lbf/in² boiler pressure, dynamometer car

Cylinder power maximum 6625 hp at 90 mile/h
Drawbar power maximum 5000 hp at 65 mile/h
Starting tractive effort 62 500 lbf


S-1a #6000, 79 inch drivers, 290 lbf/in² boiler pressure, Selkirk test plant

Cylinder power maximum 6800 hp at 85 mile/h
Drawbar power maximum 5290 hp at 62 mile/h


It was also recorded that the S-1b had noticeably better acceleration performance than the S-2b. That was not insignificant, as one of the diesel advantages was fast acceleration away from stops and speed checks. On the other hand, the S-2a was said to be 18% better in terms of coal and steam consumption and thermal efficiency. That aspect was not further quantified, but in round terms it would mean a shift from say 6% to 7% thermal efficiency, which is anyway within the scatter range for thermal efficiency measurements on late steam locomotives. Some steam engineers/writers, such as E.S. Cox, suggest that such differences tend to get lost in the vagaries of everyday service.

In hindsight it is all somewhat academic. There was a paradigm shift in process. The incoming technology (diesel) was probably not fully debugged, but reasonably could be expected to improve in leaps and bounds. The outgoing technology (steam) was mature. Some improvements were possible without getting into the weird and wonderful, but any advantages over the incoming technology so obtained would be but temporary as the inevitable early-life step changes were made to the latter.


Cheers,