RMEthe issue with short limited cutoff is torque falloff
Everyone agrees instantaneous tractive effort on an 85%-cutoff two-cyl engine peaks when the two crankpins are on the front quarters-- right? And that maximum TE is something like 1.4 times the minimum TE during one driver rotation-- right? It seems you think a 50% cutoff engine's maximum is more than 1.4 times its minimum, during one driver rotation. So at what driver positions do those minima-maxima occur?
Why do you keep talking about maximum torque when the issue with short limited cutoff is torque falloff causing unevenness?
The increased 'slip propensity' is not due to any enhanced high pressure, connected with rod angularity or not; it's due to needing more throttled steam to achieve the same effective average starting power (net of the expansion and condensation losses) ... so the admission peak effective pressure will be closer to boiler pressure.
Pressure rise to peak resultant (very likely throttle- and slot-port determined) will be relatively quick once the valve has opened to steam, and once the cylinders and pistons and transfer port area have warmed up and are not producing massive condensation (to be blown out the cylinder cocks) the torque rise will not be a significant issue adjustable separately through cutoff control, independent of admission timing and some relatively small unshrouding effect.
RMEearly cutoff... will produce drops in the wheelrim torque
If a 50%-cutoff engine produces X pounds of tractive effort when one piston is at midstroke, what do you figure its maximum instantaneous TE will be, during one driver rotation? (Total from the two cylinders, still assuming infinitely long mainrods.)
timzFor both engines, the minimum of the curve is when one cylinder is at one end of its stroke-- at what point in the stroke would the 50% engine have a higher maximum than the 78% engine's maximum?
That is not the only determinant of 'smoothness' - these are not internal-combustion engines. Net of the contribution of all four 'cylinder ends' to one revolution's torque smoothness, the early cutoff (particularly if there are cold environmental conditions anywhere except the air to which the steam is exhausted) will produce drops in the wheelrim torque and it is these that cause the 'less even' torque product. This can be exacerbated if the engine is starting or being worked at slow speed, where the early cutoff leaves a cooling shot of steam in the cylinder losing effective pressure by the second (pressure accelerating downward due to nucleate-condensation phase change).
The slot starting ports go a reasonable way toward making this less dramatic, but in practice there can be relatively little actual mass flow in them, lest compression control and some other things can become amusing. As noted, there is an enhanced effect on nominal starting TE from the longer cutoff, but it's only on the order of 6000lb or so. This might be the same sort of 'heating artifact' that was used to explain the early N&W 'booster valve' for LP on Y-class compounds.
RMEthe I1sa (modified for 78% cutoff) - the difference in starting TE for a nearly 30% lengthening of cutoff was a nominal measly 6000lb, but I suspect the real value was more uniform/less peaky torque
Why would the 78%-engine's curve be smoother than the 50%-engine's?
To keep things simple, assume the main rods are infinitely long. For both engines, the minimum of the curve is when one cylinder is at one end of its stroke-- at what point in the stroke would the 50% engine have a higher maximum than the 78% engine's maximum?
E7s in 1945 were NOT particularly reliable and clean-running, - RME
Well not in the minds of marketing men, rail passengers, the general public and the roundhouse foreman! BS and perceptions go a long way in what's what in the world and the Diesels were going to be winners, come hell or high water.
I wonder what Paul Keifer really thought, as in deep down, memoir type thinking. He watched all his Hudson's go to scrap, the magnificiently designed Niagara junked, the Mohawks all gone. He was a brilliant and dedicated NYC Motive Power guy but I suspect he had some emotions about all of this. I even wonder if he could see the demise of the NYC overall.
Even as a kid I knew something was wrong, something really stinks about all this, just did not make sense, with the scrapping of big new steam, that a mere year ago was the future. How long did the shine on the T1's star last? 3 months? 6 months? ...or more like "oh my God there are still 25 to come yet".
MiningmanIt is strange about the 5500. Quite shocking really but suspect that the thinking had radically changed. Scrapped in '51 I believe, probably just not going to do any more modifications or monies spent on locomotives they knew were doomed.
Aside from the issue of using type A gear, there was a circumstantial problem, I think related to assumptions in the original Niagara design, that added to the problem. I do not think le Massena quite worked through this when discussing the Niagara in the '80s.
Remember that at the close of the War the C1a Duplex was to be the high-speed passenger engine, and the 4-8-4 was a 75"-drivered design reasonably comparable to Rock Island and D&H designs: good, but not particularly high-powered or notable. Turning that into a 79"-drivered 6000hp locomotive raised the locomotive into a different category, which made very good sense ... as long as there were passenger trains in number and profitability to keep them largely busy. 5500 was intentionally 'derated' to match the output of a conventional Niagara on less coal and steam, and I think the 'where's my big savings' attitude after 1947 or so, associated with the boiler changeout expenses, made the compromise design unattractive after NYC engaged in its Dieseliner marketing push.
Note that PRR could have assisted in converting the S2a "back to piston valves" - they had gone so far as to patent the conversion system that produced the T1a - but apparently were not asked.
Arnold Haas had a story about crews using abusive operating techniques to wear out or break Niagaras in the mid-Fifties; I won't repeat it here as some will still be as sensitive as I am about that.
The Niagaras were magnificent but stillborn, obsolete the minute they arrived as were the T1's.
They did not become obsolete, far less stillborn, at the time of delivery, in fact all the way up to the point Kiefer's survey of motive power went to press. A combination of things doomed them, not least the progressive failure of trains 'worthy of their steel'. E7s in 1945 were NOT particularly reliable and clean-running, cost more than three times as much per hp, and were fragile in a number of ways (be sure to shut off going over grade crossings, and watch for their back transition); it was not a sure thing then that NYC would dieselize quickly, as it was for a road like ATSF with much longer distances and far less easy watering.
More to the point, perhaps, are the changes between 1947 and 1949, leading to the surprisingly abrupt termination of all the happy big-steam proposals happening right up to then.
It is strange about the 5500. Quite shocking really but suspect that the thinking had radically changed. Scrapped in '51 I believe, probably just not going to do any more modifications or monies spent on locomotives they knew were doomed. The Niagara's were magnificient but stillborn, absolete the minute they arrived as were the T!'s.
As they "allowed" the firebox wrappers to fail one by one they were quite deliberately sent to scrap, as if they could not wait to be done with them.
RME has alluded to corporate "sabotage" with the T1's as well. Poor coal being used and internal memo's "suggesting" leading to an early demise as examples. They went downhill rapidly, at least in cosmetic appearance. 52 of them sitting on a property with everymore E7"s getting the juicy assignments. Bring me those reliable, modern and clean running E7's!
From the decision date to build the Niagara's and T1's"s to delivery on the properties the world of locomotive thinking was firmly and forever set on Diesels.
They just "used 'em up" to get something out of them while they were still new but their star status never really happened.
timzThe 7% efficiency for the PRR I1s was on the test plant, and maybe not at full cutoff. (Or maybe it was 50% cutoff-- but at 14.24 mph, not dead drag speed.)
Two runs are mentioned at 7.1% thermal efficiency: the one at 14.24mph (and 50% limited cutoff) and another at very slightly over 22mph at 40% cutoff. It might be interesting to see efficiency of the I1sa (modified for 78% cutoff) - the difference in starting TE for a nearly 30% lengthening of cutoff was a nominal measly 6000lb, but I suspect the real value was more uniform/less peaky torque at low speeds.
Note the fairly extensive use of starting ports (Porta called them 'Weiss ports' in the ACE3000 patent discussion) in improving slow-speed TE on limited-cutoff engines. At higher cyclic the additional mass flow passed by the slots is minimal, albeit the steam quality is compromised somewhat via 'wire-drawing' effects.
The patent discussions of Franklin "type C" (the variable-cam setup described in the '47 Cyc) are one of the places I find a discussion of what Lima called "Long Compression" (see Townsend's ad for the stillborn 4-8-6 for a reference). This concerned modulating the exhaust differently from the inlet poppets - something difficult to do with type A oscillating, but comparatively easy with shifting continuous-contour cams with spherical roller followers.
I am still a bit surprised, given the history with the C1a, that Kiefer did not revise Niagara 5500 with Type B-2 (rotary cam on the type A valve alignments), which would have corrected the major problems that sent that locomotive with eerie rapidity to an early grave.
Paul MilenkovicOne problem with demanding high tractive effort at limited cutoff is that the torque becomes less even.
The 7% efficiency for the PRR I1s was on the test plant, and maybe not at full cutoff. (Or maybe it was 50% cutoff-- but at 14.24 mph, not dead drag speed.) But it seems it was the average for an hour. (Hopefully the Wikipedia article still has the test results https://en.wikipedia.org/wiki/PRR_I1s )
Paul MilenkovicWhat I am speaking to is both multi-cylinder and compounding making separate contributions to lower torque ripple, of benefit to drag service at the adhesion limit.
The majority of the torque-ripple reduction comes from three-cylinder DA: if you plot the resultant for Swiss drive (which a three-cylinder is) every 15 degrees, and then accou nt for angularity, double-acting cylinders, compression etc. (as David Wardale has for 2-cyiinder DA engines) you will see comparatively little variance. If the compounding is not fully IP-"equalized" you will have drops in the curve due to falloff in LP working efficiency and perhaps receiver backpressure, and perhaps variability inthe effective backpressure on the HP exhaust at unequal intervals, too. I'd expect this might be slipperier than a three-cylinder simple developing equivalent DBHP even though some of the antislip-blocking features of Mallet compounding might carry over with fixed phasing.
Yes, the situation is likely better than a two-cylinder engine, but that's easily understood. A somewhat better improvement (and recognized as such on 160 A1) is to keep the cylinders hot so that full expansion from relatively short/early cutoff occurs with limited condensation; this also helps (in theory) a 2-cylinder simple exert more torque per revolution, but the benefits of a short cutoff are pretty much 'all had' somewhere between 25 and 15 percent, with anything more precise 'lost' before the relatively small admitted charge has finished expanding and nucleate condensation has sufficiently set in.
The 'right' answer to slipping was, is, and remains some form of fast-acting driver brake that allows the locomotive to be knowingly controlled just shy of its effective adhesion limit. Anything 'cobbled in' to achieve that goal with steam-circuit characteristics either leaves horsepower money on the table much of the time, or gives T1-style low- or high-speed slipping under possibly unpredictable or variable poor conditions.
What I am speaking to is both multi-cylinder and compounding making separate contributions to lower torque ripple, of benefit to drag service at the adhesion limit.
Multiple cylinders per engine set contributes phasing. Compounding (with a good IP live-steam injection scheme) contributes efficiency achieved at longer per-stage cutoff.
For example, a two-cylinder "balanced compound" on a rigid frame, or the N&W Y-class Mallets for that matters, achieve efficiency at higher cutoff, but they don't have the advantage of multi-cylinder phasing. Simple expansion 3-cylinders such as the UP 9000 class achieve phasing to reduce torque ripple but suffer poor efficiency at long cutoff.
That the D&H weirdness didn't work doesn't disprove my point. I am suggesting that compounding is beneficial at conventional boiler pressures when the service is lugging near the limit of adhesion because it allows longer cutoff with lower torque ripple to be used. I am thinking that is what is missing from the claim, "A simple can achieve similar economy -- just shorten the cutoff." Developing high power at short cutoff will slip the wheels -- think T1.
If GM "killed the electric car", what am I doing standing next to an EV-1, a half a block from the WSOR tracks?
Paul Milenkovic I am beginning to think this is missing a point, the point being that a three cylinder compound could operate efficiently when worked close to the adhesion limit in a way that a two-cylinder simple cannot?
This is really the point that you might want to concentrate on, considering the principal thing that doomed the D&H 1400s wasn't so much their complexity as their relative inability to generate ton-miles on a fluidly operating railroad.
Look first at the features of Chapelon's 160 A1 that would translate to a good American-style compound - you might not need six cylinders doing the thermodynamic work of five, with proportional multiple expansion, but the HP steam circuit and the method of implementing LP reheat stand as two relatively good ways of improving things with comparatively low maintenance implication.
Meanwhile, B&O and Baldwin also recognized by the early '30s that pressure above roughly 300psi required compounding for best utilization. With the advent of practical welded boiler shells and as-fabricated stress relieving, I think the promise of cost-effective higher pressures was very clear. But even the most sophisticated mechanical poppet-valve systems can't provide "improved power" in a quartered two-cylinder reciprocating engine (for reasons well described on the Duke of Gloucester web site in connection with British Caprotti's nominal cutoff precision) so, even absent the better torque-ripple profile of a multiple-cylinder engine with appropriate phasing, there starts to be an advantage in having more than two cylinders ... an advantage leveraged by the increased difficulties with reducing reciprocating mass with lightweight thin-section rods and getting mainpins to live longer, etc.
Note that the 'first best use' of IP injection (also a la Chapelon and Porta) on something like a N&W Y class is not "reheating" the LP steam for lower condensation -- although that is valuable -- but in equalizing the power contribution and balance in the LP engine to permit it to make an equal contribution while running freely, and thereby allowing the locomotive (with a very high proportion of its weight on drivers) to actually run at road speed with performance on a decent part of the HP curve -- clearly no longer making less cylinder HP than a NYC Hudson above about 35mph.
I have argued repeatedly that a conjugated duplex with appropriate phasing (easier to do with Deem's or my version of conjugation than Withuhn's if high-speed balance is important) is the only multiple-cylinder reciprocating configuration that maintains the 'advantages' of two-cylinder American construction, but there is no great reason why modern materials and methods would rule out a cranked axle for three-cylinder service, or a piston-valve drive to a compound middle cylinder that would not 'stretch events' for it, or act mechanically opposite to it so inertial effects progressively screw up its performance.
Something you didn't mention, but I think bears consideration, is the use of proportional "IP" injection on a two-cylinder compound locomotive, like a glorified von Borries (or early 1400!) so that it is free-running and effectively balanced at all speeds. Here, a bit conversely, very low speeds are a bit of a problem, but only because of the number of pulses per revolution; you'd need to do an analysis to see if the savings from compounding outweigh the disadvantages ... relatively lower as they might be with modern construction ... of compensating lower drivers.
The big advantage for modern steam in fleet use isn't so much fuel saving as it is water-rate consumption when treated water is in use ... as it would have to be for high boiler pressures. The predominant thing that killed the V1 turbine, even at 8000hp, was its water rate, which even at the higher efficiency of multistage compounding emptied a coast-to-coast-size tender in under 150 road miles. (This with a modified Q2 boiler that represented reasonably good practice for 300psi and no real problem that welded shell construction wouldn't have addressed.)
It's possible that a good compound design with good in-cab instrumentation would reduce the water rate from a reciprocating design reasonably well, and permit less makeup-water use (and hence treatment cost). Other means of using the lower-pressure exhaust remain as they were toward the end of steam (including the possibility of exhaust-steam injector systems and the use of "spent steam" in Snyder preheaters) and have their own effect on lower water rate. But I think it matters less that a locomotive 'goes further' alone, as it does that the locomotive can do so while keeping other operations on the railroad fluid. And remember that an idling steam locomotive consumes far more fuel than the amount technically needed to keep it pressurized ... and far, far more than the equivalent diesel hp would be requiring.
I believe the only lococmotive ever built in this country that was three cylinder and compound was Baldwin 60,000. I believe there are extensive reports on its perfomance. This engine used an experimental water tube boiler as well. It never caught on even after an exhaustive cross country tour to sell the principle. However, what is probably the least used retained steam engine in the US happily still resides at the Franklin Institute in Philly. I would love to see some of the studies and why it never caught on as opposed to the conventional three cylinder designs used on numerous engines in this country.
I will not dispute the theoretical efficiency of a compound rigid-frame locomotive at drag freight speeds. That being said, a realistic cost-benefit analysis would need to be made to see if such a locomotive would be worth the extra operating/maintenance expense. The experience of the D&H 1400's suggests that it wouldn't.
In discussions of whether a particular steam locomotive, or any steam locomotive, is suited for drag service, the points brought up are whether there is enough adhesive weight compared to pilot-truck and trailing-truck axle loading, whether the locomotive has high drivers for speed or low drivers for lugging, and so on.
Could there be a fundamental flaw with simple expansion in drag service? A simple expansion engine appears suited for service where starting, you only temporarily waste steam using very long cutoffs, but in high-speed operation, you can generate a lot of horsepower with very economical short cutoffs. In drag service, you want high tractive effort at low speeds, which implies long cutoffs and high coal consumption
So why can't you build bigger cylinders (or use lower diameter drivers) and operate them at low speeds at shorter cutoffs? I read on the Web that such is what the Pennsy did with their drag-service Decapods that they gave a "limited cutoff" -- 50 percent cutoff at nominal full gear. The claim is they averaged 7 percent thermal efficiency in drag service, which is actually pretty good compared to other steam engines doing the same thing.
One problem with demanding high tractive effort at limited cutoff is that the torque becomes less even. To operate closer to the tractive-effort limit, you want closer to 100 percent cutoff instead of 25 percent, where a single expansion engine is generally most efficient, or even the 50 percent cutoff where the Pennsy drag engines were operated reasonably efficiently.
Using compound expansion at 50 percent cutoff at each expansion stage, this is equivalent to simple expansion at 25 percent cutoff? Only you are getting the smaller torque ripple of 50 percent instead of 25 percent cutoff working? And in a 3-cylinder compound where the center cylinder drives a cranked axle at 120 deg in relation to the outboard cylinders and cranks, you get even smaller torque ripple? And you can operate even closer to the adhesion limit of each driver axle?
OK, I admit all of the criticism of compounds in steam locomotive service, of the pressure losses between the two compound stages, condensation in the low-pressure stage, the Rube Goldberg starting arrangements (L.D. Porta argued that you should not switch to simple operation on all cylinders for starting, rather, you should admit boiler steam, after pressure reduction, to the low-pressure receiver to balance the high and low pressure stages at starting).
But as to dismissing compound steam locomotives as not-worth-the-trouble when subject to non-condensing cycles, the steam-pressure limitations of practical locomotive boilers, and loading-gauge limits on the size of the low-pressure cylinder or cylinders, I am beginning to think this is missing a point, the point being that a three cylinder compound could operate efficiently when worked close to the adhesion limit in a way that a two-cylinder simple cannot?
Our community is FREE to join. To participate you must either login or register for an account.