A simple-expansion locomotive has to operate at short cutoffs in order to get a high degree of expansion of steam in the cylinder and in turn high efficiency. Long-travel valves and valve gear is meant to reduce the "wire drawing" restriction on steam flow at short cutoffs that effectively throttles the steam flow and works against getting that efficiency. To that end, the valves are given more "lap" -- a deadzone in the middle of valve travel -- along with a valve gear with a greater stroke to counteract that dead zone to avoid this efficiency robbing restriction.
Some locomotives were also built with a short maximum cutoff -- Woodard's pioneering Super Power Berkshire was limited to a maximum of 60% whereas the Pennsylvania Railroad built a 2-10-0 with a 50% limited cutoff. Wardale also went to a limited cutoff with the Red Devil. All of these locomotives 2-cylinder simple-expansion locomotives can have problems with starting. The Lima A-1 Berkshire along with the Pennsy 2-10-0 were said to have "starting ports", where they "leaked" a small amount of steam into the cylinders to surmount getting stuck in a dead spot whereas the Red Devil had Herdner valves to control this supplemental steam. Starting ports were said to have problems with getting clogged whereas Wardale's Herdner valves never worked quite right.
In rereading The Red Devil, Wardale suggests that the wire-drawing reduction of long valve lap could still allow for a higher starting cutoff less prone to stalling, but that would have required reworking the whole valve gear for longer travel, which he didn't have time nor budget to do.
So is there a purpose to a limited maximum cutoff for higher expansion and hence efficiency in a 2-cylinder simple-expansion steam engine, or is the proper solution a long-travel valve gear that doesn't require limited cutoff and its attendent headaches?
If GM "killed the electric car", what am I doing standing next to an EV-1, a half a block from the WSOR tracks?
Keep in mind that long-lap/long-travel valves are meant to improve the transfer of mass flow at high speed (and relatively high effective cutoff associated with making the necessary power to accelerate to and then sustain high speed with a meaningful train). There's an increased parasitic loss to drive them, although I don't have numbers to quote; an interesting example is the gear fitted to the late French de Caso 4-6-4s which has tremendous nominal travel and corresponding peak machinery speed for the valve tribology.
In my opinion, fixed limited cutoff was an excuse for lack of engineer discipline in starting; what it made for was extremely peaky torque which also almost demanded very quick full throttle opening to start a 'rated' consist; the result being pronounced slipping under much less than perfect conditions, where a more sensible approach would be to keep long cutoff and modulate the throttle so that any slipping would either self-correct or be easy to accommodate. Of course as soon as the reverse was hooked up past 50% you had the same effect of limited cutoff on running economy.
To my knowledge, Herdner valves were never a good solution if your intent in providing limited cutoff was to decrease water rate. Slots/Weiss ports in any decent size piston valve had to be made so thin that coking effects, especially from 'too much superheat', would rapidly occlude them, and of course cleaning them would be an exercise in irritation for any shop crew. We might do well to remember the earlier era of the riding cutoff (e.g. 'Cuyahoga cutoff') which could accomplish much of the goal of restricted cutoff even close to starting without requiring compromise of the physical dimensions of the valve gear proper.
The 'correct' answer for limited cutoff was the use of some appropriate form of poppet valve drive -- the problem then being that many of the 'advantages' of being able to have very short valve openings did not actually work the way they were touted, particularly in certain oscillating-cam schemes, What was "needed" was some analogue to Corliss-like snap action on the poppet valves, that would unshroud them quickly (the spring return onto the seats being more direct and dependent only on the falling profile of the cam(s) involved -- if the gear were set up for 'limited cutoff' this could translate directly into inadequate physical opening of the valves relative to their seats as the cam profile for lift rotated through less of a physical arc and perhaps needed a more aggressive profile to lift the valve out of shrouding in the more limited time available.
It should also be mentioned that a good poppet valve drive will have independent modulation of the exhaust valves from the intake (see the Lima "long compression" advertising in the late '40s) which gives some of the theoretical benefit of long-lap/long-travel piston valves that are independent (e.g. two spools, one for admission and one for exhaust, in two parallel valve cylinders communicating with appropriate transfer porting).
In my opinion the 'correct' answer does not lie in limited valve-gear cutoff, but in more responsive throttling close to the cylinders -- Porta's Wagner throttles as postulated for the ACE 3000. These are fast acting enough that they can be modulated almost like riding cutoff, independently of the SHM from a properly robust conventional valve gear.
Paul MilenkovicTo that end, the valves are given more "lap" -- a deadzone in the middle of valve travel -- along with a valve gear with a greater stroke to counteract that dead zone to avoid this efficiency robbing restriction.
I am looking to see if I can find Churchward's explanation of why long lap and long travel go together.
Remember that 'conventional wisdom' in early slide-valve design was to minimize the physical travel of the valve body, as far as possible given desirable precision of steam-edge control, with comparatively long ports to the cylinder ends. Where additional mass flow was desired the solution was multiple, or Trick, porting; where a more precise control over cutoff was desired, one of the riding-cutoff arrangements acting separately over the slide valve would be employed. (Perhaps a better approach would be to duplicate the valve at each end of the cylinder, much the same way a practical locomotive piston valve is fabricated, with the Trick porting working more or less directly over the transfer ports at each end of the cylinder and the cutoff arrangements duplicated over the valves; periodically one designer or another would start looking at better materials and tribology and try dusting off slide valves of one kind or another as a 'better' alternative to the limitations of conventional piston valves ... historically, to little practical effect.
Neither of these arrangements works in the transition to piston valves, in which the port area at the steam edge (net of bridges) can be enormously multiplied, but the physical area of the bearing surface and rings must be relatively large even with tailrods, and so multiple Trick-like porting can involve a substantial increase in the overall length of the valve and complexity of the transfer port (see for example some kinds of Willoteaux valve), and riding cutoff would involve either sleeves surrounding the valve piston, working a bit like the sleeve valves applied to Bulleid's Leader, or some other external arrangement either internally or externally concentric with the means providing the physical steam-edge precision and timing. You could also use an external poppet-valve arrangement analogous to that used for additional exhaust relief on Uniflow engines, and in fact this remains a potentially attractive option for extreme high speed on reciprocating steam expanders ... assuming you want to retain the physical limitations of SHM valve drive, large piston-valve inertial mass, etc. for lower-speed economy and practical service.
The great advantage of Churchward's approach in my opinion is in the effect of the long travel, not the long lap. Recall that the purpose of Trick porting is to increase the effective opening to steam, through the 'wiredrawing' unshrouding region, as quickly as possible. At high cyclic rpm the time available for this becomes very short, so the valve (body, drive and all) must be moving as fast as possible as it crosses the admission steam edge. This means in essence that it's valuable to accelerate the valve as long as possible before the admission port begins to be encountered (it should follow to you fairly quickly why Chapelon et al. liked trapezoidal ports) and it reasonably follows that closing the port to steam will also be reasonably quick, and that the same precision will follow for exhaust where it is somewhat less necessary and somewhat less wanted. (You will also recognize another reason why I think inside admission is a desirable thing for high-speed piston-valve locomotives ... but I digress.)
Reasons for the longer lap are more involved. I think of it more as a consequence of accelerating the steam-edge motion as a design principle in itself; there is value in having more 'dead time' between exhaust opening near the end of a completed stroke at one end of a DA cylinder and the start of admission at the other end, some of which is more "thermodynamic" consequence (for example, in the absence of cylinder jacketing arrangements) than physical 'waste' of expansive energy in the admitted steam.
I should note here that I don't consider the longer lap a source of 'wire-drawing reduction' in itself; only the events between admission opening and admission closing affect that as a practical thermodynamic consideration on engine performance. Once the steam mass is confined in the port and cylinder, and starts to do expansive work, we get the two advantages of superheat (in reducing wall condensation, and nucleate condensation toward the end of the stroke expansion) but that is, in my opinion at least, a very different set of things from wiredrawing, and of course more distributed in even near-exhaust conditions in the cylinder charge.
There is an additional consequence involved in this setup, which is that the exhaust events are the 'mirror' of admission and exhaust cutoff may occur cleanly before a considerable volume of exhaust steam mass has made it out of the cylinder. Up to a point this is good, because the residual compression effects will provide cushioning of inertial force, keep the cylinder end 'hotter' than it would otherwise be, and reduce the inrush 'wiredrawing' as the valve opens to steam on the following opposing stroke ... but after that point, too much compression is a very real evil, more so as the peak pressures build very rapidly and can start stretching cylinder-head studs and so forth even when present for a tiny fraction of overall stroke duration. One 'catch' in designing exhaust ports for better release is that, in most piston-valve setups, increasing the area also involves increasing the dead space in the transfer porting (see the PRR Q2 for some possibly-muddled thinking on the relative importance of this) which has effects on the cycling flow of inlet steam through the ports that will add to those of pressure-to-velocity conversion and other effects of wiredrawing at the valves. Note that often in practice a limitation of this kind will affect poppet-valve setups that nominally have separate inlet and exhaust event duration and timing - I particularly invite your speculation and opinion on effective valve provision and design to avoid problems with it.
Two reasons for limited cutoff -- more efficient at low speed (which is why the PRR 2-10-0 had it) and larger port openings at high speed (which is why the 6-4-4-6 had it). Sounds like a lovely idea, but apparently it didn't work out -- lots of RRs tried it, but few locomotives built after 1930 had it.
On the test plant, PRR's 2-10-0 averaged 7% "drawbar" efficiency for an hour, putting out 63000 lb "drawbar pull" at 14 mph. Since it's the test plant it's not truly drawbar pull, but that efficiency is still hard to beat.
The PRR 2-10-0 had a 1-1/2 inch by 1/8 inch slot near each end of the cylinder that was supposed to admit steam to the cylinder to help the engine start. Maybe it constantly got plugged with gunk? And if it was plugged, starting was too tough?
Limited-cutoff engines built after 1930: SFe 4-8-4 and 2-10-4, SP 4-8-4s had around 73%, N&W 2-6+6-4 had 75%, PRR 6-4-4-6 had 70%, and maybe the N&W 4-8-4 was a little short. Any others?
(Maybe the CNW 4-6-2s rebuilt for the 400s?)
A few quotes from The Steam Locomotive (1944 ed.) by Baldwin's Chief Engineer, Ralph P. Johnson might be useful. He wrote, "Long valve travel, which produces long port openings at short cut-offs and longer exhaust port openings, is desirable." But -- "Valve travels greater than 8 1/2 inches have generally been avoided with the Walschaerts gear as they would entail too great an angle between the center lines of the link slot and ecentric rod which would cause excessive stress in these parts, as well as excesive eccentric throws. While these same objections do not hold true for the Baker gear other bad angles occur with this gear which become quite serious as lost motion accumulates."
Johnson also wrote, "Increasing the steam lap does not necessarily increase the efficiency of the engine, but it reduces the flexibility of the valve gear with respect to range of cut-off. It therefore is usually comparatively small."
Regarding the starting port, ALCO's A. W. Bruce wrote, "The starting port is set about 1/4 in. out from the steam edge of the valve and is about 1/8 x 1 1/2 in. in size. Naturally it must be cleaned frequently to provide a free steam opening , since the slit is so small. But because it is small, it has little effect on the total engine steam consumption."
All valuable issues to consider, from someone who particularly understood the issues as presented.
Ralph Johnson, via nhrand"Valve travels greater than 8 1/2 inches have generally been avoided with the Walschaerts gear as they would entail too great an angle between the center lines of the link slot and eccentric rod which would cause excessive stress in these parts, as well as excessive eccentric throws.
Some of Wardale's thinking on efficient valve design is associated here, too, as much of the problem is technically solved with better materials and bearing technology and tribology, but the two evils (both related to inertial stresses greatly increasing with speed) of higher valve peak speed for the longer travel and larger eccentric-crank circle as driven) do remain, and this is one feature of the duplex principle as applied both to the PRR S1 and the NYC C1a.
I would direct anyone interested in this particular issue to look at how de Caso managed his very long technical valve motion, or in fact how the French combined long travel with Willoteaux multiple porting (while preserving precision at all steam edges relative to port edges and port-opening geometry). I will also mention in passing that there is no real technical discussion of the duplex principle, either theoretical or actual, in even the latest editions of Johnson's book (mine was the 1983 (!) edition) although he certainly knew and could teach many practical aspects of its superiority (see for example the Atlantic City discussion of the PRR T1 design, where he clarifies among other things why a 26" stroke and not something smaller is used on these engines). Much of the late Lima preference was likewise very ill-documented and improperly explained, leading to all kinds of railfan consternation later on.
... While these same objections do not hold true for the Baker gear other bad angles occur with this gear which become quite serious as lost motion accumulates."
If you wonder why the discussion of all-pin-jointed gear specifically involves things like Multirol bearings and not just 'rolling element' bearings -- this elimination of lost motion through bearing and tribology optimization is a large part. Some of the nominally-amusing tendency of some N&W valve gear to 'unravel' at high speed at longer cutoff (so well documented in King and Newton, among other reliable recountings) can be attributed to excessive inertial force expression as well as unexpected loading back through the gear at some parts of the stroke. That is far less to criticize N&W than to point out the truth in extending the action of Baker or other pin-jointed gear to larger valves acting through a longer travel.
Of course, poppet valves were seen as a general way to avoid most of the evils of longer travel, while preserving the higher steam flow and achievable MEP that Caprotti-style hollow streamlined valves can produce ... with prompt unseating and equally prompt debounced closure, and with careful attention to the effects of late compression and possible reversible recovery of overpressure (see Jay Carter over on the SACA side of high-speed steam development).
It's important to comprehend what this means, particularly in relation to fixed admission cutoff (which I think is normally implemented by setting the effective lap) where this whole thread jumped off.
I do wish Johnson had expressed this sentence a bit less ambiguously, as most of the discussion of longer travel has to do with more effective high-speed performance, whereas the effect of long lap is (I think) on low-speed flexibility and not anything associated with positive long travel.
Please remember the Australian experience about lower slipping and better starting achieved by keeping the throttle effectively in the 'wiredrawing' range (effectively running the engine with sliding-pressure admission) and keeping the cutoff correspondingly longer for the same net torque demand. In essence keeping 'long lap' moderated achieves the same effect both in reducing potential torque peakiness and the need to keep effective MEP high via a quick charge of relatively high-pressure and well-superheated steam.
I repeat here that Wagner throttling close to the inlet ports on each valve is a much, much better solution than limited cutoff, even in the relative absence of fast-acting proportional servo technology to modulate the Wagners. In my opinion well-insulated ducts from the Wagners around the cylinder or cylinder head to external ports also represents an easily-maintained alternative to slots necessarily involved with cylinder-oil tribology and its effects... which brings up
Note the key point here: that the slot is always open to steam, but the effective throughput is highly throttled relative to pressure differential, so anything approximating peak boiler pressure might be found there when the port is fully closed, but will rapidly fall with opening, and perhaps rise (with negative net mass flow) should there be excessive compression as the slot is opened by piston travel. Backflow will carry some cylinder oil up into the slot on each stroke, where it will likely at least start thinking about starting to cross-link, and in bad cases to coke. This is a Bad Thing in a slot geometry that is probably cast with internal asperities with some convoluted cores, and it probably doesn't 'pay' meaningfully to extrude-hone the slot ports or improve the 'steam streamlining' to them. (I would note that one intended purpose of Herdner valves is to improve the relative flow of 'starting' steam during the presumably short period that corresponds to locked-rotor torque in traction motors)
An alternative way to implement the increase in MEP during starting is to use the analogue of pilot injection, something that might be facilitated mechanically by slight additions to Trofimov- or Nicolai-style drifting arrangements where derived motion from the valve gear is used to trip small (and easily-accessed for maintenance) valves connecting to the upper cylinder head. This also represents a logical place to include 'auxiliary valves' for higher steam distribution effectiveness at very high speeds, corresponding a bit in action to a reverse 'riding cutoff' arrangement - this constituting almost the only 'moral' way to implement electronic valve control on a high-speed reciprocating engine as well as giving reasonable performance should a full-servo valve gear suffer any sort of debilitating failure on the road.
Engineer Discipline/Valve Pilot
Overmod commented about lack of engineer discipline. I imagine a lot of advanced steam locomotive technology was not effective because many engineers had some odd ideas about how to operate an engine, had their own styles at odds with the science, or often took the easy way out.
The valve pilot seems like a tool that should have helped an engineer get the most out of an engine but if installed, I wonder how often it was ignored, or disabled because it kept a tape record of how the locomotive was operated? The valve pilot could have been useful in aiding the engineer in finding an optimum cut-off but I suspect a lot of railroads didn't use it because they knew it wouldn't be used or maybe felt an engineer should be the only judge of the right cut-off (actually, the valve pilot only provided guidance, the engineer could still use judgement). I mention the valve pilot as a tool for determining cut-off because the original post was on the subject of using steam expansively and setting the cut-off appropriately is a necessary requirement for using steam expansively.
I saw Boston & Maine 4-6-2 3713 "The Constitution" in Boston a number of times and rode behind her on her last run to Portland --- the 3713 had a valve pilot. The engine is being restored at a snail's pace at Steamtown in Scranton, PA. I wonder if any restored, operating steam locomotives today are equipped with a valve pilot ?
nhrandThe Valve Pilot seems like a tool that should have helped an engineer get the most out of an engine but if installed, I wonder how often it was ignored, or disabled because it kept a tape record of how the locomotive was operated?
This would make an interesting historical paper. I have seen a number of 'anecdotes' about alleged very high speed where it is implied that 'every tape wasn't read every time', or where any technical 'overspeed' was ignored instead of reading to see only if the physical cutoff had been adjusted to 'match traces'. I have also seen a couple of home-remedy ideas about how to disable the speed recording without opening the 'sealed box' with the tape ... one as I recall involving a strategic number of wooden matches.
The Valve Pilot could have been useful in aiding the engineer in finding an optimum cut-off but I suspect a lot of railroads didn't use it because they knew it wouldn't be used or maybe felt an engineer should be the only judge of the right cut-off (actually, the Valve Pilot only provided guidance, the engineer could still use judgment).
I suspect engineers were encouraged, and perhaps cajoled, into operating with 'greatest efficiency' then as much as some of our forum members currently indicate they are by their carriers.
It is probably worthwhile to repeat what the Valve Pilot actually did: it compares speed with valve-gear position as reported by a calibrated cam to get the 'needles' that are supposed to be matched to actually line up in the speed range where the information would be valuable. This is very different from the automatic cutoff control that was tried in the early 1920s, which used back-pressure in the exhaust as the 'control signal' and therefore could modify the cutoff directly to suit developed power demand, not just economize on steam for a particular speed.
There was a 14-page article on the system in Classic Trains Steam Glory 3, and a very good thread here on the forums in 2013 which explained how the magic cam profile was determined.
There have been some discussions, a few here, about why the Pennsylvania for example never bothered with the device, even on the T1s which were fitted with more accurate mechanical speedometers. They were an added capital and maintenance expense, and I think we could probably get a history of how and why the railroads that used them were 'sold' on various aspects of its results.
I mention the Valve Pilot as a tool for determining cut-off because the original post was on the subject of using steam expansively and setting the cut-off appropriately is a necessary requirement for using steam expansively.
This is very right, although I do have to wonder to what extent it was, in fact, an excuse for getting a speed recorder on locomotives so that speeding and discipline could be better documented! I am particularly interested in seeing the cam used on B&LE 643 (now going to restoration at Age of Steam) as the practical use of cutoff in that relatively low-drivered engine in its 'usual' anticipated service might be more than usually valuable economically in a different sense than, say, on a NYC Hudson.
I saw Boston & Maine 4-6-2 3713 "The Costitution" in Boston a number of times and rode behind her on her last run to Portland --- the 3713 had a Valve Pilot. The engine is being restored at a snail's pace at Steamtown in Scranton, PA.
We have offered access to the basically-intact equipment installed on Frisco 3751 when the time comes to 'go through' the Valve Pilot on 3713. My suspicion is that producing an operating Valve Pilot is decidedly secondary to getting the engine back in service, as I think for most modern excursions the device is much more a luxury than a necessary operating tool. That won't stop me from volunteering to work on restoring it functionally, though...
I wonder if any restored, operating steam locomotives today are equipped with a valve pilot?
I'd suspect there are, but I'd be surprised if anyone actually fires it up and uses it. There are better methods of comparing various parameters to adjust cutoff than were possible for 'pure analog' devices from that era.
On the other hand (no pun intended, but on reflection pretty good!) there is probably no better way to display the information than as the Valve Pilot does, a simple needle that is simply matched to that on the speedometer, without having to confer back and forth between different indicators. It might be possible to develop something along the lines of a four-needle stoker-engine display to include matching needles for cutoff and back pressure, which would increase 'economy' for something the Valve Pilot doesn't: actual load changes associated with changes in grade profile or train resistance.
The implementation of automatic throttle control became vastly more simple with the advent of air throttle systems, like the ThrottleMaster or the Franklin Precision equipment on the PRR T1s. That would apply to load-following (no longer just back-pressure servo) as well as automatic train-control stops of freight consists. Note the relative (at least implied) assumption that a 'modern' superheated engine is driven by opening the throttle as wide as possible, as quick as possible, and subsequently only adjusting the cutoff to get power that balances train resistance. This would characterize operation at almost any range where the Valve Pilot indications would be valuable for economy. The questions then arise whether the indicated setting is "best" for required draft (as would not have been the case on UP 800s in part of the late Thirties) or for tolerable slipping or peak stresses in operating gear. (Compression indication and control is, I think, quite a separate issue and best handled with its own indicator, controls, and valves.)
Three Cylinder Limited Cut-Off on the PRR
It may be of interest that the Pennsylvania RR considered a three-cylinder locomotive to achieve the economy of the limited cut-off on the I1s 2-10-0.
Following is from the Introduction to Test Report 31 dated 1919 of an I1s at the Altoona Testing Plant. (The first I1s was under test at Altoona from December 1916 to March 1917 -- this report is of the form of the I1s after modifications were made based on the initial tests.)
"It has been the custom to design locomotives with cylinders small enough to be oprated at a cut-off of 90 per cent, or more, without going beyond the limit of adhesion, and while efforts are made to operate these locomotives economically, it is often found that the requirements of the service make it necessary to perform the larger portion of the work at a cut-off far beyond the economical limits.
Many of our locomotives, in helping service especially, are worked, almost continuously, with a cut-off near the end of the stroke and if they were designed as to work at but 50 per cent cut-off, when in full gear, without a sacrifice of drawbar pull, there would be gained the difference between the coal and water rates at full stroke cut-off and those at half stroke, or a saving of approximately 25 per cent.
In a lesser degree, it follows that wherever our present long cut-off locomotives are being operated at a cut-off short of full stroke perhaps, but beyond the most economical range, a savings in fuel and costs can be made by eliminating the range of cut-off within which the water rate of the cylinders is excessive.
One of the principal means proposed to accomplice the desired results in steam and coal economies which are possible with restricted cut-off, has been the use of three cylinders cutting off at half stroke, but the mechanical complications and undesirable features of such a scheme are so great that a solution has been sought in other directions." (That is, a limited cut-off 2-10-0)
The following is from Conclusions: "The engines of this locomotive, compared with long cut-off locomotives as to economy in coal and steam, have met expectations. The tests have shown that restricting of the cut-off has had the desired effect in that, in full gear, where the bulk of the work is done, this locomotive operates much more economically than the L1s. This advantage, as expected, is reduced as the engines are cut back, but it is not until below the most economical cut-off for both locomotives that the L1s becomes more economical, at a given horsepower, than the I1s." (The L1s is the comparison because it was a reason for the 2-10-0: "The locomotive needed at a number of points on the lines is one having a drawbar pull 25 per cent greater than that of the class L1s, or having a pull of about 75,000 pounds at 10 miles per hour.")
Lots of fun assumptions and implications in that test report!
The most interesting is that there doesn't appear to be any consideration of either ride quality or propensity to slip in the assessment; it is purely a mention of conservation of fuel and water when the locomotive is fully loaded (meaning operating 'down in the corner' at full throttle, almost certainly)
If we think of the I1s as a logical extension of a L1 lollipop (which it essentially is, in the same sense a Chapelon 4-8-0 is an extension of a Pacific) then it makes some sense to try the implementation of extreme limited cutoff to ensure ... well, the absence of temptation to flog the engine with longer cutoff to get too much load over the road. And, perhaps, the temptation to rate the engine with too much tonnage based strictly on its indicated cylinder power and adhesion weight.
There are, still, better answers to this quandary (particularly on a 2-10-0 largely intended for heavy mineral service) than 50% LC and a grudging concession of starting ports. It just requires looking at a somewhat larger picture of engine performance and maintenance. I suspect PRR was well aware of this, as we see the amusing-in-retrospect aluminum side-rod experiment tried on one of these engines, which 'otherwise' makes little if any technical sense.
Yes, the three-cylinder subject is highly interesting (including for a PRR that just built the heavy 2-8-8-0 to avoid inside cylinders) and would have allowed the long limited cutoff with far fewer operational difficulties. Can someone find the torque diagram for a three-cylinder simple with comparable cutoff setting? Has to be far less peaky, among other things.
Just remember this about the Limited cutoff engines about the Santa Fe. They were underrated even by the builders on what their true Tractive Effort produced was. There are several recorded instances of the 5011 class of 2-10-4's without helpers in Kansas taking an overtonnage train over the ruling grade without the need for a helper during WW2. The same with the Warbaby 2900 class of 4-8-4's some call them the most powerful Northern's ever produced with more power than the J's or Niagras if they ever would have been properly rated. It is estimated they were underrated by as much as 20% of their true power. When the Pennsy borrowed the Santa Fe Texas types to haul Iron Ore in the 50's the Santa Fe engines hauled about 20% more than the J class 2-10-4 of the Pennsy.
"Limited cutoff" is not the first, or even one of the more significant, things I think about in connection with late ATSF power.
I am rubbing-hands-together delighted with the prospect of having 2926 finished, as we will be able to get firsthand test data on a number of things to establish some of what these engines in 'outshopped' condition could do.
As a note, the effects of limited cutoff were probably not that dramatic for the cases named, as I suspect you'd have the reverser past the percentage of inherent cutoff reasonably quickly (and the engine would then build horsepower up to the peak of the 'curve', which as I recall was above 40-45mph for the 2900s -- somebody find the actual numbers). These had every bit the proper valve and passage design that the 3460s did not, and I suspect some of the reported high speeds for them are no exaggeration. It's a pity we can't see an unrestricted speed test run, which I suspect would involve somewhere around the 'magic cutoff' of about 40% for best achievable speed -- well past the zone of effect for any formal limited cutoff designed into the valves and gear.
The N&W J remains my all-time favorate steam locomotive, but I believe both the AT&SF Ripley 4-8-4s, especially the 2900s, and the later and last group of UP FEFs are more powerful, like 844. And are not these two groups comarable? The J and the Niagra are comparable, despite different driver diameter and other dimensions. All terrific locomotives, and it is a shame no Niagra was saved, along with some other 4-8-4s.
Beyond the Call of Duty
It does not surprise me when a steam locomotive proves to have greater capacity for work than normally required or can be pushed to exceed normal boundaries. I am reminded of tests on the New Haven in 1925 of two 4-8-2 locomotives, one with a radial stayed firebox and one with a McClellon water tube firebox. The test results included the following comment:
"These figures are for the same train loading and operating conditions, but do not represent the maximum capacity of the McClellon boiler locomotive, which is unknown on account of limits in train size imposed by track and siding conditions."
In other words, a tonnage rating, for example, may not reflect how much more a locomotive could actually do when pressed to the limit. The fact that the train the New Haven's R-2 class 3500 pulled on test was only 87 cars long and weighed 4640 equated tons and its running speed averaged only 25.3 mph (not so bad for 1925) did not indicate that was all it could do. The train was essentially what the New Haven usually ran on its Shore Line. If circumstances required it, maybe a heavier train could be pulled faster even under adverse conditions such as snow. If the engine was operated on a different railroad, with different siding lengths, and a different profile it may have done better or worse. It seems to me that the measure of a locomotive is whether it does what the buyer wanted it to do, not whether it is more powerful than what the other guys run.
Our community is FREE to join. To participate you must either login or register for an account.