Some comments inline, instead of differently-ordered or prioritized:
Paul Milenkovic
Wardale was trying to instruct locomotive drivers (British usage, locomotive engineers in US) to arrest a slip by retarding the reverser to shorten the cutoff.
The primary issue here is the difficulty of doing that effectively under working conditions. First, the implication is that you're dealing with high required TE under uncertain conditions (in other words, close to the expected point of slip propagation) which means you want to return to previous throttle and cutoff as promptly as you departed them. This is difficult -- tedious, at best -- to accomplish merely with a screw reverse of adequate precision. whether or not the reverser has 'power assist' or even power implementation. If you have something like a lever-operated Hadfield, or one of the Franklin Precision installations with a quadranted lever (as, I believe, on NKP 765) vs. a wheel (as on most of the early NYC Hudsons) the advice is reasonably correct, as there can be as little reflected force or misapplied 'force feedback' as there is in the controls for a Hulett, and even if practice is required there are reasonable haptics for returning the reverser to steady-state high power even a beat before the physical slip has come to a stop (eliminating some of the effective control latency effects).
To me, the most obvious action is backing off on the throttle. In my car, I lift my foot off the accelerator pedal; I don't shift the transmission into a higher gear.
If steam-locomotive throttles worked as easily as accelerators, and on the same kind of compensated spring-return linkage, modulating the throttle would indeed be a preferred way to handle slips, although perhaps with slightly longer latency as steam mass flow 'downstream' of wherever the throttle is will continue to have some effect. This can in some examples of design (I consider them 'misdesign') produce pronounced continued slipping even with the throttle closed, the 'Blue Peter' accident being one particularly poignant thing. There, by the time the driver recovered the presence of mind to close the regulator, there had been sufficient carryover into the elements to keep the slip propagated to the point of severe inertial damage. A modern American engine with front-end throttle doesn't have that problem, but there may still be issues with the mass flow downstream of the throttle poppets.
And there is considerable inertia in the linkage even to a proportional air-throttle arrangement like the one on PRR T1s -- and issue with effective response from the physical throttle, due to the sequential poppet actuation and some concerns with back pressure on required opening force. Most of the time it is relatively difficult to get a smooth, proportional 'feel' to a steam-locomotive throttle, and it can be especially difficult to modulate the thing effectively between 'closed enough to stop the slip' and 'open exactly as wide as I had it before'. (There are some ways around both this and the problem of quick reverse changes followed by positive return to previous vernier setting, but they're probably too technical for interest here)
On the other hand, maybe working the reverser makes the dips in the cyclic steam-locomotive torque curve even lower and halts a slip that way?
It's the peaks, not the dips, that matter. Controlling the excursion of effective pressure is more important than the concomitant shortening of the torque peaks (and the implied larger throttle opening/mass flow needed to maintain torque with the gear in such position). Shorter peaks give lower average power, which gives lower average torque, which implies shorter and less-sustainable slip duration.
The idea of letting the wheel slip to get traction in mud or deep snow is one of those vain hopes -- my wheels are spinning, but I still have some forward momentum so if I keep going, maybe I can make it?
The point is that you're maintaining some forward tractive effort -- it may be at a lower percentage than with a 'static coefficient of engaging friction' but still at a percentage of the greater 'available horsepower' represented by the higher rotational speed. It's only justifiable -- to the extent it's justifiable at all -- if done in the expectation that slips are being caused by short or transient conditions and once the section in question has been passed, or the train accelerated out of a speed range where the slip recovery manifestly involves ability to start, the problem will stop.
It's a dumb practice anywhere and any time, and you can expect an actual owner of a given steam locomotive to react strongly to avoid or prevent it if they can.
However, this is very different from the idea of 'microslipping' or creep, which became a staple of diesel-electric operation as early as the 1970s. There the idea is to permit torque well above that corresponding to a given conservative adhesion limit, but arrest any detected slip within a tiny fraction of a rotation to re-establish physical adhesion. The sum of the torque communicated through the railhead adds up to be 'greater' than the average continuous torque that could be supported shy of any slip -- or so goes the explanation. It evidently works, and doesn't produce too much wheeltread or railhead damage or lateral corrugation or more railroads would eschew it...
Doing this level of precise slip arrest on a steam locomotive of conventional design is difficult. About the only type I've seen that could do this via the throttle or reverse would be the Besler motor locomotive. What's needed is something that can implement a quick clamping of the driver(s) involved, and equally fast release of the applied additional friction; I've proposed several ways to implement this including the likely method for T1 replica 5550. The best use of this would NOT be in conjunction with either the reverser control or the throttle per se; you'd either associate it with the same detection that triggers the wheelslip light on a given engine, or use a pushbutton 'shot' like a Franklin air-actuated booster control.
[quote]With respect to the variability of adhesion, which applies to the varying conditions for roads, for hayfields and for rails, John Kneiling wrote in Integral Train Systems that 15% adhesion was what you could count on, regardless of "what the vendor is telling you."[quote]
Traditionally, of course, any effective 'variability of adhesion' would have to be factored into the assigned train resistance for any particular run, with an implied tolerance for a certain amount of transient slipping -- or an understanding and acceptance of potential slowing or delays regarding from an enforced doubling, need for assisted 'shove', etc. In the old days before 'lazy' PSR this would have been fairly well known 'empirically' when assigning power to trains knowing the likely prevailing conditions. "15%" translates into a FA of 8, and if you tried to sell a locomotive so designed to any business-oriented railroad they would laugh you out of the office.
Which is not to say I haven't had my arguments with prospective manufacturers who think grossly-derated traction motors and the like are 'attractive propositions to preclude slipping'. Those that didn't listen got woodshedded by the buyer community in fairly short order!
This would be for diesel electrics, which don't connect the powered axles to rotate at the same speed. You are essentially going by a worst-case low value because once one axle slips, the remaining axles need to pick up the train load and eventually all axles will slip.
What normally happens if one axle slips is that the reduced TE quickly produces a fall in speed. That by itself precludes additional slipping until the torque 'rebalances' at the lower speed -- at which point the slipping axle is likely to have recovered merely because of the decrease in speed. So the more likely scenario is that you continue to see various axles 'blink' into slip and then appear to recover, as the speed falls, rather than massive and progressive slip until the train stops cold. Of course if the train is so overloaded that the axles slip even at lowest speed, even instantaneous speed, you'll get a stall. I'm tempted to add that only the stupid allow protracted slip, of axles or whole locomotives, of the type you sometimes see graphically illustrated in videos or 'after' pix.
One of the selling points of the Krauss-Maffei diesel hydraulic locomotive was a 30% adhesion factor. The three axles in each truck were interconnected by drive shafts to the single torque converter transmission link to a diesel prime mover -- there were two diesel engines in that locomotive, one for each truck.
There are more problems with this operationally than may first appear -- you have to consider the shock of torque coming on and off the bearings in the Cardan-shaft universals, and on the effective contact patch at the crowns of the gear teeth. Since there is usually low shock absorption, even to 'quill drive' spring standards without damping, in the whole drivetrain of a diesel-hydraulic downstream of the converter array, you should not be surprised to see bad effects, particularly as there will be a tendency for a particular axle 'at the point of slipping' to repeatedly bang its bearings and teeth, perhaps several times per second, at the full torque differential between slip and adhesion. Issues as you'd expect get dramatically worse as loose motion begins to open up at the universals, teeth begin to spall or fret, and foreign matter begins to affect the tribology.
You may have noted the rousing success of conjugated truck axles in North American railway practice since the early 1960s. This despite its relative success in a number of other fields, including HSR with low unsprung mass (which almost demands some kind of shaft final drive even with small PM-AC traction motors).
... the wheels needed to be turned on the shop lathe to within 1 mm of the same diameter -- a rod-drive steam locomotive along with the one-inverter-per-truck EMD design for their early AC-drive locomotives have a similar requirement.
All this is correct -- note that you could happily operate with even substantially-mismatched wheel diameters, just without having excessive microslipping and interaxle drive-means shock and wear. Note that the ACE 3000 in particular called for regular 'underfloor' dressing of the driving-wheel profile even though the effective 'higher wear' (presumably up to an equivalent of a 'magic wear rate' for freight-car wheels) would result in more frequent changeouts of the 'one wear' cast drivers...
The figure of 25% for a two-cylinder steam locomotive probably takes into account cyclic torque, and for three-cylinder steam with smoother torque, I heard 30% adhesion claimed as with the shaft-drive diesels.
There are some arguments about the adjustment to a "FA of 4.00 criterion" when using something giving better cylinder operation, like good RC poppet-valve gear or Willoteaux valves. Most of these have at least the tacit effect of increasing the effective pressure peak, or shortening its effective duration, during much of the anticipated 'higher economy' operation. Usually the response from the design community mirrored Wardale's installation of that ridiculous Herdner valve: throw away 'economy' when starting integrity, or lower slipping in poor conditions, became an emergent priority.
Keep in mind that many three-cylinder designs don't have the geometric elegance of a pure Swiss drive. Much of the effect on 'average wheelrim torque' can, and turns out to be, practically achievable with a two-cylinder DA if you control the rise and peak of cylinder thrust more carefully. A method to accomplish this was inherent in the operation of the Franklin type D 'automatic cutoff' setup as implemented in the Army 2-8-0 project ... admittedly, not very sophisticated from an armchair-thermodynamicist's point of view!
... the maintenance requirement for wheels was probably the least of the troubles Rio Grande and Southern Pacific experienced with their KM's.
It wasn't really mentioned as a prime concern in any of the early discussions I had on these in the early '70s, when perhaps there was still some future for large conjugated-truck power. (We might remember that in this context the Ingalls Shipbuilding 2000hp passenger-unit proposal also used conjugated trucks like this, and a modern locomotive using a Bowes drive instead of torque converters would have been practical at early-Seventies horsepower levels)
The initial thing that would have killed them was Germanic service requirements. Under the guarantee, all the lubricant in the multiple final drives had to be changed every thirty days ... almost certainly using some German-spec lubricant with relatively colossal delivered cost. There was a comparable requirement for the fluid used in the converters. All the shaft balances and universal-bearing lubrication had to be checked and (probably expensively) kept patent on some regular basis. (This was against the requirement for a typical nose-suspended alternative: run some Crater into the gearcase and let it go until you have to service the wheelset for excessive wear...)
Then there was the fun if you had to do any kind of switching, or even reversing, with one of these things. If you think a GE is slow to load, imagine having to wait while all four transmissions mechanically reversed (a Trains article on them referred to "thump ... thump ... thump ...thump"). Of course we all, of a certain age, remember the grave consequences for the torque converter if we reversed while still rolling forward ... so be sure the engine is completely stopped and won't be bumped by, say, slack rollout with the converter engaged to the driveline. The omnipresent threat of driveline or converter damage if you were so imprudent to rush this stately process ... which you will notice involves repeated stops of the consist using full service-brake application ... which of course wouldn't be covered by the guarantee is again a whole world of consideration you don't have on a practical diesel-electric.
Of course with AC drive and computers to catch a slip much faster than any human locomotive driver, 40% adhesion is claimed.
Again, the practical import of this begins, and ends, with the people determining load factors for practical trains. It is much likelier that the practical effects of higher adhesion are manifested in very-slow-speed operation than in allowing 'heavier' trains on grades ... but any railroad thinking it could, in fact, operate heavier consists taking advantage of 'better' adhesion would pretty quickly have to test that assumption against reality -- or train its crews to know when to reject a given job with a given power consist...
... on the subject of transfering weight to get more traction, Wardale writes that the Chinese QJs had a "traction booster" that was an air cylinder that transfered weight through the equalizer beams from the carrying axles to the drivers.
If it's deemed cost-effective, someone can always try it -- as noted, the concept dates back almost to the very first use of practical locomotives for long-distance service. As with boosters, there are 'vogues' for traction increasing that die out as the cost and relative worth change (in North America, with better technologies and increasing engine size -- see precise riding cutoffs as another unfortunate casualty with periodic attempts at functional renaissance.
Someone on this Forum once mentioned that the 6-axle AC locomotives having only 4 traction motors -- apparently BNSF uses them in intermodal service -- have a similar thing. The idea is that you overload the drivers only at low speeds to help with starting or help with not getting stuck, but the drivers have their normal loading at higher speeds where impact damage to the tracks and bridges is the limiting factor?
Well, aside from semantics, that's a logical interpretation of what that system is intended to do; we had a pretty good explanation a couple of months ago that what the system does is move the spring seats for the idler axle up in the truck frame, not 'jack' the axle itself, thereby allowing more of the weight to come on the other axles naturally without linkage concerns. This is in my opinion a much better way than throwing additional load on powered axles as most of the active 'traction increasing' methods in legacy steam practice appear to have done.
I don't remember the full mechanical details, but it's easy to check: one of the earlier implementations of 'articulated power' in Germany around the beginning of the 20th Century involved a jackable axle, I think a powered one, in one of the engine's trucks. You can probably get to a copy of Wiener for the bloody details quicker than I can. There too, apparently (at least as I recall) there were ride-quality issues that nipped the idea in the bud fairly quickly.
The practical value of the 'traction increaser' in modern diesel practice might be assessed by noting that I haven't seen any discussion about removing (or disabling) the apparatus from the engines equipped with it, but equally no deep discussion about implementing something like it on the 'other' type of idler-axle C-truck replacement (the EMD 1-B+B-1 approach)