LastspikemikeThere will be a speed at which the locomotive is steered by gravity in a superelevated gauge. Probably quite a low speed. Neither side wheel flanges will need to exert steering force.
The (gravity-driven) lateral resultant of the coning in the engine-truck wheels allows it to steer both the truck frame and the front of the chassis into curves even if there is insufficient 'passenger' superelevation. This in turn simplifies the job of the leading wheels of the rigid coupled wheelbase. If the curve is correctly spiraled the progressive transition up the flange fillet gives substantial stable restoring force before actual flange to railhead-side contact.
AAR shot movies of this in the 1930s, and it would be much easier to make a video record of guiding today. I confess I'm looking forward to watching this on the TTCI Fast Loop while following the data off the instrumented wheelsets on 5550.
LastspikemikeWe're sliding a long way off your original topic, which is still worthy of further exploration, imho.
The moment you introduce momentum on a curve, gravity ceases to be the only perturbing force; the moment you introduce sprung lateral... as in a good lateral-motion device on the first driver pair, where one is essential... you have additional concerns that gravity only indirectly affects. Consider what shock absorbers/dampers are for if gravity is supposed to be the only 'restoring' force... and what fills in in their absence on functioning high-speed locomotives.
Keep in mind that absent very stiff lateral on the engine truck, the leading driver pair (and its flanges, unless you're very careful) does an important part of actually steering the chassis. Here a careful inspection of Stroudley's arrangement on the Gladstones will serve you in good stead.
In a good guiding arrangement, you want relatively stiff on-center compliance, then soft compliance well-damped for curve swing, then harder compliance (usually proportional via cams or eccentric sector gears or the like) for further shock and accommodation. I believe this, in principle, was behind the N&W testing a Fabreeka composite spring in the primary suspension of a 2-wheel lead truck, I think probably in response to that arrangement being used on pedestal tenders in the '40s.
It is interesting that Kiefer subsequently realized why the idea didn't work on tenders, and N&W must have observed comparable results as the experiment is as 'carefully undocumented' as B&O's apparently wholesale adoption of that late-'30s dream project the front-end coal stoker.
LastspikemikeCut the mass in half right through that CG and where does it go?
Cut it vertically, you'll get two new centers of mass displaced laterally but touching at the bottom. That isn't even metastable.
Cut it at an angle and you get something more interesting ... go look at Levi Bissel's original pony-truck centering idea. Which way will the system move, and will it stop spontaneously?
my understanding of why judges do not want lay people to defend themselves ... they don't understand the terminology commonly used in the court
greg - Philadelphia & Reading / Reading
LastspikemikeWe' re on the same page even though you may not think so.
hard to agree when you argue against textbook terms
LastspikemikeCG also doesn't exist but it's very convenient to think it does.
?? are you arguing symantics, should be center of mass
LastspikemikeSuperelevation permits higher cornering speeds.
the angle of superelevation results in a centripedal force, the lateral force that pulls the train toward the center of the curve and depends solely on the angle of elevation.
the (imaginary) centrifugal force depends on the speed (radial velocity) of the train around the curve.
there is a specific speed where the centripedal and centrifugal forces equal. any faster or slower and the wheel flanges rub. of course not rubbing at all is an ideal situation, but maintaining that speed minimizes the friction.
LastspikemikeThat is what I said.
you also said
LastspikemikeFor high speed rail the only reason to superelevate is to improvemaximum feasible cornering speeds.
there's a very specific reason for superelevation. what you're saying is not clear.
LastspikemikeCentripetal force is just an arithmetic convenience. It isn't real.
i think you mean centrifugal force, the outward force trying to maintain the objects linear motion. the centripedal force is due to gravity because the CG is not centered over the upward forces generated by the rails
OvermodI have always thought a significant reason for superelevation was to help preserve track geometry. Minimizing lateral force in typical track structure is a good thing...
using gravity to balance (cancel) lateral forces
LastspikemikeUnder the train. Pointing towards the center of the Earth.
obvious. but that "resultant vector" is not thru the center-line of the body and results in a lateral centripedal force opposing the centrifugal force due to inertia
In part the superelevation is to preserve the action of the wheel-rail geometry as the resultant that is "centrifugal force" acts on the center of mass. Various techniques over the years in high-speed equipment deal with this as well.
Another purpose of superelevation is to keep the vehicle stable if its suspension permits the longitudinal center of mass to deflect out. Advanced negative-cant-deficiency (tilt) systems account for this (they roll the carbody around an axis close to passenger inner-ear height, so the actual lateral deflection of center of mass is controlled or even shifted inward as the 'pendulum' action reduces lateral component of force as felt by the passengers as nearly as practical to vertical, decidedly unlike many classic passive tilt approaches (e.g. the high pivot on Cripie's TurboTrain).
I have always thought a significant reason for superelevation was to help preserve track geometry. Minimizing lateral force in typical track structure is a good thing...
LastspikemikeFor high speed rail the only reason to superelevate is to improve maximum feasible cornering speeds.
where is the force that accelerates the train toward the center of the curve?
LastspikemikeIn my mind's eye experiment, I'd like to understand how on Earth one of these machines can negotiate a steel rail curve at 100+ mph.
the speed depends on the radius of the curve (larger radius, higher speed)
Fundamentals of Railway Curve Superelevation may help with the math
LastspikemikeAdded weight, whether dead weight presumably in search of additional tractive effort or accessories needed or beneficial to add in some fashion. How much weight? Where relative to the driver axles to put the weight?
Most of what you add to a locomotive will be on the rigid frame or the boiler which moves in known ways in thermal expansion. Weights can initially be treated as beam moments relative to the frame when 'designing', and then effectively checked by weighing afterward as desired.
(Incidentally one of the supposed key secrets of Alco design was the treating of the hinged frame of a Challenger as if it were rigid in the vertical plane, with the equalization handling what was then essentially a long 4-12-4 for effective suspension ... turns out N&W did the same thing with the A but without the crowing.)
Note that the leaf springs can be arranged to tilt with the equalizers, and the arrangement on at least some locomotives is metastable -- if the engine is lifted incorrectly the equalizers can tilt to a 'stable' position that is no longer properly weight-distributing. (I believe there is a diagram in Ralph Johnson's book that shows this; the solution is to jack up the engine and level the suspension by hand, after which just lowering the weight down on the wheels fixes things...)
Leverage effects from adding weight within the driver wheelbase or outside it, outside the total wheelbase including either the pilot or trailing truck or both?
If you have a copy of the MR Locomotive Encyclopedia (steam) you can quickly compare a great many rigging arrangements.
Springing of each driver axle and the separate issue of springing the trucks. Equalization connections between sprung axles.
Brings to mind the equalizing spring pivots between the twin and triple leaf sprung axles under the mobile homes I built in the early 70's
LastspikemikeThe lead and trailing ends of the pair of leaf springs on twin axles were connected as usual by shackles to spring hangers welded to the frame.
Not a far stretch from what the locomotive designers had adopted:
Rigging_EQ by Edmund, on Flickr
Combined with the functions for the pilot wheels to both lift and guide the front end and the stabilizing effect of the trailing truck while negotiating curves and provide added support it all comes together as a well designed support and distribution system.
Straight-track by Edmund, on Flickr
Pilot-Curve-merge by Edmund, on Flickr
Regards, Ed
LastspikemikeI'm not going there. Pneumatic tires with rubber contact patches are pretty complex devices. I do know my stuff in that area.
I'm not sure about the context for the slipping driver situation. The usual engineering of steam locomotives means all drivers slip at the same speed or none of them do. Up to the limit of slip of the lightest loaded driver, no driver will slip.
Assume for a moment that torque deflection in the axles is minimal and there is no tolerance in the rod bearings, and the rods do not deflect. If the locomotive encounters oil or a patch of that peculiar plastic material that leaves become when trains run over them, some of the drivers will experience diminished -- sometimes radically diminished -- "static" coefficient of friction, while others do not. At some point the balance between resultant of thrust and adhesion friction can 'go negative' at which point the drivers will begin to slip. Very (very) shortly after this the conditions between the rotating contact patches and whatever is under them on the railhead assume the characteristic of "sliding" friction, and if nothing changes at that point, the breakaway torque would now be sufficient to accelerate the spin (this doesn't happen proportionally in reciprocating locomotives unless you have enough admission, precise enough valve gear, etc. to 'make it' around to the following admission, but it happens dramatically with electric motors or hydraulic drives) and you will have to reduce torque until the speed differential between patches is reduced enough for the asperities to engage, etc.
Now, it is also possible for enough coefficient of friction to be established to arrest differential spin as well as re-establish traction. This results in what can be severe driveline shock.
The great emergent issue with unconjugated duplex-Atlantic locomotives was that some defect in track -- an oil spot, a frog, a low rail joint or cross-level defect -- would momentarily unload a wheel. On a normal eight-coupled the other 7 drivers would help the engine 'retain its footing' but on the duplex it's 25 percent. Hence the phenomenon of high-speed slipping when the locomotive is operating at high speed at near its rated load -- one or the other engine will momentarily break and spin, the engine and train will flow down as traction is lost, and the slipping engine will re-establish itself (hopefully with no more than a bit of a bang). Note that it need not be predominantly the forward engine that lets go (as in the pictures of low-speed slipping on PRR T1s) and there is very little tendency for runaway destructive slip (as the 'other' engine does not have either the adhesion or the horsepower to keep the train at speed)
Once that limit is reached for the lightest loaded driver then weight distribution does matter because of the difference between kinetic and static friction coefficients.
The phenomenon of stiction is also related to this transition zone. Driving on ice gives an ordinarily skilled driver a good lesson on stiction effects.
Ice is a little funny because it involves near-modal change to zero coefficient of friction (and minimal viscosity) in a thin layer of melted liquid; physical displacement or refreezing of that layer rather promptly reestablishing rather strong friction... until enough melts again. The effect, though, is that friction is there and then it's not, with very little transition.
The classic weight distribution modification was on the M-1sa class where the Reading aded a box of lead on the pilot deck to weight the nose of the engine for better tracking and to reduce derailments.
Dave H. Painted side goes up. My website : wnbranch.com
LastspikemikeCounterintuitively but understood empirically even by toddlers less contact area between materials does not cause traction problems. Reason is friction is proportional to weight, well force actually.
Interestingly enough, in railroad adhesion the contact patch actually does deform elastically under axle load, and it is 'enough' larger for larger drivers that references mention the difference as significant.
It is possible to make a heavy track car perform as well as an 'optimally light' one ... if for example you put skirts, fans, and a 40hp snowmobile engine on it for downforce. But you will immediately and instinctively recognize your tire life will be measured in seconds if you do... which is also the issue with decelerating heavy road vehicles with the right weight on the right area of contact patch...
LastspikemikeIt does not contribute to traction if actually off the rail but the total friction force remains unaffected provide the drivers don't transition from static friction to kinetic friction by slipping.
agreed if the driver is actually off the rail (not likely) the weight distribution changes
if the driver is actually slipping while on the rail, the weight distribution is unchanged but the force applied by the cylinder is redistributed to the remaining wheels which is not very likely to cause them to slip.
gregcwhat types of auxilliaries? and are they trying them in different locations?
...obviously later models learned from the mistakes of the earlier.
when did "weight distribution (?)" get figured out? was weight distribution just a matter of adjusting driver or truck suspension or did it require moving components?
In the old days this was a very tedious operation. Essentially the locomotive was divided into transverse 'slices', the weights of which were then figured; side-to-side lever arms and torque moments were likewise doped out, then everything was correctly added up and any longitudinal moments calculated. Even small nominal changes might result in a need for substantial recalculation -- all of it charged to motive power, whereas impact (no pun intended) on track expense would not be.
was dead weight ever just added to a loco?
I would not be surprised to see some weights applied to one side or the other to counterbalance an auxiliary moved to a different location, e.g, compressors moved from hanging off the side of the boiler to the pilot deck. But that would still be a kludge compared to using something more beneficial.
If the railroad in question valued its track 'correctly' there would be far more attention to issues of weight distribution and transfer, various kinds of shock loading, and augment reduction than on more primitive ones...
[/quote]when did "weight distribution (?)" get figured out? was weight distribution just a matter of adjusting driver or truck suspension or did it require moving components?[/quote]All the above, as required.
In the old days this was a very tedious operation. Essentially the locomotive was divided into longitudinal 'slices', the weights of which were then figured; side-to-side lever arms and torque moments were likewise doped out, then everything was correctly added up and any longitudinal moments calculated. Even small nominal changes might result in a need for substantial recalculation -- all of it charged to motive power, whereas impact (no pun intended) on track expense would not be.
OvermodWeight distribution is also affected by where various auxiliaries and components are placed.
what types of auxilliaries? and are they trying them in different locations?
obviously later models learned from the mistakes of the earlier. when did "weight distribution (?)" get figured out?
was weight distribution just a matter of adjusting driver or truck suspension or did it require moving components?
was dead weight every just added to a loco?
ndbprrNobody has discussed the spring rigging which is critical including the lead truck and trailing truck supporting the firebox
I could go into the finer points of spring rigging and snubbing but there are people here who would pay to keep me from it...
Nobody has discussed the spring rigging which is critical including the lead truck and trailing truck supporting the firebox
Keep in mind that most of these kinds of change in modern practice involve equalized locomotives, where tinkering with the lever geometry is what determines the effective load on drivers and weight transfer to associated carrying wheels. This also fully addresses any issues with drivers supposedly 'raised' above or below their neighbors. (this was a timeless topic on steam_tech where John Knowles would advocate the English practice of individually springing each wheel 'just right' without the added weight of all those levers)
Weight distribution is also affected by where various auxiliaries and components are placed. This is sometimes discussed in detail (as with some of Larry Brashear's comments about ATSF design 'engineering' in the early '20s, for example about what was necessary for 325psi boiler pressure). Since freight motive power design in that era was more a matter of low-speed expediency (with necessary frequent track repair for other reasons) changes that gave unequal driver load might be tolerated if fixing them would cost more than the perceived inconveniences.
There is such a thing as 'tapered loading' Some references call it a misguided design principle, but I've always liked the concept in principle (one perhaps naive idea being that both axle load and peak augment would be rolled progressively onto less-than-perfect track, another being that less load on the lead driver produced lower shock going into curves).
If wheel arrangements were modified with different trucking, I would assume that the equalization would be one of the first concerns.
Reading did very famously get things awful wrong from time to time, but not (to my knowledge) with weight distribution. They very famously tried a pin-guided Adams truck under the back of an otherwise-good Atlantic (it guided and teetered more or less exactly as you'd expect it would looking at its pictures) and tried spring lateral on a 2-10-2 without damping, with likewise highly predictable and (in retrospect from a non-career-endangering perspective, comical) maiden voyage apparently culminating in some field welding...
woodoneThe one set off the rails would not have any traction, would it?
of course, it would not contribute to the tractive force of the loco
I do not understand? If you were to lift one set of drivers 1/16 inch off the rails the rest of the weight would be on the remainder of the drivers. The one set off the rails would not have any traction, would it?
LastspikemikeSince total tractive effort is not affected by weight distribution on the drivers the consideration would be the strength of the rails.
that's right, but it affects when slip occurs
what kinds of loco modification resulted in changes in weight on drivers resulting in less than ideal balance?
presumably, the ideal weight distribution is the same on all drivers, but i see in the chart below that modifications (e.g. I5 abc) sometimes exacerbate weight distributions. the weight distribution of the I5b & c are worse than the I5a. The I7b, k and i improved weight distribution. The I8b improved over the I8a but the I8sd which presumanly added a superheater resulted in an acceptable (?) degradation. There were no improvements to the I10sa which had a near ideal distribution
when did loco designers recognize the value of balanced weight on drivers?
4th 3rd 2nd 1st mx/mn tot adhWt 37627 43818 32937 33638 1.33 74.0T 16.5T I5a 41025 35391 34084 27450 1.49 69.0T 13.7T I5b 41740 37447 32404 29309 1.42 70.5T 14.7T I5c 32984 29457 38197 27072 1.41 63.9T 13.5T I7a 34087 39613 43112 32313 1.33 74.6T 16.2T I7b 42347 39409 43784 27110 1.62 76.3T 13.6T I7j 41150 36300 31475 35225 1.31 72.1T 15.7T I7k 34710 36060 41983 32672 1.28 72.7T 16.3T I7i 28737 53163 33425 32675 1.85 74.0T 14.4T I7m 47500 54250 57800 44425 1.30 102.0T 22.2T I8a 52325 56075 52150 50600 1.11 105.6T 25.3T I8b 53400 55250 50750 48875 1.13 104.1T 24.4T I8sd 63000 64690 61150 63080 1.06 126.0T 30.6T I9sa 64600 65100 63400 64500 1.03 128.8T 31.7T I9sb 70910 71510 71580 70190 1.02 142.1T 35.1T I10sa