Friction for Percy

The question Percy and I have been grappling with is this: I strongly believe that 622 and 10 both spontaneously lost their quartering, causing them to lock up; how is this possible?

Let’s review what we’ve learned so far:

• There is substantial play between the original crank pins and the connecting rods, as well as lateral play between the wheels and the frame.
• If there is too much play between the crank pin and the connecting rods, the chassis will bind when pushed, but not under power. If there is little play, the running quality is unaffected.
• With the original clearances, Percy runs fine with the wheels mis-quartered by up to 6 degrees, slightly worse at 8 degrees and starts to bind at slow speeds at 10 degrees out of quarter.
• With one crank pin as much as .015″ closer to the axle than the other three, the chassis runs fine, but at .022 there is clear binding, and at .03″ the chassis locks up.
• Reducing the size of the holes in the connecting rods to 1mm while reducing the crank pins to match does not make a difference. Percy runs about the same with one driver mis-quarterd by 8 degrees, or one crank pin out by .022″ regardless of the size of the crank pins or the size of the crank pins relative to the holes in the connecting rods.
  • Something did cause a 1mm crank pin to shear off.
• Compensation seems to amplify any rough running, but does not appear to cause binding.
• Losing a connecting rod can make a wheel get out of quarter sufficient to bind it thereafter.
• The precise 90 degree angle is less important than a matching angle among the wheels.
• If one axle gets out of quarter, and the other axle is strongly connected to the wheels, then the fixed axle will tend to correct the loose axle remarkably quickly.
• When one axle is loosely connected to the wheels, a derailment can be coincident with both axles losing their quarter and bind.
• When both axles are loosely connected to the wheels, a stop and start at dead centre can cause both axles to lose their quarter and bind.

The most surprising result of all of these is that mis-quartered wheels can actually self-correct! However, apart from violence by derailment or coupling rod failure, the only cause for wheels to lose their quarter seems to be a dead-centre start with very loose wheels. Now, I’m not saying that #622 and #10 never derailed or suffered any other violent effects, but it still seems that they quietly tied themselves in knots over the course of time. Certainly, they never exhibited Percy’s blithe self-correcting behaviour.

Reflecting further on the differences between Percy and my eight-wheelers, there is also the lateral sideplay in the four wheels. My engines have scale-width frames, which admit for almost no sideplay. Indeed, it is possible, especially after painting, that they are so wide they introduce friction with the wheels.

If there is friction with the wheels, perhaps that explains the lack of self-correcting behaviour, and may even explain the spontaneous corruption. So, I designed a new keeper plate and wheels specifically so they would rub.

Sadly, while I could certainly make Percy stutter and struggle against the friction, the wheels themselves stayed true. Loosening a wheel on its axle, I found that Percy still self-corrected, although the rubbing largely stopped as the wheel moved out of gauge to stop rubbing. The loss of gauge made for a derailment, of course, but still no poor quartering.