Track friction is wonky.

[Justin-Fisher]

Make a 45 degree angle ramp and a duplicate conveyer belt (track device) made with the same contact material, with belt velocity zero. Expected behavior would be that these two should behave identically -- a stopped conveyer might as well be a ramp. However, an object massive enough that it will barely sit still on the ramp will slide down the conveyer.

I suspect that the problem may have to do with conveyers not correctly obeying the "static friction" / "kinetic friction" distinction -- instead it seems like they typically have something akin to the "kinetic friction" of a ramp made out of the same contact materials. If a conveyer is moving upward, an object that is heavy enough to slowly slide down it will slide to the bottom and then when it stops at the bottom, it mysteriously gets a big kick to start going upward until slippage makes it slow and slide down again. I suspect that this mysterious "big kick" is static friction kicking in, even though the object is static with respect to the conveyer object itself, not with respect to the moving belt.

(I posted videos of some of this oddness in Discord, but apparently github doesn't accept such videos.)

A partial work-around can be to amp up the Coulomb friction for the conveyer, though there are obvious drawbacks and annoyances to proliferating contact materials, and I don't think this would reinstate the static/kinetic distinction on conveyers.