Looks like I implemented another one of those weird solutions that work
although they're not quite right. :-) I don't use any damping. I simply
calculate the absolute value of the lateral force from the slipangle, and
give it the sign of the lateral velocity of the tire instead of the sign of
the slipangle. So only the magnitude of the lateral force is delayed by the
effect of relaxation length, not the sign.
Also, I clamp tanSlipAngle at 20.0. If I didn't clamp it somewhere, it could
just increase and increase towards infinity in a full sideways slide, for
instance, and you don't want that because it could take a long time for it
to return to a reasonable value. A tanSlipAngle of 20.0 corresponds to a
slip angle of about 87 degrees, and there's practically no change in lateral
force if you go from that towards 90 degrees (which you'd only reach when
tanSlipAngle went to infinity, anyway). Woohoo, I've invented my very own
magic value now. :-)
Could also be that either one of these by itself would do the trick. I don't
really remember why I put both in there because it's been a while since I
worked on the tire model (yes, I know, I should comment my code properly
:-)
Great, yet another one. :-)
I dunno, the four OpenGL windows in Lightwave are working fine for me,
although I only have one of those ancient TNT cards. :-)
If pi were 3.5, circles would look pretty strange I think. :-)
Petri Blomqvist
