shadowmage45
Motor / torque integration has a +/- (l/r) bias
Can be seen during accel with a motor; always pulls to the right. If motors are reversed, pulls to the left.
Likely there is a >=0 check somewhere that should be only >, or other similar discrepancy in handling between positive, negative, and zero values.
Could also just be floating point imprecision, but the consistency of the problem points to it being fixable in one form or another.
To test -- does the KF-TrackSimple exhibit these problems? It is rigged in such a way as that wheel colliders will be oriented in the same direction when the part is used in symmetry.
If the KF-TrackSimple exhibits left/right bias problems, then the error is not in KSPWheel or its physics model, but something deeper-rooted in the Unity and/or PhysX code.
Might still be fixable (read: worked-around in hacky fashion) if the root source of it can be found.
Hello Shadowmage45, this bug is very important to me. I love KSPWheels. Thank you for doing all this stuff. So I tested a lot. There is no left/right bias in KF-TrackSimple. So I assume we can find a fix here. I activated KFWheel Debugging options and watched the numbers. The Problem seems to be uneven angle, torque and velocity numbers. I have made a video for you. Watch it in .25 Speed and pause on the numbers when the rover is pulling to the right. There you can see that a/t/v is very much off. I hope this can help to fix code here. Thank you for reading. I wish you a merry christmas and a happy new year. With best regards - Martin
Video: https://www.youtube.com/watch?v=mPgNLDNlgwY
