Problem solved:
Apparently our programmers had initialy written their traction control software to address a bug in the initial encoder routine. Then, in one of the updates the bug was addressed by NI (or whomever). We left our fix in place not realizing the original bug had been addressed. The net result was that we took the absolute value of the direction vector (multiplying -1 by -1). We've removed our fix leaving only the NI fix, and now things are working properly.
Thanks for following along with our little drama
