OK, by differential, are you talking A, -A, B, -B because that is what I have. They are also shielded cable to a certain extent, the only mistake I made was to use unshielded plugs and not get the twist as close as I can to the end of the plug. So you don't think that final jump is caused by noise being picked up by the drivers as a step signal. This is my second tuning session I have all my axis's down to 5 steps max occured error during a program. My next PID tuning I will use the official method of bringing gain down to zero and bring dampening and derivative up till I get oscillation back off and then start increasing gain and the others until all is well. I have shielded cable on the way. Should I try my hardest to ground them or is shielding enough?
AshB82, its not as simple as you think. If you plan on sticking with the old pneumatic speed control you off to a good start as those VFDs are really good noise generators. The 5VDC for TTL logic is very susceptible to noise, Get yourself something like CS Labs parts don't go with the cheap stuff as they only know you till you hadn over the CC # and then all you can do is e mail the vendors and they give you stupid replies to your problems.