hmm i can see the twist being a potential for the bearings not touching.. i wonder if maby the 2 issues are in some way related... i would think that you would be more likely to loose steps going up rather than down.. unless the plunge rate is too high.. but reducing the speed would eleminate that issue... maby motor tuning? ramp speed? but then again i would think it to be more likely to loose going up than down... just a couple off the wall ideas to consider... maby the screw or nut has something funny about them so that there is gobs more friction in one direction that the other? i did run into a machine at work that had an antibacklash nut, and in one direction i was smoothe, and in the other direction it would bind.. hmmm sure gets me to head scratching....
Grizzly X3, CNC Fusion Ballscrew kit, 3 500oz-in bipolar steppers, 3 203v Gecko's, Linear power supply from Hubbard CNC, Mach 3, BOBcad Pro Art V22, Rhino.