Mario, that would really be interesting! It would be a good comparison showing a new machine compared to an older one.
FYI I talked yesterday to my local friend who runs a CNC shop filled with Milltronics machines, and he suggested I measure the backlash on the three axes. X axis was .0023", Y was .0021", and Z was zero backlash as far as I could tell. So the backlash is not the primary issue with this machine; the ballscrews appear to be reasonably good. I think it is the shock of the table vibrating when the cuts are aggressive that is causing the problem; I'll bet the encoders are skipping steps, or double counting pulses or something when this is going on.