Okay, so what I was expecting here was a nice basic facing operation.  Climb milling with the passes along the X-axis.  .04" on the first pass, .01" on the second.

I've been using Mach for a little while now, but this is a new one on me.  Never seen this happen before.  Only thing I can think of that changed from the previous cut to this one is that I turned on backlash comp for Y and Z axes ... but this was happening in X, not in Y or Z, and I did enough testing (x0, x-1, lather-rinse-repeat several hundred times) to verify that the adjustment in backlash was returning me to 0 on the dial indicator reliably and repeatably.  Anyway, on each pass, the X zero was moving in the positive direction by something like .098", according to the DRO. (I've got one of the early Bridgeport CNC's, it still has the mechanical DRO's on it.)

So, here's the picture.  If I can figure out how to attach a file I'll attach the XML and the gcode.

Anyone got any ideas?

(yes, I know the finish sucks - this cutter is just about worn out.  I didn't need this part to look pretty though, just be mostly flat.)

Hm.

Turned off backlash comp, re-homed and re-ran the cut - you can see where the toolmarks are slightly different, but there's no difference to the touch across the surface so I'm not too concerned with this part.

The underlying problem on the other hand - that concerns me.  I've got about .002" backlash in both Y and Z after 400 moves of 1" each.  I'd like to see that get a little lower and I was hoping that backlash compensation could help me out with that ...