[EvolveES]

on our PCNC 1100 we keep it adjusted so that all three axes are less that 0.0005" if it gets above that we will re-adjust it. It takes patience, but it is possible to improve the adjustment. To get all that you can out of the machine can involve looking at all aspects, from the gib adjustment, to the bearing pre-load, to the level of the base of the machine (which i learned the hard way). The bearings you can measure the preload by indicating the end of the ball screw while changing directions back and fourth, if it moves then the bearings are not tight enough (it's kind of a pain to do the Y because of removing the bellows). if you adjust the bearings too tight the machine will make a kind of a bad grumble sound when slowly moving. I think with attention to detail you can achieve much better performance than the tormach baseline specs. We have had little difficulty hitting +/-0.001 and we rae new to machining also. good luck

[adamvs]

I'd be exstatic to get .0005.

My backlash / lost motion is:
X: .0015-.002 ( it is more in the middle of the travel, not sure what that's about )
Y: .0012
Z: .0004

My Y axis needs some gib adjustment, but it is a pain as Evolve mentioned. I'm going to take another look at the preload on the X axis, It seems a little more noisey than the other two at low speed.

[adamvs]

I got into it yesterday and did the same thing MichaelHenry did to expose the X axis ball nut, and try to improve on the .002 backlash in the X-axis. It takes about 10-15 minutes to remove the shrouds, uncouple the X-axis motor, unbolt the X-axis coupler housing and motor mount. Getting the gib removed is a little tricky as the you need to remove the oil line ferule nut, which is easy to get to but difficult to unscrew as it is positioned in the sharp angle of the dovetail way. I have a 10mm open wrench which is ground down to something that looks like a fork, this works well enough. From there with the coupler block supported ( so you don't bend the ball screw ), and the gib removed, the table easily slides off the end.
My X-axis was pretty dirty, so I took the opportunity to clean it up. Some observations:

1) The ways coating that Tormach touts was in excellent shape, no detectable wear.
2) The ball nut mounting screws and ball nut yoke screws were actually ok. Some needed a little snugging-up, but I wouldn't say any were loose. BTW, the yoke is held down with four 6mm allen screws, which use a 5mm hex key. The Ball nut itself is mounted with six 5mm screws, which use a 4mm hex key.
3) The "hand scraped" gib, was not hand scraped in the sense you would think of on a Bridgeport mill way. It looked more hand chiseled, if anything. It seems the idea here is a gib that has limited friction by limiting the contact area on one side. Imagine a flat surface gib vs one that had a pebbled surface.
Attachment 272814 Attachment 272816 Attachment 272818

Got it oiled, put back together and went to readjust the gib. Best I could do was .0018. But I noticed I could move the table by pushing on it about .0015, I think this is due a minute gap between the gib and the way. If I tightened the gib, this 'pushing' gap went down to about .0005, but lost motion due to friction on the gib went up to about .0024! I thought maybe the there was some looseness in the Y axis gib that might account for this, but if I pushed the saddle with a 2x4 I couldn;t see anything on the DTI.
So there it is. It seems there is some sort of sweet spot between loosening the gib and creating a gap ( therefore lost motion ), and getting the least amount of backlash/ lost motion. I'm stumped. The problem seems to be in the gib, and the fact I can only get the best result by having this .0015 gap.