Backlash compensation - circles vs. squares

[ScullMFG]

Hey Guys/Gals,

I have a mill (Fryer MC-50) that is setup with an Oak board. We run a KP-3 probe that is pretty well dialed in (within .000x"). We have found that running square contours gave us a little deviation in relation to the probe measurements and x dimensions were slightly off compared to y dimensions after cutting. As an example, a 2" square would cut somewhere around 1.996 on x and 2.001 on y - We used the backlash compensation to fix this and it also fixed the measurement deviations we were seeing with the KP-3 probe.

Moving on to critical circles. With the lash comp that we entered for dialing in our square cutting, we found that interpolating circles did not produce the same result. We are basically seeing an exact mirror when doing full circular interpolation. So, in order to make circles actually round within .001", we have to remove the lash compensation that we entered for cutting squares. Is there a setting somewhere to address these things individually. Some of the parts we make are critically dimensioned and forgetting to add or remove the comp based on a feature of the part can cause missed fits and scrapped parts.

Maybe I'm just missing something completely - does anyone else experience anything like this?

Thanks!
Andy

[cnckeith]

please post a fresh report.zip so we can see all the setting you have made in the control configuration.

[cncsnw]

There is no provision for applying different amounts of backlash compensation depending on how the axis stops and reverses.

The only resolution available to you is to reduce the actual mechanical backlash, and also reduce any stick-slip tendency, so that you need less backlash compensation in all cases.