Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: fusion on October 11, 2019, 09:45:24 AM
-
Hello,
My mill has about .005" backlash with ballscrews. Backlash comp works good and my round pockets and bosses are within .003 in diameter. But I noticed when I change direction at .001" increments, the little bump from backlash comp moves the table about .002 - .003" past the .001 I clicked. It's in a dead zone for a few thousands, then catches up. I think this is caused by the inertia of the table.
I have the speed set to 50%. If I lower the speed %, will it help to tame this overshoot? How slow is too slow?
Thanks!
Mike
-
Consider the the 50% value a suggested starting value. I question setting the the value higher or lower will help much. Play with the setting as you wish. Inertia of the table should not be a problem if the axis has adequate power to move it for desired operating conditions.
Backlash can vary, no backlash would be ideal, so I will say that one should minimize
the backlash, but to do that, you need to know what is causing the backlash.
Consider:
- Idealy a ball screw should provide linear movement over the total length of the screw.
You would need to profile the total screw screw travel in both directions to know what you
actualy have.
- What part of the backlash is caused by the nut?
- What part is caused by support / thrust bearings? Often overlooked but can be a large
contributor to backlash.
- Any axis component involved in moving the axis contibutes in some way to backlash.
I must also remark that the steps per and tuning should have been set /tested with backlash
removed.
Recall that the smallest theoretical move that you can do is equal to 1 over the steps per unit. Unlikely that you can accurately move at theoretical value or measure that movement accurately.
In general any adjustment done ie; bearing pre-load, gib adjustment, etc should be done in a way that the total torque required to just move the load is minimized and the remaining torque is maximized.
RICH
-
Thanks Rich!