Backlash comp causes motor stalling

[dugpits]

Yeah, the title says it all.

When I am running at rapid (and I've adjusted my rapid down some)...say anything over 100 IPM, if I am running the Y only, then bring on the X additionally, the X motor will stall out and make that terrible noise and lose position and all that fun stuff. Same thing when running X, then start Y...X will stall out. Primarily seems to be a problem with the X, but I think Y does it a little bit if I screw around with it enough.

If I turn backlash comp off, I can't stall it at any speed no matter how frantic I pound on the keys...so it has SOMETHING to do with the comp...but I don't know how to fix or why it is happening.

I've tried all kinds of different speeds for the backlash setting - doesn't seem to help.

I've adjusted my rapid down to a lower level as mentioned above - no dice.

I've adjusted my acceleration setting in motor tuning - no dice - ALTHOUGH I don't remember just how slow I put the accel...

I've loosened my gibs - no help. Tightened - no help.

It doesn't SEEM to have problems when I am at a G1 feed command...only on rapids.

Anyone else suffer this or know what I'm doing wrong?

-Doug

[TXFred]

This is happening while you're jogging manually, right?

It's a bug. The solution is, don't do that. Jog one axis at a time and you'll be OK. In my experience, the problem happens when jogging, but not when running g-code.

A better solution is to eliminate the backlash on your machine so that you can run without backlash compensation. Backlash comp is a band-aid solution anyway. It's better to eliminate the slop in the machine.

Frederic

[dugpits]

Frederic,

Unfortuntely no, it doesn't happen ONLY when I am jogging manually. (But *mostly* when manually)

When making some interpolated holes, when it came to clearance then went to rapid to the next hole, it would stall about 50% of the time on the X.

Had to have a quick hand on the estop or I would have banged up a couple of parts. For the last couple I watched really close then hit feed hold before the rapid move, manually jogged at a slow rate to take out the lash, then resumed.

I totally agree that the backlash comp is a band aid. From what I can tell, the ballnuts are giving me about .002 of my .003 backlash. That is with oversized balls. X, Y, Z have all about the same lash....003 to .0034.

The rest is with the gibs. I can move the table with my hand and it will stay about between .001 and .002 - if my wording makes sense. However, if I tighten the gibs any more it will actually get worse and it goes pretty quickly up to .010 to .015...so that's about as good as I can get it.

I plan on getting some oversize .1265 ballbearings now, the ones I have were .126. That should help.

Aarghh. Not really in the mood to rip the machine apart. Again. Not right now anyway. But if its gotta be done...

Thanks,

Doug

[photomankc]

There's only so far you can go in getting that lash cut down. Adjusting the gibs tighter just makes for stick-slip and that's even worse. There might be a little slop in the screw mounts and I'd look for that (bearings not completely seated and can move in and out a little). If you can get it down to 0.001 that's pretty liveable even without comp but really doesn't seem excessive to have 0.002 or 0.003". Have you tried backing down accelleration? Not sure in Mach but in general controllers will accelerate 1.5 to 2 times as fast to take up backlash as it is presumably almost no load on the motor to do that and it gets done faster. Might be that your accelleration is already high but crosses the line when it gets ramped up higher to take up the lash.

Just a straw to grasp at.

[ImagineRobots]

Yeah, the title says it all.

...

I've tried all kinds of different speeds for the backlash setting - doesn't seem to help.

I've adjusted my rapid down to a lower level as mentioned above - no dice.

I've adjusted my acceleration setting in motor tuning - no dice - ALTHOUGH I don't remember just how slow I put the accel...
....

-Doug

Have you tried adjusting the shuttle acceleration under the general conifg settings? Hoss video below shows how to do it:

http://www.youtube.com/watch?v=A27nOkTFYDY]Mach 3 Backlash Compensation - YouTube

[dugpits]

Have you tried adjusting the shuttle acceleration under the general conifg settings? Hoss video below shows how to do it:

Ah..no, I hadn't seen that one amazingly enough.

My shuttle was already adjusted down to .004.

However, I did start tweaking that number in conjunction with my accel, rapid speed, and % speed backlash - and found a pretty good compromise. Maxes me out at 160 IPM on rapid but that is plenty fast for me and I couldn't get it to lock up. Had to move the shuttle accel back up to .02 seconds - so I'll have to see if it effects the cutting at all.

Thanks for that!

Photoman - thanks also for the tip, I didn't realize that it would accel FASTER to take out the lash...that's very interesting. I did back it down some now also. I know that it is in the ballnut balls because with the motors powered off, I can get a very slight amount of turn on the screw by hand (measures about .002-ish table movement), but no movement when under power (locked). I also rigged a belleville on the bearing to give me a positive pressure for another check. The .1265 balls should be here next week....maybe then the lash comp won't be an issue.

Thanks again all,
Doug