How do you deal with backlash???

In the help section of PNCconf there is a reference to "backlash", and "compensation". It says there is a section in the manual, I haven't searched hard for it, but didn't find anything in a quick look?

After redoing the X axis, trying to reduce the torque required to move it. I now have about .010" of backlash. Also after doing the "test/tune" on my Y axis I have about the same amount of backlash.

I think I should try to reduce it mechanically. .010" for a machine with ballscrews is terrible. Not much better than my manual mill with leadscrews.

Can backlash be dealt with in the software?

I'll try to get some pics posted. Andy Pugh suggested this earlier, and I thought I had it figured out. I would like to minimize it mechanically if possible.

thanks,
Ken

Here is a picture of the back of the end plate with the ball screw up against the washer I installed.

In the instructions it says that the retainer ring on the ball screw should contact the bearing. There is no way for the retainer ring to fit through the hole in the end plate. Instead it was binding against the end plate. I thought I would install the washer to act as a shoulder for it to ride on.

Last night while trying to fall asleep, I was trying figure out how to machine a larger hole in the end plate. I started wondering why the end plate on the Y axis had a large through hole, and the end plate on the X axis had such a small through hole in comparison. It finally dawned on me that one of the previous owners had switched the end plates on the X axis. So this AM I pulled both X axis end plates, along with the Y end plate. Sure enough the other end plate had a larger through hole.

Here is a picture for comparison.

When I pulled the end plate on the Y axis I noticed that there was a gap between the outer race of the bearing and the shoulder that it should rest on in the pocket. Somehow the last time I assembled it I must have pushed the bearing back off of the shoulder. I reseated the bearing, measured the depth to the bearing from the front of the cap and reassembled it. I checked the depth after assembly and it was the same, so hopefully I've cured the backlash in the Y axis.

The backlash in the X axis was because the beveled spring washer was not tight against the retainer nut. It had actually started to wear the back of the nut, where it contacted the spring washer. Everything is tight now so hopefully I've cured the backlash in the X axis also.

thanks,
Ken

BigJohnT wrote:

Great! and now you know more about your mill.

John

Well you would think so. Apparently though, I am missing something.

I tested the X axis, yesterday morning. X axis moved 1.000" according to my setup and returned to between .001" and .002" for a dozen moves, while testing the X axis.

I got back to the Y axis this morning and it moved .998" according to my setup. (I'm thinking I might not everyjthing exactly square, parallel, or perpendicular, as needed.

However, when it traveled back to zero it stopped at .017". That's .017" of backlash.

The only difference in these 2 axes is that X has open ball bearings that the spring washer is pushing against. The Y axis has a sealed ball bearing and the spring washer is pushing against the seal. I checked the coupling and it is tight. I also switched the motor leads, to make sure the drive wasn't missing steps. It is running at 10 IPM, and 1 IPSPS. So I know it's not missing steps due to missed signals, stalling.

Would a "thrust" bearing make any difference? Should I try to find an "open" ball bearing, with no shield. Obviously, I've got to tear back into the Y axis, I'll look for any thing else that could have been loose. The original ball bearings were the open ones. The original owners had lost 2 of the originals, I ordered new ones from Enco and got 2 shielded bearings.

thanks,
Ken

customcutter wrote:

The Y axis has a sealed ball bearing and the spring washer is pushing against the seal.

If something is pressing directly on the seal, that will definitely cause backlash, and possible missed steps if it is binding the bearing.

BigJohnT wrote:

Maybe they lost more parts besides the bearings... the only difference between open and shielded is you can't grease a shielded bearing otherwise they are the same. Time to investigate where the slop is... can you push the Y axis and get that much backlash? If not then look for loose couplings etc.

John

Threw and indicator on the y axis this AM, and I can get about .012" of backlash just pushing/pulling along the Y axis. Looks like I'm going to have to pull the X axis stepper again, just so I can slide the table far enough off to get a peek at the ball screw nuts and hopefully see where the play is and what is going on. I was hoping I wouldn't have to do that again.

Ken