Setting up new motor/encoder on Wesel bed mill (RESOLVED)

[nop]

Hi again. I got back up to my dad's shop today. After reversing polarity of the power to the Z-axis motor and enabling direction reversal, I reconnected the motor to the machine. The Z-axis now jogs as expected.

I then re-ran autotune. It seems like it comes close to completing, or perhaps does complete, but ends with "Warning! Excessive friction detected on Z."

Looking over the manual, I'm guessing my next steps should be to set up the limit switches, then run the drag utility. Sound reasonable?

Thanks!

[cncsnw]

Yes. Verify that you have the motor revs/inch and encoder counts/rev correct, so that it really does move an inch when it says it moved an inch.

Then set up the limit switches, and verify that you can slow jog away from a tripped limit, but not towards one.

Then let it home off the Z+ switch (either with the built-in sequence, or by running "M92/Z" followed by "M26/Z" in MDI).

Then find out what the software travel limit should be, and enter that in the Travel (-) column of the Machine Configuration -> Jog Parameters table.

Then run the drag test.

[cncsnw]

Looking back at the history in this thread, it is very likely (since you had to change the direction reversal setting) that you will need to swap the wiring connections to the plus and minus limit switches.

If there is a terminal block in a cabinet somewhere, you can do it there.

Otherwise you can do it in the switch body itself (assuming you have a dual switch unit so both sets of wires go to the same switch body).

[nop]

Thanks for the quick and detailed response. I'll do what you described and report back ASAP.

[nop]

It wouldn't have been hard to change the wiring in the cabinet, but I ended up swapping the limit switches on the machine, which seemed easiest. No problems.

Thanks again for providing detailed instructions. They made setting Z home a snap.

Re-running the drag test on Z, I still get an "excessive drag" error. The other two axes check out fine. Kind of a bummer.

I should re-iterate that this is a bed mill, so the Z motor is responsible for some literal heavy lifting. My dad installed counterweights to eliminate this issue (but not immediately, which is why he thinks the old motor expired). I wouldn't be surprised if he hadn't done an autotune after doing so. He may have, though- he can't remember. Anyway, I thought I'd try the drag test without the counterweights installed, and also with only half the weight. Unfortunately, I get the same "excessive drag" error for both scenarios.

Is there a procedure I can perform to troubleshoot the reported excessive drag, or is this likely a problem with the motor?

[cncsnw]

After the Drag Test has finished, you can press ESC back to the main screen; press F2/Load; enter ".." to go up one directory level so you are looking at C:\CNC7; arrow over to highlight "DRAG_Z.OUT"; and press F8/Graph.

That will show you how high the drag is, compared to a reference box at plus and minus 25% torque, mapped over the axis travel. Since this is your Z, the left side of the graph will be maximum Z-, and the right side will be machine zero. It will draw separate traces for the downward move and the upward move.

[martyscncgarage]

Can you take a picture of the PLOT and post it here? I would be interested to see it and how Marc interprets it.
Marty

[nop]

OK, here we go. Thanks for helping with this, Marc.

[cncsnw]

That's pretty bad.

Plus or minus 32 units would be 100% motor torque, so by the time you get to the bottom few inches of travel, you are using more than half your motor torque just to overcome drag/binding/resistance. Even at the top of travel it is still taking about 1/3 of your available torque just to move.

At 10.0 revs/inch with 40 in-lb motors there is no excuse for that much drag, even with the limited current capacity of the M15 servo drive.

It is not a counterweight problem. You can tell that because the plot is fairly symmetric top to bottom. On the way down it is taking far too much motor torque to push the head down; and as soon as you turn around to come back up, it takes equally excessive torque to pull it back up.

Does the Z axis ballscrew have a support bearing at the bottom, or is the bottom end of the screw free? If there is a bearing, is it perhaps significantly misaligned, so that the farther down you move the ballnut, the more it puts the screw in a bind?

[nop]

Hey, I think we got it. The drag test completes without error now. How does this graph look?