servo instablility

[Karl_T]

I've got a Hardinge CHNC lathe, camsoft control, galil 1840 card, dc brushed servos.

The X axis on this machine has picked up a drive instability at the end of a move. The servo will vibrate back and forth at the end of motion. Doesn't do it every time and the fequency of the vibration goes from just a couple cycles per second to very rapid, 25 cycles/sec? Amplitude is one or two degrees each way, maybe more when vibration is rapid.

So far, I've double checked all the encoder and servo wiring for obvious problems. Did a logic probe test of encoder pulses both at encoder and at Galil board. Then I re-ran WSDK servo tuning, no problems found. But I got considerably different tuning parameters - KD from 1210 to 2160;KP from 246 to 236; KI from 40 to 15. Note: never seen the vibration problem in WSDK

Any suggestions on what I should investigate next?

Karl

[One of Many]

Karl,

You mention the controller, but not what amps you are using. If the amps have their own gain settings, this could give WSDK some problems.

You should be able to play with the manual WSDK tuning features to reduce the gain and damping values. As I understand it, WSDK has a few options for tuning formats. Use what gets you close, then finish it off manually if need be.

I am not registered on the galil site, so cannot peek at the 1840 manual, but did see some on the WSDK tuning that you may already have access to.

WSDK manual

Best of luck.

DC

[Karl_T]

This is what Al-the-Man said earlier about my Galil servo amp manual tuning:

"The Galil drives are re-packaged A-M-C drives (25A8), the pots should be adjusted so P1, loop gain is fully CCW in current mode.
P2, Current limit should be adjusted fully CCW, then clockwise to the appropriate value, If the servo amp peak is 20 amps but the desired limit is 10amps then turn to 50% point.
P3 should be adjusted fully CW or back it off if the motor oscillates or is unstable.
P4 Offset, I normally hook power up to the drive with a motor attached and the analogue signal input shorted so it is equal to zero, the motor should not move, adjust until no motion. "


I guess I could try this again, but I don't see a problem in this step. Anything else to look at?

Karl

[fkaCarel]

Maybe no use to you, maybe usefull to others:

http://www.embedded.com/2000/0010/0010feat3.htm

[One of Many]

I suppose backing off P3 just a tad bit more until it stops the periodic oscillations, would be appropriate then. The rest might be handled with the manual controller tuning.

DC

[keithorr]

In the AMC documentation for the driver http://www.a-m-c.com/download/datasheet/25A8.pdf it is suggested that "switch 2 should be off for most applications.
I couldn't get my motor to stop oscillating until I turned it on while the motor was jittering around. Don't know why it worked, don't care.
With a power supply and 600oz motor from Dan Mauch, I use switch 2 and 3 on.

[Karl_T]

Thanks for the tip. And thanks for the link, I didn't have that. Turning on switch 2 makes sence when you read the manual," switch 2 on shorts out the current integrating capacitor" Too much integration is what causes the trouble.

I'm in the middle of moving the old manual lathe out to make room for the CHNC. I'll try this just as soon as I get the shop re-arrranged. This (moving) is turning in to a huge job. The whole shop has to be re laid out. And now, a lot of the electrical drops have to be moved.

Karl