Servo PID tuning over a large XY Table

[rainman]

Hi all,

Need some advice and opinions over a problem that I faced in my CNC cutting application.

I have a XY table (both linear motors) and of strokes 1500mm.

Both linear encoders are 1um.

I tuned both axis over 9 locations. From location nearer to 0,0 (home position) to the near end strokes and the following error I get range from 30 counts to 1000 counts.

Where the worst is at the near end location.

How do you guys deal with this kind of situation where you can have only 1 set of PID parameters?

What could have been wrong? Mechanical alighment, encoder alignment?

Need some advice from you guys.

Thanks!

[HuFlungDung]

Do you notice any binding during the travel of the axis? Any uneven loading due to worn ways, gibs that are loose, sticky or jamming up could sabotage your PID filtering, especially if the motors are underpowered, or the gain on the amplifiers is not turned up enough to give the motors the juice they need to do the work.

Also, try moderate settings of accel/decel ramps before you do the tuning, to see if you can get better consistency.

[rainman]

Do you notice any binding during the travel of the axis? Any uneven loading due to worn ways, gibs that are loose, sticky or jamming up could sabotage your PID filtering, especially if the motors are underpowered, or the gain on the amplifiers is not turned up enough to give the motors the juice they need to do the work.

Also, try moderate settings of accel/decel ramps before you do the tuning, to see if you can get better consistency.

Thanks for the reply. I'm using analog torque mode for the amplifier, is there any special thing I can setup from from there?

if let say the motors are undersized, will it have good tuning on few corners but very bad tuning on the others?

Or undersize problems should give consistent bad tuning?

Thanks.