Servo PID Tuning - can't clamp down on overshoot

Here are more screenshots

• same acceleration (10in/s^2)
• no FF2
• just dropping P from 20 to 15 to 10 to 5 to 3 (see the file names)

I was able to recenter before & after with Igain after these screenshots, but nothing seems to really help the initial lag at all.

Hmm if accceleration doesn't affect the overshoot then its not likely to be lag in the resolver
mechanical?

Just to make sure the acceleration wasn't a problem at all, we dropped the acceleration down to 3 in/s^2. Halscope screenshot is attached -- no real change. The initial lag is still around 0.002-0.003 inches.

It exactly matches the acceleration time period, and then jumps into position almost vertically. It's almost like something is holding the motors back during acceleration.

You might try increasing the drives velocity loop gain (increase till it oscillates then back off 20% or so)
while linuxCNCs P term is low to see if that helps the lag.

the fact that its unaffected by acceleration does suggest mechanical issues though

Just tried increasing "speed loop gain adjustment volume." It did help a teeny bit. We increased it 2 'gradations.' On the 3rd gradation, it started buzzing as soon as the drive was enabled, so we backed it back down to 2 gradations.

Still not changing the basic problem.

What mechanical issues would you look at?