Servo Wiring and Tuning detailed How To example Mesa 7i77.

PCW,

Good call! I have Max error in there for sure. Are these other lines other than Max error ok to have?

setp pid.x.maxoutput [JOINT_0]MAX_OUTPUT
setp pid.x.error-previous-target true
setp pid.x.maxerror .0005

only the

setp pid.N.maxerror .0005

is harmful

So I made the suggested adjustments and this is where I'm at currently. When I jog the error jumps up and then stays fairly consistent, then as soon as I let off jog it drops back in line. Identical but opposite reaction in the opposing direction.

Adding additional P causes it to go into a condition chasing a tenth back and forth. Adding any more then the .75 FF1 causes a overshoot in the opposite direction when I let off the jog.

I have attached a video of the scope results.

photos.app.goo.gl/GH7n5Lm8qcZPwpeL7

Sounds like FF1 is not dialed in
It should be set to within 1/10 1%
Too much FF1 means motion leads the command
Too little FF1 means motion lags the command

photos.app.goo.gl/VtVqi5K1CTXwk5wZ8

So my machine appears to be working fine. I go to a position and it accelerates and decelerates to get there and does't over shoot it.

When I am cutting however, I often get joint following errors.

So I thought I would revisit the tuning, since I didn't have much luck with that initially. For example, I can increase P and not see much of any difference until I get to about 500, then it starts loudly oscillating and chattering. So then I have to bump it down to 300 or so. As in the tutorial I then FF1 I adjust to about 3.5, anything much higher and it starts acting poorly as well.

Am I missing something when it comes to tuning,

I watched this video which I found interesting and long:
Not directed an LinuxCNC so may not be applicable.

Curious if any had any further resources for tuning, and when it's necessary?

Pcw,

Dose what you said here not apply to me since I'm using step/for rather than analog?

You can either calculate FF1 (10/velocity@10V) or scale the analog output
in velocity in which case FF1 = 1.0

If not how do I measure the velocity?

Note that FF1 is not like P that you adjust up till you have too much and then back off.
FF1 needs to be set accurately and the best way is to see if the feedback position leads the commanded position or lags the commanded position. This is best done with a low P setting
(this is assuming velocity mode drives, torque mode drives are tuned very differently)

Type_Zero_Design wrote:

Pcw,

Dose what you said here not apply to me since I'm using step/for rather than analog?

You can either calculate FF1 (10/velocity@10V) or scale the analog output
in velocity in which case FF1 = 1.0

If not how do I measure the velocity?

So you have a step/ dir driven machine with encoder feedback?

If not and its a plain step/dir machine (no external feedback), FF1=1 ,P=1000, and make sure the maxerror settings are commented out

I can't really seem to find out of my drives are torque or velocity, I will have to take a closer look at them when I am at the shop tomorrow. Is there any easy way to tell?
Thanks!!

PCW,

Yes I have servo machine with encoder feedback using step/dir.

Currently I have gone through and set all my PIDS and FFn to 0 as Tony recommend in his tutorial and adjusted my P up until I got oscillations after joging and then backed down until they cleared.

Then I started running FF1 bring my error in to be inline.

Currently though when I get FF1 high enough to get my Error in line, after releasing the jog I get osscilations as shown in the video in the link here: photos.app.goo.gl/eiimAkZXWb9CGVqG6

Also attached are the current running settings that were running in this video.