Servo Tuning detailed How To.

More
29 Aug 2019 22:42 #143612 by PCW
Replied by PCW on topic Servo Tuning detailed How To.
Sounds like is it running in torque mode ( are there tachometers at all? )

Please Log in or Create an account to join the conversation.

More
29 Aug 2019 22:49 #143613 by randypetersen
There is a "TACH" input on the driver, but I don't have anything connected to it, and I checked the old ribbon controller that was connected to the old controller and the TACH pin wasn't used either. Only thing coming from motor is Power and Encoder (i believe) and brake for Z.

Please Log in or Create an account to join the conversation.

More
30 Aug 2019 02:25 #143627 by randypetersen
Having difficulty finding any servo tuning information on servos operating in TORQUE mode. Does anyone know of any tuning resources for me?
Thanks

Please Log in or Create an account to join the conversation.

More
30 Aug 2019 03:12 #143628 by PCW
Replied by PCW on topic Servo Tuning detailed How To.
A _very_ rough guide is this (for a single PID loop/Axis)
Start with FF1=0 and FF2 = 0 and I=0

Chose a small amount of P (say 50 = full torque with 1/5" error)
Then increase D until it gets too "buzzy" and then back off about 30% on D
Then increase P until you get oscillations and again back off about 30%
Then adjust FF2 so that you neither overshoot or undershoot during
acceleration (the FF2 setting is inertia dependent so its better to do this with a typical table load)
Finally increase I until motion becomes unstable (Watch out for WIDE oscillations here)
and back off about 50%
The following user(s) said Thank You: randypetersen

Please Log in or Create an account to join the conversation.

More
30 Aug 2019 18:09 #143677 by BigJohnT
Is any of this still valid?

gnipsel.com/linuxcnc/tuning/torque.html

JT
The following user(s) said Thank You: randypetersen

Please Log in or Create an account to join the conversation.

More
30 Aug 2019 18:16 #143678 by PCW
Replied by PCW on topic Servo Tuning detailed How To.
Its still valid (just more conservative)
The following user(s) said Thank You: randypetersen

Please Log in or Create an account to join the conversation.

More
30 Aug 2019 19:06 #143684 by Type_Zero_Design
Just Bumping this up for any feedback on next step to eliminate the osculation following the command in the video linlk.

Type_Zero_Design wrote: Todd,

I posted a link to a video of my hall scope with the error in the post with my current settings a post back, but I know it's kind of an unconventional way to post a video so I've included it here as well( It won't let me upload directly)

photos.app.goo.gl/eiimAkZXWb9CGVqG6

Hopefully that's viewable for others, I'm not 100% sure it's from my Google drive.

Please Log in or Create an account to join the conversation.

More
01 Sep 2019 19:30 #143804 by andypugh
That's pretty slow to be P-term oscillation. Do you have any I term?

Please Log in or Create an account to join the conversation.

More
01 Sep 2019 19:47 #143807 by randypetersen
My X and Y are both:
P:1500
I:1000
D:4
BIAS:0
FF0:0
FF1:0
FF2: 0.1
Deadband:.0001

Z:
P:50
I:1
D: 0.7
BIAS:0
FF0:0
FF1:0
FF2: 0.02
Deadband:.0001

Still got work to do.. but performing way better now. Still has issues when command a move less than 1", so I will keep working on it, thanks for everyone's help!

Please Log in or Create an account to join the conversation.

More
01 Sep 2019 19:51 #143809 by andypugh
I=1000 with P=1500 looks wrong. I is typically a very small fraction of P.

I think you are seeing an I-term oscillation there.
The following user(s) said Thank You: Type_Zero_Design

Please Log in or Create an account to join the conversation.

Time to create page: 0.165 seconds
Powered by Kunena Forum