Servo Tuning detailed How To.

More
01 Sep 2019 22:05 #143814 by Type_Zero_Design
Andy,

I'm not sure if you were referring to my post or Randy's with the oscillations but I'm guessing mine.

This is a shot of the settings in that vid.

Attachments:

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

More
01 Sep 2019 22:43 #143817 by andypugh
I have no idea who I am talking to, or about what any more.

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

More
02 Sep 2019 03:26 #143853 by Type_Zero_Design
Andy,

Sorry for the confusion, for clarity and not to muddy the waters anymore I have made my own post to continue my tuning endevor.

forum.linuxcnc.org/10-advanced-configura...ther-tc-225-retrofit

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

More
09 Oct 2019 19:50 #147592 by HueyHQ

HueyHQ wrote: On my retrofit, I had a lot of trouble tuning following all the advice I had found because I had entered an arbitrary value for the acceleration parameter (way too high). It wasn't until I bumped it down to something more in line with what I had seen in others config files, that I could even start to tune my servos. My machine works fine now, but doesn't move as fast as it did before the retrofit, so I could go back and refine the tuning with higher accel params.!

Recently, I found my X-axis was oscillating really badly when routing out complex shapes. Straight X and Y cuts were fine but diagonal was the worst, unless I really slowed down the feed.

I was never really satisfied with the tuning, so I re-visited it yesterday.
What I found is, if I reduce to P = 0.1, the movements are now as smooth as silk, leading me to believe that the PID is possibly done in the servo amps? To those who know more about this than I do, does this sound right? (I have attached the driver manual in case that sheds any light, for anyone who is THAT interested!)

Setting P = 0 causes a following error, so what would be the lowest I should set it to, or is 0.1 OK?

I've also tuned FF1 = 0.00727 and FF2 = 0.00001 to get a reasonably flat (albeit very noisy) halscope'd f-error.
Attachments:

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

More
09 Oct 2019 20:21 #147594 by tommylight
If it works at 0.1, leave it at that.
The following user(s) said Thank You: HueyHQ

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

Time to create page: 0.134 seconds
Powered by Kunena Forum