Challenges Tuning Servos on Brother TC-225 Retrofit

More
19 Jul 2020 04:14 #175108 by PCW
It looks to me like the drive is limiting the acceleration
Did you check your drive manual for setting that may either filter
input or limit the acceleration?


To try open loop you would change (X axis for example)

net x-pos-fb <= hm2_7i76e.0.encoder.00.position
to
net x-pos-fb <= hm2_7i76e.0.stepgen.00.position-fb

(note that P should be 1000 for open loop and FF1=1)

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

More
19 Jul 2020 14:23 #175147 by Type_Zero_Design
PCW

I went through the manual and found these two parameters that might fall in line with what you were discussing. Do either one seem like something that I should look into? If so what kind of adjustment do you recommend?



Thanks for the help, I will report back my open loop results in a few.
Attachments:

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

More
19 Jul 2020 19:52 - 19 Jul 2020 19:53 #175192 by Type_Zero_Design
Quick update.
Param 7 is set to 300
Param 15 is set to 0 (off)

Attached a screenshot of the same joint0 with the change to the HAL to run open loop.
Adjusted p=1000 and left FF1=1


been reading the Chinese/english manual today and there is a section about tunning the drive. If based on the scope you think it should be tunned better itself.. what exactly am I shooting for on the scope?

**Not sure why it keeps rotating my photos when they post. Sorry. **
Attachments:
Last edit: 19 Jul 2020 19:53 by Type_Zero_Design.

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

More
19 Jul 2020 20:45 #175197 by PCW
Can you also plot the commanded and actual positions?
An error plot by itself is a bit hard to interpret

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

More
20 Jul 2020 12:22 #175249 by Type_Zero_Design
Quick note... I realized on the open loop scope I posted earlier I'm pretty sure FF1 was set at 0 so may want to ignore that.

Reran the open loop jog with the requested data points P=1000 FF1=1



Below is a Closed Loop jog with the requested data points and P=8 and FF1=1
Attachments:

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

More
23 Jul 2020 01:33 #175439 by Type_Zero_Design
Any insight on the results above? Is this an accel decel issue?

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

More
29 Jul 2020 11:57 #176366 by rmu
I wonder what kind of drive / servo system you are using? The english in the manual is hilarious.

Your halscope pictures are bit hard to read, I would try to create actual screenshots on the linuxcnc computer, and use proper scale on all channels (nothing to see on position command and feedback because zoomed out too much).

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

More
29 Jul 2020 12:32 #176377 by tommylight
The last picture looks like way to much lag or to loose of P term and/or incorrect FF1.
"Take screenshot" should be somewhere in the menu, most probably in the "accessories"

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

Time to create page: 0.261 seconds
Powered by Kunena Forum