RPi4 & 7i96 XY2-100 Following Error at High Velocities

More
05 May 2021 23:18 #208039 by Limedodge
I checked on the scope. Neither PosX-overflow nor PosY-overflow are tripped when following error occurs

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

More
05 May 2021 23:25 #208040 by PCW
Your min and max limits do exceed the galvanometers range however

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

More
05 May 2021 23:29 #208041 by Limedodge
Yes, you're right. They should be +/-75. Total foot print is 150

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

More
06 May 2021 00:05 #208046 by Limedodge
I noticed something interesting

These are both scope screenshots from the same run. F.error vs vel_cmd vs pos_fb.

The first is running "normally" (albeit with ringing). The second SS show the moment it fails (zoomed way in)





It almost seems like the pos_fb is "late". It has already been commanded to change velocity but the fb doesn't show any action.
Attachments:

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

More
06 May 2021 00:12 #208047 by PCW
It may be that at 2 KHz on a RPI
you occasionally drop packets

You might back it down to 1 KHz and
see if you get the same errors

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

More
06 May 2021 00:28 #208048 by Limedodge
Same types of errors at 1Khz servo P=1000

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

More
06 May 2021 00:33 - 06 May 2021 00:51 #208049 by PCW
is the error random or always at the same place?

Might also see if you are losing DPLL sync

Check for packet errors by monitoring packet-error

Check for loss of DPLL sync by monitoring dpll.phase-error-us
Last edit: 06 May 2021 00:51 by PCW.

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

More
06 May 2021 01:15 #208053 by Limedodge
I'd say a little bit of both. Seems to happen at the usually sticky spots, but if I change anything (max velocity, PID settings, etc...) it moves to different spots.

There was a couple of times when I had it set right on the edge of a max velocity that would work, where it would execute the Gcode to different spots without changing any settings.

So I would say it's random, but some code is more susceptible than others

I'll see what the dpll.phase-error-us numbers and packet errors look like

Also got me thinking I haven't changed from timesyncd to NTPD on this OS yet. not sure if that will matter, but might as well get that knocked out too

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

More
06 May 2021 01:24 #208057 by PCW
Yeah timesyncd is almost guaranteed to cause this kind of issue

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

More
06 May 2021 01:33 #208059 by Limedodge
Apparently the OS image i used already had timesyncd disabled (NTPD enabled), so no change there

No lost packets

Here's the scope of the dpll_error. I have no idea what's good or bad

Attachments:

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

Moderators: PCWjmelson
Time to create page: 0.370 seconds
Powered by Kunena Forum