ASD-A2 EtherCAT on LinuxCNC

Hi @jirkam,
Well I do not know the truth, it may have to do with the gmoccapy screen.

Greeting
Chimeno

Hello.
I had a Delta servo service technician.
He tried to set this up.

Autotuning:
The axis portal
J-L = 4, Setting bandwidth from 25Hz to 150Hz, vibrations still occured.
Measured resonance frequency via Scope was 12,3Hz = setting parameters P1-25 = 123 and P1-26 = 1 (higher value causes worse noise) but it did not help. Then I set P1-29 = 1 to automatically find and filter low frequency vibration but vibration were still here with small movement command.

X axis
J-L = 4, Setting bandwidth from 25Hz to 150Hz, best result was around 70-80Hz bandwidth.
Measured resonance frequency via Scope was 2,5Hz = setting parameters P1-25 = 25 and P1-26 = 1 (higher value causes worse noise) but it did not help totally.
Also I tried to set P1-08 = 2 and P1-68 = 50 to make it more smooth, no big change.

When the service technician left, I tried to disassemble the servos and test them on the table.

The result is still the same:



The green curve is vel-cmd.

Is it possible that I have something wrong set up in Linuxcnc?

Thank you jirka

chimeno wrote: Hi @jirkam,
Well I do not know the truth, it may have to do with the gmoccapy screen.

Greeting
Chimeno

Hi.
I tried it in "axis" and more lines with number 5 is even when starting "axis".
Have a nice day Jirka

Hi to all.
I enclose the configuration. If anyone wanted to look.
File "YYXZ servo X PROFCOM.hal" please rename to "YYXZ servo X PROFCOM.par".
You may notice some errors.

Can anyone help me please?
Say the price "for help".

Thank you very much for your comments.
Jirka

Greetings to all.

I have some new information.
My more experienced colleague found that when "MAX_ACCELERATION" is increased to 1200, the "lcec.0.X.srv-vel-cmd" curve is smooth.



And this is what it looks like when crossing manually.



But "MAX_ACCELERATION" 1200 is too much for my machine. I need a maximum of 750.
There is still a configuration error and we can't find it.
The PID loop is tuned and the f-error is 0.015 mm. And that's too much.

Can you think of a solution?

Thank you very much and have a nice day.
Jirka

To me that looks like not enough FF2