pncconf 2.8
14 Jan 2021 10:57 #195303
by Jabbery
pncconf 2.8 was created by Jabbery
Hello,
I am running into a problem and lack of understanding of the jitter test.
I configured the previous version of linuxcnc and was able to get everything on my router working fairly clean, steppers stepped smooth no jitter or quivering.
In version 2.8:
Question 1 - the jitter test, what number do I insert into the wizard, it implies the max jitter of the servo thread but it expects a number around 1,000,000 so do I put max interval of the servo thread?
Question 2 - This is part of question 1, my steppers when idle keep quivering and moving. I used the same settings I used in 2.7. I also noticed the PID P=6000 and read it should be 1000 (base thread speed) this reduced it but did not eliminate it.
Question 3 - assumed 2.8 is like 2.7 the linuxcnc opening display shows axis 0,1,2,3 (xyyz) and once individually homed they change to X,Y,Z. My Y needs to be homed together.
I have no hal file yet as I am in the middle of the wizard.
Thanks
I am running into a problem and lack of understanding of the jitter test.
I configured the previous version of linuxcnc and was able to get everything on my router working fairly clean, steppers stepped smooth no jitter or quivering.
In version 2.8:
Question 1 - the jitter test, what number do I insert into the wizard, it implies the max jitter of the servo thread but it expects a number around 1,000,000 so do I put max interval of the servo thread?
Question 2 - This is part of question 1, my steppers when idle keep quivering and moving. I used the same settings I used in 2.7. I also noticed the PID P=6000 and read it should be 1000 (base thread speed) this reduced it but did not eliminate it.
Question 3 - assumed 2.8 is like 2.7 the linuxcnc opening display shows axis 0,1,2,3 (xyyz) and once individually homed they change to X,Y,Z. My Y needs to be homed together.
I have no hal file yet as I am in the middle of the wizard.
Thanks
Please Log in or Create an account to join the conversation.
- tommylight
- Away
- Moderator
Less
More
- Posts: 19219
- Thank you received: 6440
14 Jan 2021 11:12 #195304
by tommylight
Replied by tommylight on topic pncconf 2.8
Servo period should be 1000000, do not change that, and P should be 1000 for stepper systems with no feedback.
The jittering might be caused by
pid.maxerror 0005
Or something similar, am on phone, so
Edit the hal file and omit or delete lines with maxerror for each joint/axis and test again.
The jittering might be caused by
pid.maxerror 0005
Or something similar, am on phone, so
Edit the hal file and omit or delete lines with maxerror for each joint/axis and test again.
Please Log in or Create an account to join the conversation.
14 Jan 2021 15:29 #195322
by PCW
Replied by PCW on topic pncconf 2.8
P should be 1/servo_thread_period so 1000 for a
normal 1 ms (1000000 ns) servo thread period.
Pncconf should probably not make the servo thread
timing settable as this causes a lot of initial setup issues
If you have oscillation its likely you still have P
too high on some axis.
Also as tommylight says: pid.maxerror 0,0005
is incorrect for metric systems (this is a bug in older
pncconf versions) pid.maxerror 0.0005 is fine for imperial
systems, pid.maxerror 0.0127 is appropriate for metric
systems.
A P higher than 2/servo_thread_period will always
cause oscillations
normal 1 ms (1000000 ns) servo thread period.
Pncconf should probably not make the servo thread
timing settable as this causes a lot of initial setup issues
If you have oscillation its likely you still have P
too high on some axis.
Also as tommylight says: pid.maxerror 0,0005
is incorrect for metric systems (this is a bug in older
pncconf versions) pid.maxerror 0.0005 is fine for imperial
systems, pid.maxerror 0.0127 is appropriate for metric
systems.
A P higher than 2/servo_thread_period will always
cause oscillations
The following user(s) said Thank You: tommylight
Please Log in or Create an account to join the conversation.
14 Jan 2021 18:57 #195359
by Jabbery
Replied by Jabbery on topic pncconf 2.8
Thanks tommylight & PCW.
It reduced the problem but did not eliminate it. At the end of the deceleration it now quivers for about half a second. Changing any of the values on the test axis window made any difference. I started messing with the PID tab and put the value of 1 into one of the fields and the X rapid to the max and kept going so I cut power to the machine and that was the last it moved. The ethernet now shows disconnected and will not connect, I will reboot.
I have a few other issues happening slowing my testing. I will work on them later today. The home switches are producing switch errors when hit. I will check the 2.7 hal and compare to the 2.8 hal to see. I did notice it was different. In pncconf the switches are configured as home/limit.
It reduced the problem but did not eliminate it. At the end of the deceleration it now quivers for about half a second. Changing any of the values on the test axis window made any difference. I started messing with the PID tab and put the value of 1 into one of the fields and the X rapid to the max and kept going so I cut power to the machine and that was the last it moved. The ethernet now shows disconnected and will not connect, I will reboot.
I have a few other issues happening slowing my testing. I will work on them later today. The home switches are producing switch errors when hit. I will check the 2.7 hal and compare to the 2.8 hal to see. I did notice it was different. In pncconf the switches are configured as home/limit.
Please Log in or Create an account to join the conversation.
14 Jan 2021 19:05 #195360
by PCW
Replied by PCW on topic pncconf 2.8
Can you post your hal and ini files? That would speed debugging any motion issues
Please Log in or Create an account to join the conversation.
14 Jan 2021 19:12 #195361
by Jabbery
Replied by Jabbery on topic pncconf 2.8
attached is hal & ini
Attachments:
Please Log in or Create an account to join the conversation.
14 Jan 2021 19:19 #195363
by PCW
Replied by PCW on topic pncconf 2.8
a couple errors:
#setp pid.x.error-previous-target true
this should not be commented out (for any axis)
and
on Joint 0, you have
FF1 = 0
this must be
FF1 = 1
#setp pid.x.error-previous-target true
this should not be commented out (for any axis)
and
on Joint 0, you have
FF1 = 0
this must be
FF1 = 1
Please Log in or Create an account to join the conversation.
15 Jan 2021 02:52 #195396
by Jabbery
Replied by Jabbery on topic pncconf 2.8
Thank you that solved that problem.
Please Log in or Create an account to join the conversation.
Moderators: cmorley
Time to create page: 0.072 seconds