Following errors after upgrade.

More
24 Feb 2020 16:03 #158388 by sliptonic
I've been running linuxcnc 2.7 on my mill for a couple years. I recently upgraded to build both linuxcnc and machinekit from source so I can test development work being done in FreeCAD Path workbench.

After building LinuxCNC 2.9, all the tests pass. When I launched the mill configuration, the script prompted for automatic upgrade which I allowed. Now the mill config starts and shows gmoccapy. But when I try to enable the machine, the Z axis moves momentarily and then throws a following error.

I assume this is related to the 2.8 joint/axis merge. Since none of the joint configuration numbers changed, it must need some additional configuration for the joints that I'm missing. Any ideas?

Working machinekit configuration: github.com/sliptonic/millstone/tree/gmoccapy

Failing Linuxcnc configuration: github.com/sliptonic/millstone/tree/linuxcnc-2.9-gmoccapy

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

More
24 Feb 2020 16:15 #158391 by tommylight
Usually a discrepancy between axis/joints in the ini file.
I had such issues only once a long time ago when 2.8 was still in testing, cant recall the exact error.
Am on the phone, will try to have a look at the files later.

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

More
24 Feb 2020 17:56 #158401 by sliptonic
I found a minor discrepancy between joint and axis in the max_velocity. Corrected this but it had no effect. Still get the error.

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

More
24 Feb 2020 18:21 #158406 by PCW
Replied by PCW on topic Following errors after upgrade.
deleting all lines like:

net N-vel-fb => pid.N.command-deriv

Should fix the issue

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

More
24 Feb 2020 19:22 #158415 by Henk
What PCW said worked in my case.

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

More
24 Feb 2020 21:44 #158424 by sliptonic
Yep, that did it. Thank you. I'm curious what's going on here. Perhaps there should be a note on the upgrade wiki page.

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

More
25 Feb 2020 08:21 #158443 by bbsr_5a
is this caused by the Axis/joint conversion updater or by a new wizard config

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

More
25 Feb 2020 17:30 #158473 by Todd Zuercher
I think it was caused by an error made by the old configuration wizard, but because of an old bug in the PID component, wasn't noticed until updating to the current version where the PID comp was fixed.

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

More
25 Feb 2020 19:00 - 25 Feb 2020 19:00 #158481 by PCW
Replied by PCW on topic Following errors after upgrade.
Yep that's basically it. The old PID component ignored its command_deriv input and
always calculated it internally. The new component uses the command derivative
input if its connected to a signal, and uses an internal calculation if its command_deriv
pin is unconnected. The problem is that some versions of pncconf connect an undriven
signal to the PIDs command_deriv pin, making the FF1 tuning term inoperative.
Last edit: 25 Feb 2020 19:00 by PCW.

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

More
25 Feb 2020 20:24 #158489 by cmorley
Well I'd suggest using black magic rather then a switch for this PID behavior was the real problem.
In fact I still think it should be changed.
In the very least the black magic should have checked if there was a driving pin connected.


Chris

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

Time to create page: 0.087 seconds
Powered by Kunena Forum