Advanced Search

That got me up - but now I've got "Unexpected realtime delay on task 0 with period 1000000"

Latency test gives 60624 ns jitter on the servo thread and 70111 ns jitter on the base thread.

I've been using Linuxcnc for the past 5-6 years (I use to be Clark22 but I got locked out).   More accurately, I've been making baby steps towards getting my cnc to work properly before taking an extended break (build the machine and get it moving, stop for a long break, get the spindle working properly, take a break, get endstops working then take a break ......).  The last big step was getting the spring probe and edge probe to play nicely with each other and then to get the auto tool change and updated tool.table working.  I've don't tend to ask for help but rather keep plugging at it until I get it figured out.  The past month has been going around and around, but today I finally got it working.  I've seen you ask for people to test out the probe_routine.py and the versa_probe.ui and I wanted to let you know that I got it working today and I really like how it's doing.  I have a lot more testing on it to make sure it's actually working but after a month the tool.tbl finally updates.  For anyone else wanting to try these out there's on part that was causing a lot of confusion for me.  the python_probe.py goes in the /usr/lib/python3/dist-packages/qtvcp/widgets/probe_routines.py directory and the versa_probe.ui goes in the /usr/share/qtvcp/widgets_ui directory.  For some reason I had the versa_probe.us in the first directory (who knows how it ended up in there). I thought clarifying that might help somebody.

That is not really helpful, need to know the QtPlasmaC version in the window's header.

For what it's worth, I ran your G-Code on v2.10-009.071, it runs fine.

I added a feature in 2.10 that changes the "VEL:" label to show when a reduction is present. In this case, the line read "VEL@60%: 1020"

It's a program that's intended to pause machine motion (halui.program.pause) anytime the wireless probe controller detects an error (hm2_5i25.0.7i84.0.2.input-24 == 1).

It creates a few HAL pins, one of them being: c.newpin("show-message",hal.HAL_BIT, hal.HAL_OUT) # -> qtdragon.probe-error

I added code to the probe_routines.py file which will set motion.digital-out-02 anytime the probe is in motion (through the MDI command 'M65 P02').

I've been working on the best way to implement error handling while a wireless probe is in motion. I initially tried setting HAL pins directly by patching the probe_routines.py file, but I couldn't get that to work. I then tried tying the same QTDragon dialog to motion.digital-out-00 and then sending a 'G64 P00' in the probe_routines.py before any probe movement. This didn't work as well - I got the same 'HAL: ERROR: signal 'probe-error-msg' already has output pin'. Also, since 'ACTION.CALL_MDI' pauses the UI, I don't think that this method would have worked anyways.

Currently, I have rudimentary error handling implemented by checking a digital input tied to the probe error output before the start of any probe motion (all within probe_routines.py). The subroutine will wait up to 5 seconds for the error to be handled, and will abort if there is no resolution. I tried adding a user interaction dialog box into this routine but I wasn't ever able to display a message to the user while the routine was active. Another issue with this approach is that the subroutine won't halt motion if the probe errors out after a MDI command has been sent, but before the probe has finished its motion.

The idea behind the newest implementation is that anytime the probe is in motion and an error is detected, machine motion will be paused, the user will be prompted for intervention and upon clearing the fault, the routine will resume.

Until fairly recently, I was using probe basic with .ngc files for all the subroutines. I updated and switch to QTDragon, at which point I tried (unsuccessfully) moving over my probe error handling to the python based probe routines.

If there is a different implementation you could think of, I'd love to hear it. I haven't been able to find a solution that's already been implemented. 

Might have helped if you mentioned this to begin with. You really wasted a lot of people's time not giving the whole story.