Advanced Search

Careful now, I grew up 500 miles from the coast and 90 miles from the nearest town! The brown mulga snake I grew up with is supposedly the third most deadly in the world. I tripped over a log once to find one sunning himself under my chest when I landed. We both got a fright, I froze and he took off. Another time, I met one in the house coiled up 2 feet in the air ready to strike at me. Maybe I was about 10. Took me a couple of weeks before I got him with the .22 under Mum's dressing table!

Yet I only ever knew two people who got bitten, both on the foot while riding a horse. One was my grandfather and they had to hit the snake with a stick to make him let go and the other was a young girl about 12, the daughter of a drover. Both lived through the event! I don't think much venom got past the kangaroo skin upper on their RM Williams riding boots!

Its quite tricky building a custom kernel for preempt_rt. What we have done here is look at the kernel version the RT patch was released on at kernel.org, then we have to find the actual commit in the Rpi kernel git that matches the same release. If the kernel version deviates from the RT patch, the kernel will fail. So we end up with a kernel in between released versions of Debian. We also need to be ahead of the released kernels to maximize support for the Pi 5 which was released just after Linuxcnc 2.9.1 was released. Rpi are also making it hard for us since the Pi 5 because they are no longer tracking each version of the Linux kernel so sometimes, we can't upgrade to a later RT patch.

So you have to wait for Linuxcnc to release new images for a  reason. Don't try and upgrade yourself without forking thr Linuxcnc builder and compiling from source

I have no idea how Ubuntu got in the conversation, these images are based on Debian with the Rpi kernel.

The kernel version used is ahead of the standard Debian versions so things break on a dist-upgrade.

is there a way to download previously uploaded nvem-rt1052.txt from NVEM?
 

Nice!

I have a MH500w2 i'm working on! I looked through your files and tried messing around abit. But i'm afraid i can't figure out how to add what appears to be different. you have some layer of "home-x" parameters which i seem to skip in my ini (made with Mesa CT). i'll add my files if you have the time for a quick look.

I'm running LCNC 2.10 with Gmoccapy.

I just tried that, not sure if I found something useful but something strange did come up.
The pins I looked at were:
joint.0.pos-cmd
joint.0.pos-fb
cia402.0.drv-actual-position
cia402.0.drv-target-position
joint.0.motor-pos-cmd
joint.0.motor-pos-fb
cia402.0.pos-cmd
cia402.0.pos-fb
axis.x.pos-cmd
cec.0.8Axis.xactual-position
lcec.0.8Axis.xtarget-position

all seem like they're what they should be except axis.x.pos-cmd, after I sent G0 X50 in MDI mode, it was 50, everything was fine, then I switched to manual control mode, it changed to 0.625. That is also when I get the motor position error.

I took some screenshots at startup, after homing, after sending a G0X50 command and after I got the error. I don't have anywhere to host the images though, so I put them in the attachment in case that helps.


 

Did you try with halshow? That is the way to debug this.
When linuxcnc is still running, start halshow in a terminal window. The machine can have e-stopped, that's ok.
Just type halshow at the prompt.
Navigate through pins and investigate the values of the pins outlined above.
This should hopefully give an understanding of where things go wrong.

I can't jog or do anything after running a program, I get motor position error on xyz axes afterwards and have to restart linuxcnc before I can reconnect to the machine. Same thing if I run a few lines in MDI mode and switch back to manual control.

Yes the x, y and z axes are driven by leadscrews.

Yes they are stepper motors.

I did double check the unit last week, we have everything set in mm. I wish it was as simple as that.

There is also an index to this thread located here.
forum.linuxcnc.org/38-general-linuxcnc-q...official-images-only

I've calmed down now and want to try the ecad.so module.
I searched around a bit after your code, but don't know how to pull code for ecad, or how to build it.
Can you provide some instruction?
Also, do I need your version of lcec and lcec_conf? Seems old, it didn't recognize float-ieee haltype.
And you have Ethercat master code, do I need that or can I use the distributed apt packaged one?

I wonder did you fully reverse engineer the commands for velocity mode?
Are you sending exactly the same values as the original controllers do for the selected RPM?
When I had a look at it last time I found out that the values don't directly translate to the requested RPM value, at least there's no linear mapping, it is a little bit off from time to time.

Read through this thread, there is an example in this very thread.

you should have homing to index enabled and working, then you have to align gantry to be perfect square, when homed. And if you use velocity mode with PID, then the linuxcnc uses real feedback from motor encoders. So if you disable servos, and move axis by hand, the position is reflected in linuxcnc and you dont have to re-home. Same with gantry. it will square back.

The best option would be to add 'XYZBC' mode to 'probe_basic.ui'. If that is not something you can do then you'll have to convert to 'XYZAC' as you mentioned or hope that @Lcvette has a look at this issue.

I have installed the PI5 image (25-01-27) and have gotten everything working EXCEPT the hal_gpio input pins. Outputs...done, inputs... no luck. I can read the input pins with GPIO from python... but not within linuxcnc.

Any ideas?