Advanced Search

The sensor is a product is from Proma Elecktronica in Poland. I am helping him to test it with Linuxcnc. The idea Proma had was to make a sensor that was easy to use for DIY projects.

I am writing a component to help integrate this into Linuxcnc. I have not even checked this will compile yet but its a start
github.com/rodw-au/proma-cv2
The idea I had was to do a fast calibration whenever the Z axis is homed.

I also built a laser cutting machine based on Linux, I just need a high-precision and fast height sensor. The height needs to be controlled in the range of 0.4 - 25 mm. Do you have a closed project? I wonder what element base it is implemented on? I tried to use ad7745, but it did not show itself very well. I want EtherCAT as an interface.

Well this is exciting. I have a bit of machining to do to get this mounted on a ball screw for further testing. Hopefully tomorrow
 

Lukasz at Proma is very impatient!  

Not sure I would bother with the 'Main' contactor.

Yes, iov2 has been removed in master.
Simple cases (like yours seems to be) can be handled by a ngc remap for M6 while more complex situations can be handled by a python remap of M6 and or T.

If you just need a safety interlock you might be able to simply add some hal logic that inhibits the relevant output pins until certain conditions are met.

That error report still refers to the 'pid.s.do-pid-calcs' not being found. Maybe you attached the wrong file?

Yes, exactly. But do you know of any example configs or even discussions on how to implement a homing origin switch? I don't think I have found any.

I didn’t think the J1900 had hyper threading. I didn’t see it but I will look agian.

In BIOS hyperthreading must be disabled, and TPM also.
And in case you find "aggressive link management" regarding PCI-E, also disable.

I went through the google dock and enlisted the help of ChatGPT and this is what I did and the results.

• CPU: Intel Celeron J1900 @ 1.99GHz (4 cores)
• Kernel: 6.3.0-rt11-linuxcnc (upgraded from 6.1.0)
• NIC: Realtek RTL8111/8168/8411 (using r8168-dkms driver 8.051.02)
• Power Supply: Switched to a more stable supply from the plasma table

What I Did:

1. Installed the 6.3.0-rt11-linuxcnc kernel.
2. Installed and configured the r8168-dkms driver.
3. Disabled EEE on the NIC (r8168.EEE=0).
4. Set kernel boot parameters:

isolcpus=1,2,3 nohz_full=1,2,3 rcu_nocbs=1,2,3 pcidle.max_cstate=1 processor.max_cstate=1 idle=poll


5. Assigned IRQ 91 (NIC) to CPU core 1.
6. Set all CPUs to performance governor (persistent in rc.local).
7. Confirmed BIOS settings: C-states disabled, Turbo Boost off, EIST off.

 

I can't say I've seen crashes in Master, any more details?

Not OK, notice the red/blue vertical lines on both sides of the histogram, those are excursions outside the visible area that do not show unless explicitly requested, so
latency-histogram --nobase --sbinsize 1000 --show
will show those excursions.
Although it just might be usable for Mesa, but i would not risk it.

What version of linuxcnc?

That is OK for a Mesa card and about where mine ran for years. I'll try and remember to look at my brix tonight.
Did you do the cpu affinity?

Hi, 

I have a touch probe that requires an additional activation signal prior to probing. I've mapped motion.digital-out-00 to the pin and would like to know the best way to add M64 P0 before probing and M65 P0 after. Should that be added to each probe routine, or would it be better to use a macro for that functionality? I'm using Qtdragon_hd and Versaprobe.   

Thank you