Printer Model: Voron 0 (mini afterburner)
MCU / Printerboard: SKR Mini e3 v2
Host / SBC: Raspberry Pi 3B+, Waveshare CAN hat
klippy.log klippy.log (127.9 KB)
Describe your issue:
Hi everyone, I’m working on diagnosing an issue with my BTT EBB36 CANBUS toolhead board disconnecting after printing for some time. The printer will successfully complete prints of 2-3 hours, but shortly after finishing the connection to the toolhead board will fail and klipper will shut down (the error is “Timeout on wait for ‘tmcuart_response’ response”).
I’ve been through the hardware and software configuration thoroughly and haven’t found the problem yet, but I did notice that the bytes_invalid for the EBB36 will only rise if the extruder motor is active. So I can turn on the heaters and home the printer and it will sit forever with no problem, but if I send an extrude command through the web interface, the bytes_invalid will begin to increase and eventually klipper will shut down. I can reliably reproduce this by restarting klipper and extruding 1mm of filament using Mainsail. After a short time (5-20 minutes), klipper will shut down. Strangely it seems to take longer to shut down if it is printing as I have successfully completed several 2+ hour prints.
Probably next I will try getting a shielded cable and/or trying a new EBB36, but I wanted to see if anyone could diagnose my issue given this information about the extruder motor and bytes_invalid. Thanks for taking the time to read this.
Mmmm, canstat_EBB shows zero errors.
Log seems truncated.
That is a little strange, that there is no debug output for EBB.
Just a wild guess: Maybe there is a similar issue with the GS USB.
Like host interrupts sometimes reorder the can frames.
Then, there is a recommendation to run kernel 6.6+.
Seems similar to the gs_usb:
Probably fixed in 6.0+, according to GitHub.
That implies that you are using the kernel built in support.
It looks like you were right, I updated the linux kernel by running “sudo rpi-update stable” in the terminal. The linux kernel was Linux 5.10.103-v7+, now it is Linux 6.12.20-v7+. I have turned on the heaters and moved the axes and extruder and bytes_invalid is 0 and not increasing, so this looks very promising. I will try some longer prints next.