Ongoing lost communication issues | VORONDesign | Page 1

twilit sable Aug 7, 2025, 1:22 AM

#

I've been having issues over the last 6 months or so where my printer stops printing randomly with lost communication issues. Things I have tried to date:

Replaced the Raspberry Pi
Using the 32 bit version of the OS
Replaced the U2C
Replaced the EBB36
Replaced Cartographer running on CAN to TAP
Converted the Box Turtle from CAN to USB
Replaced the Octopus Pro with a brand new one
Reflashed the firmware on everything
Replaced the cable from the EBB36 to the U2C
Ensured the cable at the EBB36 is secured and not moving

Regardless of all of this, I'm still seeing bytes_invalid and bytes_retransmit increase and potentially fail the print.

Is there anything else I should be looking at?

This is on a Voron 2.4. Raspberry Pi 4 32bit OS. Ebb36. Box Turtle with an AFC-Lite.

sturdy plinth Aug 7, 2025, 2:01 AM

#

Can you post pictures if your toolhead, specifically the strain relief?

twilit sable Aug 7, 2025, 6:22 AM

#

Hopefully this helps?

earnest grove Aug 7, 2025, 7:15 AM

#

twilit sable I've been having issues over the last 6 months or so where my printer stops prin...

A klippy log of when those lost communication errors happen would be helpful as well. What device does it say lost communication with? How the can network setup? Where’s the 120ohm resistor jumpers setup?

twilit sable Aug 7, 2025, 8:18 AM

#

This is the most recent file that caused and error. Sometimes, its the EBB36 that loses comms, sometimes the Octopus MCU, sometimes the Turtle. There doesnt seem to be a pattern from what I can tell.

I have resistors at both the EBB36 and the U2C

📎 klippy.log.2025-08-01

fiery raptor Aug 8, 2025, 10:04 AM

#

twilit sable This is the most recent file that caused and error. Sometimes, its the EBB36 tha...

There are no lost communication incidents in this file, from what I can see

twilit sable Aug 11, 2025, 10:20 AM

#

Try this one

📎 klippy.log_1.txt

fiery raptor Aug 11, 2025, 9:08 PM

#

twilit sable Try this one

alright, well, there's definitely a disconnect there.

#

it looks like everything was fine...until it wasn't

#

and we do see shutdown commands go out to other mcus, suggesting that it really was just the one that went down

#

one thing I do note is the apparent HARD shutdown of the pi

#

afterwards

#

don't do that

#

always shutdown/reboot nicely if at all possible

twilit sable Aug 12, 2025, 7:34 AM

#

Heres another from a print thats currently ongoing

📎 klippy.log

twilit sable Aug 12, 2025, 7:07 PM

#

I seem to notice that it's happening more frequently with this sort of movement where the movement is long and continuous

atomic frigate Aug 14, 2025, 12:01 PM

#

@twilit sable can you check your txqueuelen or just run

ip -details -statistics link show can0

or

cat /etc/systemd/network/80-can0.link

or

cat /etc/network/interfaces.d/can0

depends on your config. If you txqueuelen is to low then your EBB is running out of space in the queue so you need to increase the queue capacity that will probably be the problem if your hardware is correct.

twilit sable Aug 14, 2025, 6:48 PM

#

atomic frigate <@267903912714043393> can you check your `txqueuelen` or just run ``` ip -detai...

I've tried setting it at both 128 and 1024.

I'm going to hopefully try eliminating the U2C from the chain and connecting direct to the Octopus this weekend to see if that make s a difference

atomic frigate Aug 14, 2025, 10:43 PM

#

I have a similar setup, M8P + U2C + EBB36. I had a similar problem you mentioned but my was solve with the txqueuelen to 1024 or you can increase it a bit more 4096

#

@twilit sable did you flash the EBB36 to the same birate of the U2C I use bitrate 1000000?

twilit sable Aug 15, 2025, 12:36 AM

#

Yep, bitrate is the same

#

I'll try with the txqueuelen set to 4096

twilit sable Aug 16, 2025, 12:41 AM

#

4096 didn't help. Failed in the same place

chilly elbow Aug 16, 2025, 11:05 PM

#

How is the power supply unit' cooling. I also had random printing failures solved by a fan directly blowing on the PSU when printing starts.

atomic frigate Aug 18, 2025, 5:44 AM

#

@twilit sable did you solve it? probably that fan suggested by @chilly elbow will help.

Can you try to run inputshaper? (that will move your gantry a lot, so you can check if power drops or the ebb get disconnected, you can use the built in gyro from the EBB) That test can narrow it down to find the main bug.

#

@twilit sable are those cables you connected the EBB thick enough for the 24V power? probably the distance from the PCB to the EBB is too much for a thin cable

twilit sable Aug 18, 2025, 9:31 AM

#

atomic frigate <@267903912714043393> did you solve it? probably that fan suggested by <@1368861...

I actually haven't had a chance to look over the weekend. I'll check later this week and advise.

The wiring I've had running since I built the machine a couple of years ago. This fault has come completely out of the blue, and with no resolution. I've checked for continuity and breaks throughout with no positive outcome.

Some prints are absolutely fine. It just seems to be some that it tends to disagree with. Mainly big layer times.

twilit sable Aug 20, 2025, 10:38 PM

#

Alright, so I have checked the cooling in the electronics bay and the PSU is definitely getting coverage. I've also replaced the PSU with a brand new one. Have now taken out the U2C and using the can l and h connectors on the octopus pro.
I was able to get a piece printed yesterday that had failed before. But there were still a couple of bytes invalid and bytes retransmit throughout the print. Nothing crazy but still not nothing.

Will try another piece tonight to see if it repeats.

twilit sable Aug 22, 2025, 12:14 AM

#

And....still no success 🙁

#

Will give shaketune a crack to see if I can isolate particular areas of movement

twilit sable Aug 31, 2025, 8:09 PM

#

Is it possible for failing A/B motors to be causing issues? I'm still at a loss as to what can be causing this.

#Ongoing lost communication issues