Lost communication with MCU 'nhk

[Robzilla]

Could be voodoo, but my problem with these errors went from 2 or 3 a week to once in 3 months when I replaced SB door/hatch part with a vented version (this one https://www.printables.com/model/907170-voron-clockwork2-cable-cover-with-ventholes-for-ld ). I was printing a lot of ASA & ASA-CF so 100-110 bed and 250-270 hotend temps. Had even bought a replacement high temp board to move to but haven’t had reason to switch out so it will go on a new printer next year. There’s another version with a 4010 fan on printables but a bit of wiring involved.

Firstly I'm glad that is working for you, as its very frustrating!! I have fitted heatsinks to my board, but with no change. Now I have been doing some further testing:

I have successfully printed ABS (4ish hr prints) I am conservative and these tend to be a slower feeds (100 -150mm/s 0.4 noz 0.2 layer height) but the bed is at 105°C and the nitehawk SB runs hotter.

I have successfully printed PLA (7ish hr print) but I slowed the speed to around 150mm/s (0.4 noz 0.2 layer height)

I have successfully printed with no filament at higher speeds 200-250mm/s (9ish hr print) multiple times both bed and hotend running at proper temps (60°C & 230°C)

HOWEVER every time I try to run this print with filament, then it fails usually 1-2hrs in. VERY FRUSTRATING!!! Wasted filament too.

I have checked and cleaned all electrical connections to the Nitehawk, and PI, I am running a high quality USB cable. Nothing has made a difference. BUT I am starting to think it could be to do with the umbilical cable itself (i have a 350 size machine and the umb. cable is very long); could it be that with higher volumetric flow from higher print speeds that the higher power draw from the hotend (possibly extruder too) is too much for the cable and causing a low voltage on the NH??

I'm wondering if I could make a temporary power cable to test this, or perhaps make my own umbilical cable?

 

[Caspernicus]

I'm having this and it's driving me absolutely insane. It used to be that if I replaced the USB cable connecting the RPi to the little board that then goes to the umbilical, the issue would be solved until a few weeks or months later when the cable "burned up." Also switching ports on the RPi would extend time between failures, but I've run out of "good" ports on the RPi and don't want to replace it unless I have to, but that's probably what I'm going to try next.

 

[Caspernicus]

I don't want to speak too soon, but I may have found a solution. I got a USB-A to USB-C cable with 2 ferrites and extra shielding, apparently they're used in HAM radio applications, and I'm about 60 hours in with no errors yet. I didn't replace the RPi like I said I would, they're expensive
Here's the cable I used: https://www.amazon.com/dp/B0BJ182J9J

 

[StefanT]

Hmm, ferrites. Good idea. You can buy them also without cable to clip on existing cables. I think I will add one to the cable to the stealthchanger USB hub (the LDO version of the birds nest). Not that I have this problem but better safe than sorry.

 

[Robzilla]

I don't know if this will help anyone, but I have been doing further testing and I think it is related to the extruder; I have been using purely volumetric slicing, and I have been printing the same or very similar parts over and over. I can print these parts in PLA successfully up to 18mm3 (both 0.4 & 0.6 nozzles), but as soon as I go over 20mm3 the machine will fail with 'lost communication with MCU nhk' I am running E3D 60w heater with high flow nozzles, so this is disappointing. I have tried raising the current to the stepper with no improvement. I also flashed the nhk and leviathon board to the latest klipper with no improvement.

 

[ATVRACER]

I’m about ready to install 6 of these in my 6 tool changer I’m building I don’t want to fight 6 of these. After Reading some of these post works fine until empty till I add filament or work fine on pla it may be voltage drop to the cpu on the tool head with higher current load if the rung on the board not large enough the cpu will see a voltage drop may be going low enough to reboot the cpu with the E-motor stop and heater voltage returns to normal and all happy but with an error on the pi . If someone here can add a few wires at the cpu power in and measure it running at different temp and loads you may see the issue? Should be an easy fix