MCU Timouts

General Discussion
1

Basic Information:

Printer Model: DIY cartisian
MCU / Printerboard: SKR 1.4 turbo
Host / SBC : Pie Zero 2W
klippy.log

klippy.zip (2.0 MB)

moonraker.zip (5.4 KB)

Fill out above information and in all cases attach your klippy.log file (use zip to compress it, if too big). Pasting your printer.cfg is not needed
Be sure to check our “Knowledge Base” Category first. Most relevant items, e.g. error messages, are covered there

Describe your issue:

I am getting constant mcu timouts, I changed my Pie zero 2, I changed my cable 2 times, I am using a very short jackted usb cable for my serial connection and using the jacket as ground. I am still getting timouts mid print. my pie is not overheating and it is not stressed at all. I dont get what is going on from these logs. can someone read these and tell me what is wrong? steer me to the right direction, so I can hopefully fix the issue.

…

2

Hello @sinole and welcome to the forum!

With what? With another Pi Zero?
The Pi Zero is way to weak for running Klipper properly.

5

hello and thank you for reading my post. No, I meant Zero 2 W, just just used that as short name sorry. and I am having a bit of embacring moment on the new forum, not being able to edit my reply properly. :smile:

6

Only 1 end of the shield should be grounded and it should NEVER carry current.

7

ground on the MCU goes to ground on the pie,so i dont know how only one end can be grounded. ground of the whole system is connected anyway up to the ground of your psu power plug. and if you have anotrher device connected to the same circut of the wall, they are connected to the same ground. I remmber i slightly overclocked my pie zero 2 and even overclocked the ram. even though pie is not even being challanged, and it runs very cool. overclock always brings unstability. fingures cross that was the issue. I removed it and ran a very long print to check,

8

Normally,
GND, or common ground, is a reference voltage in DC.
Ground in the power plug is the earth.
And they are normally not connected directly. (There can be a several-MΩ resistor or a Y capacitor).
Where the cable shield is connected on one end to GND to block EMI.

image
image640×853 80 KB

-Timofey

9

so is it wrong to use the shield as ground wire? I ripped open a bunch of usb cables, a lot of them do the same, they dont have ground wire they use the jacket. and even when I opened a good usb 3 cable it was running the ground wire on the outside of shield without insolation.

10

well removing overclock didnt help.

11

I use Power to power my raspberry pie. Even though i get “vcgencmd get_throttled, throttled=0x0”, I heared buck convertors are too noisy and it could be the reason I get timouts in a long print. any thought?

12

Just my 2 cents; they can be wrong.

From your log, you use P0.3/P0.2 so direct UART.
The USB cable is just a physical medium and does nothing.
I doubt it is EMI, as EMI will result in invalid bytes or retransmits.
Lost communication happens when the board does not respond in 5 seconds.

The protocol should be able to easily handle EMI and retransmit what is necessary.

So, I guess, the real question is why your board is losing communication.
Common GND between RPI and SBC should ensure normal UART performance.

How you recover from the Lost communication?
I guess, if the board is hung (so it cannot respond to UART), you cannot recover it without a power cycle?
If the board restarted (from a power spike, for example), klippy will receive wrong data over UART, and should output that the board is probably restarted.

You can duplicate the 3.3V or 5v line between SKR and RPI, in the worst case, the step-down converter that powers RPI will power the SKR, I guess.
But it is more about duct tapes.

(X) USB
( ) Serial (on UART0 P0.3/P0.2)
( ) Serial (on UART3 P4.29/P4.28)

image
image1920×1356 814 KB

-Timofey

Maybe wires are loose, maybe they have to be soldered, or another wire for GND can add redundancy (it is not so cool/correct, as it can create ground loops, but… it can allow us to keep the connection).

Ah, and if you use a step-down from the same PSU (which powers the SKR).
There is already a common ground between boards.

I guess, I have to quote Mike and asks for photos/connection/schemes.
As you can have ground issues only if you have several PSUs.
If it is the only one, UART should work without a separate GND Bridge wire.

13

I need to mention, I have no problem in short prints. the communication issue happens after 9-11 hours into a print. that is why I am leaning towards power supply reliability. I will post a picture of my electronic box when I get home.

14

Tried an external power supply cana kit pau. Didnt help. This is the config tried to iso the gray wire aka uart connection away from rats nest but i dont know if it still picks up emi from all of those wires. I wana try ferrites next on both ends

IMG_2112
IMG_21121920×2560 1.08 MB