Klipper Updates - why do I do this to myself!?

[Repman]

Congrats!

This is now the right moment to make a backup

[TitusADuxass]

6 minutes ago, Repman said:

Congrats!

This is now the right moment to make a backup

I've already backed up the config files.

[SamppaD]

On 10/15/2025 at 10:23 PM, NikosSprocket said:

It is still very, very nice and clean installation!

The 5Amp Meanwell should provide enough juice to the Pi for all the peripherals, I think. Soon I will be adding a second camera, nozzle camera, which will also connect to the Pi5. I hope will not have any power issues as already have cabinet and toolhead LEDs. 

Just have to actually install the 3DO camera to my toolhead design. On CAD it fits, but I have found out some manufacturers provided CAD drawings of their products that do not much the actual product!

 

Hi. By the way when you connect 5v psu to RPi don't use gpio pins to power the rpi. Header pin is good for 6A but header connector and a awg 22 wire  is not. Dont ask how i know :-).. recommending one of three options. 
Option 1: make a custom USB-C cable use awg 18-20 and power the rpi via USB-C connector.
Option 2 you can solder wires directly to rpi pcb underneath the header pins 2 and 4 (5V) and ground on pin 6 or directly to usb-c connector bottom. 

Option 3 use a power hat that mounts on top of RPI
RPI has a input power limit 3A that can be bypassed in some use case scenario when you have all usb ports in use.
In that case rpi4B will draw close to 4A max. So definitely 5A PSU and a proper power cable that can handle the amps.

[SamppaD]

5 hours ago, TitusADuxass said:

I've already backed up the config files.

You got it to work ,was it the Klipper related issue that required a full reinstall?

[Savate]

14 hours ago, SamppaD said:

Option 1: make a custom USB-C cable use awg 18-20 and power the rpi via USB-C connector.

Another good reason why usb-c is a better way to power the pi : the usb-c power input is protected, gpio pins are not.

[SamppaD]

1 hour ago, Savate said:

Another good reason why usb-c is a better way to power the pi : the usb-c power input is protected, gpio pins are not.

Yep. 
What i noticed after measuring power consumption under load was a power drop on gpio pins under load. 
This can lead to unstable working conditions with rpi.
Pi needs stable 5.1V thats not the case under heavy load,voltage will drop. 
Original RPI wall charger output stable 5.2V
So if encounter strange connection issues,degradation of printing quality,printer stopping mid print,check the RPI psu and its connection to RPI. 
On top of that if someone Still want to use gpio header to power the pi then use 2 pins for power but i don't recommend it. Mine RPI was connected per Voron manual and based on my encountered issues I don't recommend using GPIO pins for power.

[TitusADuxass]

17 hours ago, SamppaD said:

You got it to work ,was it the Klipper related issue that required a full reinstall?

I think so, it after I pressed the destruct/update button.

[SamppaD]

Well,I think I finally found the problem,Octopus pro 429 board is failing after 200 hours of printing. 
This morning I made a new USB-C cable with thicker AWG18 gauge wires to connect rpi via USB-C just to get rid of the gpio header voltage drop and moonraker connection issues. 
Checked the new cable before installation,hooked it up to bench power supply and tested with tester at 5V 3A,all good,providing 5V no issues. 
Checked all the harness before closing electronic bay,run basic check that everything is working as expected.
Started a print and on XYZ homing Z0-Z1 motors are turning but started humming/vibrating and no chamber lights..hmm..
Stopped the print, checked main PSU/RPI and all Z motors/chamber lights connections. 
Measured voltages on Z motors on the control board connectors and noticed that all voltages are different,printer ON and homed,Z motors holding. Z0: 2.4V ,Z1:4.8V Z2:3.8V,Z3:4.2V? 
When printer is just turned on,no homing done all Z motors are at 24V. 
If I'm correct i would expect printer homed and Z motors holding that voltage should be about the same on the same pin between drivers, I'm I wrong? 2209 tmc drivers works from 4.8V to 29V

Main psu is at 24V,RPI at 5.2V. Checked printer.cfg and all settings are the same for all Z motors. 
Homed printer again and humming disappear.

Signal pin for chamber lights is PB6.Chamber light at 5V but no signal on and no light. Measured chamber lights harness via multimeter ,voltage and ground at led connector showing 5V..
that what i was expecting but checking with multimeter between ground and signal pin on the led connector when printer is off i get short circuit? 
Disconnecting harness from control board and still get the short across PB6 pin and ground. Checked harness no short. This looks like control board issue. Does this PB6 has any connections to Z stepper drivers or main controller?
So at least PB6 is gone.
I purchased this board from authorized BTTseller in 2023 and first board was  DOA,got replacement that had defective PC5 probe port..China QC.  That would be this current board in use. Tried to find replacment 429 board but no luck..sold out

Update: 
Thermal camera image showing  hi temperature heat signature  on Z0 driver on idle..so board is done. Authorized BTT seller that sold this board is not available anymore.
I contacted BTT support,email respond was  no warranty  because its not purchased directly from them. 
So with BTT its always the same case they are selling BTT products using authorized sellers but when the product is purchased from the so called authorized seller and product has issues then its not under warranty as its not purchased from BTT directly!?  I dont understand BTT logic behind this.... 
Ordered new Octopus pro  732,550mhz  to replace this defective board but i must say that this is the last of BTT stuff that i purchase from them. There is no QC or warranty when you purchase BTT products. Unnfortunatly all printers are running on BTT products at this point of time,but not for long..no BTT products,no issues. 

Edited October 20, 2025 by SamppaD
Update

[SamppaD]

Final Thought, so as printer was printing for 200 hours before klipper update i was convinced that issue is related to Klipper or RPI.. 
After checking all that and finding only minor issues with rpi psu voltage drop that i fixed,I can see  that it was just coincident that printer broke down after update. 
It was the logic board that was responsible for connection issues and degradation of print quality..

[SamppaD]

Update on Voron resurrection. Found a BTT 429 spare brand new board in my parts collection bin. Flashed mcu and everything is rocking,no connection issues anymore. Printed for last 5 hours and looking good. Done a autopsy on failed board,found out that controller failed. De-soldered controller and short on PB6 is gone. Checked pcb,no visible damage to tracks or controller,it just decided that its time to go. New controller is about 12 USD+shipping from digitkey,so i will be fixing this board    I ordered a faster 723 octopus but its still  in transit..

Edited October 26, 2025 by SamppaD

[Repman]

15 hours ago, SamppaD said:

Checked pcb,no visible damage to tracks or controller

What you could do (now as the controller chip is gone) is to check and trace every solderpad of the controller to the peripheral devices by means of a continuity tester. Just to see if the circuit pathes are ok.

[SamppaD]

8 hours ago, Repman said:

What you could do (now as the controller chip is gone) is to check and trace every solderpad of the controller to the peripheral devices by means of a continuity tester. Just to see if the circuit pathes are ok.

You are right,i need to find the reason behind this failure. Visibly i couldn't find any burn marks,broken tracks,mechanical damage to the pins,ports,or pcb,so out of interest i will check all controller pins next week when i have more time.

By the way  some Octopus boards have manufacturing issue at least probe port was DOA on mine and online there is more cases about this issue,so it is possible that this board had more issues,then what i was aware of. That really low from BTT to selling defective boards as working... Warranty is 6 month for BTT boards,so most likely when someone notice this issue or board fail its over 6 month old by that time and no warranty.. I will be switching to mellow components at some point of time,i didnt like BTT way of dealing with manufacturing issues or customers regarding warranty.  

[polyneutron]

Back when I had the Creality CR10S Pro V2 and the mother board failed after years of printing, BTT introduced the Manta M8P V2 with can and 8 driver chips. I have always kept a spare. When I built the Voron 2.4 R2 I used the Duet 3 6HC with LT1 tool board and the scanning Z probe which was surface mount failures. I did print a few things with the voron but the probe failed and stopped the whole process. Then I went back to Manta and the spare is going to a going into an Ender 5 Plus that will be a core XY after a few more prints. I did order another Manta M8P V2 and CM4 with 16gb emmc and cooling heat sink fan to keep it cool. The Voron is going to a Jabberwocky hot end that is a lot heavier than the Orbiter V3 Revo that was almost working. The jabberwocky has a USB pass through that works. The Cartographer probe has to be updated with USB firm ware and the Voron will be running again.

Mellow Fly SHT 36 V3 is CAN FD capable so it would fix the issues with duet as no more surface mount connections that tend to fail.

Troubleshooting electronics used to be a joy, now it is out of my wheel house. I still have a 50 power scope and a needle tip soldering iron. All collecting dust. I used to remake lexus RX 400 needles when they failed by soldering into a new circuit board mini led's used in phone screens. I have even built circuits for pulse revers power. It's all over now.

 

[Repman]

1 hour ago, SamppaD said:

Visibly i couldn't find any burn marks,broken tracks,mechanical damage to the pins,ports,or pcb,so out of interest i will check all controller pins next week

I think this would be a good thing. Imagine you reinstall the new chip and it does not survive the weekend because of other issues...

Anyway I wish you good luck!

[polyneutron]

There is only one time I have had with a bad Manta M8P was a burned resister. That prevented getting it into boot mode. I could have replaced that resister and trace but what else may have been compromised just before it got to many electrons.

[polyneutron]

Sometimes there is microscopic metal between bgm traces. It depends largely on the age of the solder stripper in the chemical processing of the board. You need bright light or backlight and 40X scope or better. The defect looks like a shadow around/between the traces. The defect usually doesn't show up until maximum power with all of the connected devices and then you develop a short. 

This is what made me change to Duet. A Duet is at least 3 times the price but they have some design issues using surface mount power connections on there scanning Z probes.  That is what drove me back to a Manta M8P and keep spares in case the solder stripper was new or old. Stray current due to a failing heater can cause issues with chemical processes, they typically quarantine products involved but it depends on what the automated inspection even looks for before it is deemed good.