You could try to flash the EEPROM with a configuration for 1x4 instead of 1x8 just to see if something changes.
1 Like
good idea
i’ve tried to use claud/chatgpt but it understand too little of the context and PCB design to be of much use
My guess is that their is either an impedance matching problem, or the pcb is susceptable to rf noise.
If there was a way to read the link training registers, it might give a clue.
The normal way to check these is with very expensive vector network analyser test equipment.
I was thinking the same. Maybe when i go back to university i can ask around to see if they have some of the equipment i would need to test that…
to connect the gpu do you connect it directly or with a riser cable? maybe its the riser, or something onto the pins like dust or hand grease/oils
or even the on board tiny psus circuits to give the pcie slot the 75w of power cause so much noise that a gpu crashes but something else doesnt
I’ve tested without the riser and its the same issue unfortunately
1 Like
I am currently in the processes of learning Altium as i want to transfer this project over to it. This might take a while. I am also moving back to college so i will be busy this week
Apologies for the radio silence. My mainboard was having issues regarding charging and dead ports. Framework agreed to replace it under warrenty
4 Likes
I’m still in the middle(just starting) transferring my project over to altium
2 Likes