My first attempt at board repair - Asus Prime X670-P

[tiagooliveira95]

Hey everyone

I decided to dive headfirst into board repair. I'm a computer engineering student, and I'm attempting to fix my first board.

I have made some progress, and I feel super excited. Still haven't fixed the board tho, but I decided to make a post here to share my progress.

This is an X670-P, seller stated bent pins and no power on, the socket has 2 bent pins, the pins are slightly bent upwards (VSS and a Channel A pin), so it's an easy fix, don't even need to replace the socket, tested all power rails on the CPU and found no shorts, the bent pins are also not in short, power rails of both southbrides 0.95v and 1.8v also show no shorts.

Given this information, I tried to turn on the board (no cpu, no ram) with a cpu fan attached, but bridging the PWR-SW did nothing, completely dead, so I flashed a new bios, tried again, still nothing.

My next step was to check if the PWR_SW signal was being received by the Nuvoton IC, which it was, so I probed the PSON pin of the Nuvoton, and only saw 2.4v ,which I believe is incorrect. This signal did not change when I bridged the PWR_SW pins. To test if the rest of the circuit was working, I shorted the gate of the MOSFET that the Nuvoton's PSON signal controls, and the PSU and CPU fan turned on.

I hypothesize that I have a dead Nuvoton chip, or some of its power rails are shorted, maybe the Nuvoton chip is not receiving all the power rails is the reason for the 2.4v in the PSON pin.

This is all I have for today, it's 3AM here, I must go to sleep, but tomorrow I will check all power rails of the Nuvoton chip to see if I find something.

I'm really happy with my progress so far, and any advice or validation of my hypothesis would be appreciated.

[tiagooliveira95]

Yeah, pretty sure it's the Nuvoton that is dead, all rails are there and stable, but it refuses to work, new one should arrive in a few weeks, after replacing it i will give an update to hopefully help someone in similar situation

[DynaxSC]

Note that not all boards start without CPU inserted in the socket. Some do monitor CPU presence, and do not start board if no CPU is in the socket. This is usually done by the Super I/O. Grounding the relevant SIO pin can allow board powering w/o CPU. This is for sure so on Intel boards, but AMD might be the same.

[tiagooliveira95]

Note that not all boards start without CPU inserted in the socket. Some do monitor CPU presence, and do not start board if no CPU is in the socket. This is usually done by the Super I/O. Grounding the relevant SIO pin can allow board powering w/o CPU. This is for sure so on Intel boards, but AMD might be the same.

Thanks for letting me know!

But even if that was the case,I still find it odd that PS_ON sat around 2.4v, this pin goes straight to the mosfet with no resistors or caps on the line, a 10k pull-down resistor barely changed the voltage, so i don't think the pin was floating, it looked to me it was being weakly driven.

I don't have a spare AM5 CPU on hand, I removed the SIO already, will wait for the replacement, and once installed i will check the PS_ON voltage again, if it sits at 3.3v and pulls low correctly, that should confirm that the original one was faulty, however I kinda doubt because the board has clear signs that someone tried to install a CPU on it, and listing stated no power up.

Since the damage to the socket is so light, I believe the SIO died for some other reason, and the socket was damaged afterward. The seller did not include a CPU cover, which tells me that maybe they were not taking proper care of the board, perhaps because they already believed it was dead.


Photo show the cpu socket, afected pins are BD32, BC31 both go to A1 and A2 ram slots, and BD31 GND

Spare CPU and SIO should arrive within 2 weeks. Still, before turning it on, I will make sure to check, and double-check if there are no dead shorts on all the adjacent pins