Help diagnosing system instability, kernel power critical message

[dr.stevil]

This has been happening on and off since first building this rig in February of this year and I’m at my wits end. I want to find out what’s causing it so I can have a reliable machine again. I should note that it only seems to happen when the machine is idle and typically in the middle of the night, so never when it’s been under any heavy load.
specs are as follows

AMD Ryzen 9 3900x
x570 Aorus Elite mobo
64gb g.skill ripjaws 3600 ram (4x16) running at the default 2133 speeds
Corsair RM850x PSU (using a UPS)
Gigabyte Vega 64 GPU
Samsung EVO+ 970 m.2

the past few days I’ve went into my Office and found the machine unresponsive, black screen and GPU fan humming away @ 100%. I used bluescreen view and this is what I’m seeing. I’m currently running the latest version of memtestx86 on each stick to rule out the ram. The GPU was running in my old machine with no real issues so I don’t think that is necessarily the culprit. If I was a betting man, I assume it’s either the mobo or cpu that’s faulty but I don’t know how to test for it.
any suggestions or help is appreciated. I keep all drivers up to date and used several bios’s with no real change In behavior. Sometimes the machine will go weeks without any problem.

another quick note, if I try to overclock the ram, it will tend to cause a crash much quicker than if I was at the default stock speeds but don’t usually get a crash dump

 

[Zepher]

Swap out the PSU if you have a spare one and see it still does it, that would probably be the easiest thing to do.

 

[DooKey]

Might be the MB or CPU doesn't like four sticks of RAM. I'd drop down to two sticks and see what happens.

 

[dr.stevil]

Sounds good!
I did have a spare PSU until yesterday
Gonna do more memory tests and go from there.
Thanks ya’ll

 

[dr.stevil]

So believe it or not, I think it was a bios issue. Went back to and older revision from earlier this year and so far so BSODs or random restarts. Keeping my fingers crossed

 

[dr.stevil]

Ok so it wasn’t a bios issue (or not completely). It turns out that I have a stick of ram that keeps coming loose... how that’s happening, I’m not completely sure, especially considering it happens sometimes without even touching the machine. Possibly thermal expansion?

either way, has anyone else dealt with this before? It’s super odd

I have my heat sync fan directly above the sticks (like a 2mm gap) so I’m thinking about wedging a piece of cardboard or something in there to keep pressure on the top of the sticks to fix the issue for now. I really don’t want to disassemble the machine to RMA the board.
Wtf

 

[Epos7]

We have very similar systems, and I have been experiencing the same thing for some time. I made a thread about it thinking it was a PSU issue.

More recently, I found this on Reddit, and have tried setting my VDDG and VDDP voltages manually in BIOS. It's been a few days now, and no issues so far. I'll have to go for a couple months without seeing the error before I consider this the solution, however.

 

[Epos7]

I'll also note that while I don't think it's related, I've been having an issue with my computer constantly waking from sleep for no reason. So far I think I've been able to resolve that by disabling wake from sleep for my network adapter in device manager. Just throwing it out there in case the two issues end up being related.