Question New PC restarts without warning and for no reason

[MRMaroon123123123]

Hey everyone,
I recently built a new pc featuring:

• RTX 3070 Inno3D Ichill 3
• Ryzen 5 5600x
• 2x 8GB corsair vengeance ram
• Asus TUF x570 plus wifi Motherboard
• 2x Samsung 870 Evo SSD (one of which was salvaged from my old build)
• Be quiet pure loop watercooler
• Corsair RM750 750watt PSU

Now, this new build sherds everything I throw at it benchmarks included. However, it has an issue that keeps popping up which I can't make heads or tails of. That being that it sometimes cuts out to black screen and immediately restarts without warning whilst doing things that do not justify such a response from a system. The most reliable way I have found to reproduce this issue is kind of weird: Having world of warcraft open and trying to watch something on prime video using google chrome. This triggers the restart most reliably.

I have already tried somethings to fix this issue. these include:

• Updating all system drivers
• running an sfc /scannow (which did find a corrupted file but that did not solve the issue)
• rewiring a bunch of stuff in my system to new ports.
• Seeing if I could find anything in the windows event viewer. This lead to me finding Kernal power ID 41 a rather unhelpful report merely saying that the system did not shut down right.

Now, whenever I try to find a solution to this problem most advice either ends up being Temps or PSU. I know for sure it's not my temps because I ran multiple benchmarks and stress tests to see if I could trigger the issue that way during these tests my temps rarely exceeded 75 degrees Celcius. I am not sure how to test my PSU but I feel that the issue would have occurred during the stress tests if it was to blame. and for what its worth, the RGB in my case doesn't turn off when the issue occurs which I feel would be a normal thing to happen in event of a PSU failure.

At this point, I am not sure what on earth could be causing this issue. My instincts say faulty RAM but I stress-tested that as well during the stability test I performed with AIDA64. Maybe a faulty Motherboard, or maybe not even hardware related. I really want to see if I can figure this out before I go to a repair shop or have to send parts back. Any insight you fine folks would have is greatly appreciated

 

[tecmo34]

Go into the Nvidia Control panel, under Manage 3D settings and change the Power Management mode to "prefer maximum performance" .

There are know issues with the RTX 3000 series cards with crashing due to voltage issues at low or idle performance. This has been a recommended fix for this type of issue.

 

[MRMaroon123123123]

Go into the Nvidia Control panel, under Manage 3D settings and change the Power Management mode to "prefer maximum performance" .

There are know issues with the RTX 3000 series cards with crashing due to voltage issues at low or idle performance. This has been a recommended fix for this type of issue.

Just wanted to thank you. This was the fix! looking back it all makes sense, why It did not happen during stress tests and only in low-intensity situations. I really can't express how happy I am to have fixed this. Thank you!

Unfortunate update: It just happened again after setting the new power preference. Maybe a one off? hopefully. Did not restart my system after editing the preference. maybe that was it. edit not a one off.

 

[Deleted member 14196]

Maybe faulty psu

 

[tecmo34]

You are running two PCIe cables to the GPU and not daisy chaining one?

 

[sizzling]

My guess is the psu, the RM range seem to have a lot of reports of issues with the 3000 series gpu’s. The RM has a lower than average transient response performance for what is supposed to be a high end psu.

 

[MRMaroon123123123]

My guess is the psu, the RM range seem to have a lot of reports of issues with the 3000 series gpu’s. The RM has a lower than average transient response performance for what is supposed to be a high end psu.

Should I just return this one for another one or is there one in particular that you reconmend

 

[MRMaroon123123123]

You are running two PCIe cables to the GPU and not daisy chaining one?

If you mean by daisy-chaining that I used one cable to fill all 16 slots on the Card then yes. The cables in the box had 16 power pins on them so I assumed it would be enough to power it.
But from your question, I am assuming you want me to take a second cable and connect it to my card. Thus making it so that each cable only uses 8 pins instead of 1 doing 16.

side question if the card was not getting enough power wouldn't I have seen it shut off during the stress tests as well? Not to be rude or anything I am wildly out of my league here just genuinely curious.

upon further and closer inspection of the PSU, I read it: 6+2 PCIe. I was using a single slot 8 slot to power 16. attached a second cable to another 6+2 slot and ran that into the GPU as well. I Will update if the issue reoccurs after this rewiring.

UPDATE: Yep did it again.