PC is restartingly seemingly at random under low loads. Tried everything (feels like).

[Xyrax]

TL;DR no BSOD. no warning. no event viewer logs. no dumps. just a power cycle. thought it was psu related but it isnt. Memtest came back fine. never reboots while gaming. Temps are fine on GPU and CPU.

Specs are:

• NZXT H510 Case
• Windows 10 Pro
• i7 9700
• Corsair RM850w Power Supply
• MSI MPG Z390 Gaming Edge AC Motherboard
• 16gb Dell OEM Ram (Swapping this out today)
• Deep Cool L240v2 CPU Cooler
• MSI 2080 Super Blower
• This Vertical GPU Mount

Some history: Started with an Alienware Aurora R8. GF and I Built a gaming pc for her and I wanted a nice case and AIO for myself. So I moved. Alienware MB has a proprietary CPU Cooler mount + Power Button, so I buy a new MB.

Install goes fine, cables were a bit short on the dell PSU (also an 850w gold rated) so I get cable extensions. It boots and works. That week it reboots randomly once. has been doing this 2-3 times a week since then.

Whenever it reboots it: Does so with no warning, no blue screen of death, no error code beyond "windows did not shut down correctly" in event viewer, no minidump in c:\windows, no memory.dmp in windows either.

Monday this week it reboots twice on me in one day. Do some reading. Seems like PSU. So I buy a new one and replace the PSU. Now using all Corsair cables no more extensions. Didn't do it yesterday but just did it to me again today.

Ran windows memory test on ram. Came back with no issues. installing new ram today when it gets delivered. (It was on a good deal on amazon)
It always reboots when I am doing "low load" stuff like listening to music/browsing reddit. Never has it rebooted on me while I am gaming. I don't think it's temp related either because I monitor them and they always stay well within acceptable ranges.

I am at my wits end here. Not sure what else can be wrong besides my motherboard. I have seen others claim that the CPU going into low voltage or something causes the MB to reboot itself? but I put my PC into high performance mode and I thought I went into the power settings and set it to minimum 100% so not sure what's going on there.

Please help me, I have nothing left to try other than totally wiping my drive and then if that as well fails, returning my motherboard and getting something else.

EDIT: The only lead I am chasing right now is this. I don't understand Vdroop and cpu multiplier so I didn't touch those but I did change voltage to 1.25v (that was just under the stress test number) and turned off turbo boost.

 

[umeng2002]

When I was under-volting my Ryzen 2700x, I ran all the stress tests in the world to get it "stable," and it passed all those tests; but I would get freezes and crashes every now and then when idling on some website... My voltage was too low for idle stability when the CPU would throttle and under-volt even more to save power. Under full/ heavy load, the CPU increases the voltage to normal levels and it was stable... so I under-volted too much. I reduced the offset reduction to decrease my under-volt to fix the issue.

Make sure you have enough voltage under idle/ low loads.

 

[joemama]

So manually setting the CPU voltage didn't fix the problem?
It could also be a windows problem, I have seen some problems going on with the recent windows updates

 

[DeathtoGnomes]

Did you change any BIOS settings recently? You could reset the BIOS, after that go back in and load default settings in the BIOS.

 

[nguyen]

First I would take the entire PC apart and put it back together (including CPU), use only the cables that come with your Corsair PSU and nothing else.
Second remove your vertical GPU mount and plug your GPU into the motherboard slot
Third use HWmonitor or HWinfo64 to check if any part of your computer is overheating (CPU, GPU, VRM, Chipset, SSD, etc..), go to BIOS and load optimized default.

From my opinion your vertical GPU mount is most likely the culprit, so maybe check that first if you like.

 

[delshay]

Check all sockets has no foreign objects lodged. Use a torch & shine a light down all sockets. If nothing is found, do not clean.

Yesterday I found two human hair lodged between the CPU pins. This is the second time I have found hair here, so you may want to check here too. Use a torch here too. Again if nothing is found, do not clean.

I used my front mountain bike light, as my hand torch light is broken.

 

[GeeBee]

The Vertical GPU mount thing looks very suspicious indeed.

I'd get rid of that if I were you.

 

[milewski1015]

First I would take the entire PC apart and put it back together (including CPU), use only the cables that come with your Corsair PSU and nothing else.
Second remove your vertical GPU mount and plug your GPU into the motherboard slot
Third use HWmonitor or HWinfo64 to check if any part of your computer is overheating (CPU, GPU, VRM, Chipset, SSD, etc..), go to BIOS and load optimized default.

From my opinion your vertical GPU mount is most likely the culprit, so maybe check that first if you like.

Adding on to this, I would breadboard it first before rebuilding in the case. Disassemble the PC, and then just build the bare minimum outside the case: one stick of RAM, use the iGPU instead of the 2080 Super, etc.

 

[Deleted member 193596]

do you have a small hdd/ssd left over?
after checking the hardware,cable connections and a cmos reset... install windows from scratch on another drive.

why? i had the same issue as you after kaspersky messed with my installation on a prebuild that was filled with bloatware. the PC just shuts off completely while watching a youtube video or sitting on the desktop. and then it reboots. (no dumps aswell)

 

[milewski1015]

do you have a small hdd/ssd left over?
after checking the hardware,cable connections and a cmos reset... install windows from scratch on another drive.

why? i had the same issue as you after kaspersky messed with my installation on a prebuild that was filled with bloatware. the PC just shuts off completely while watching a youtube video or sitting on the desktop. and then it reboots. (no dumps aswell)

Another good suggestion. If OP doesn't have another drive to use, they could try running sfc /scannow in an elevated command prompt to check for corrupted OS files