[SOLVED] System Instability, Random Reboots

[elderdrake]

Just rebuilt a new system after an old motherboard went bad and ever since I have been experiencing random restarts almost timed exact to 22-24 minutes on different games. I did a lot of research into this and after resetting my CMOS and changing some power settings in Windows (dealing with fast power on), the system seemed stable for about 24 hours and now I am back to being able to do normal task without a restart but any gaming causes the restart. I have tried a lot to solve this:

-Event Viewer shows it as a generic error 'Event ID 41' with nothing leading up to it that I can see
-checked RAM with memtest86
-reseated all power cords
-changed the GPU port
-ran Prime95 for 30 minutes on the CPU (to simulate a hard load), no crash
-made sure all drivers are up to date including Windows
-cleanly removed my NVIDIA drivers with DDU (in safe mode with no internet access) and cleanly installed the newest driver (same environment)
-ground tested the PSU to see if the fan works (it does) and checked the pins with a multimeter with all pins showing stable and correct voltages
-removed 3rd party apps like iCUE and disabled EVGA Precision

I am stumped and considering this is basically a new unit I do not know what to do. Everything is new but for the RAM, hard drives, PSU, and GPU (the later two were RMA'd from EVGA with the last 2 weeks). I am attaching* links to my recent Speccy report and I ran HWINFO from starting a game till crash with full monitoring and providing a link for that as well.

*I do not know the proper procedures for the forums here on links and to upload a file nor what is an acceptable link. I will provide basic links for a .txt file for Speccy and the .csv file for the monitoring I did on HwInfo up until the system crashed and restarted.

Speccy:
https://www.dropbox.com/s/88flvscgxfps32p/DESKTOP-1.txt?dl=0

HwInfo:
https://www.dropbox.com/s/b3not82tl0f88i6/Hwinfo_log.CSV?dl=0

Update: Reset the CMOS again and started a different game up and almost to the exact minute it restarts. I can run anything that does not put a decent load on the PC but anything that does causes it to restart just around the 30 minute mark. Check the Event Log, nothing again pointing to anything and I had HwInfo logging two events now back to back but I don't really know what may be out of line there or not.

 

[elderdrake]

Update 2: System is running a game on round 3 now going on 50 minutes, not sure if it is stable yet but I will keep responding and updating this post for at least a few days. Making sure to update in case others have issues akin to mine with a similar setup. I am not sure if Speccy list it but I have an EVGA 850W G3 PSU.

To rule out power issues I switched wall outlets (wifes idea), and made sure I was not an idiot and had say a CPU power plug plugged in wrong. So I went over all the cords and made sure the fittings were tight and not stressed and while doing so I thought of my EVGA 1080ti FTW 3. My GPU was being powered off one cable with two 8 pin adapters. Thinking my card might be starved somehow on that one cord, or the cord was having issues with the load, I kept it plugged into one 8 pin on the GPU and dug out another VGA power cord and connected it to VGA 2 with the though being on splitting the load between plugs and wires. I am not sure if this is even a realistic repair but at this point I am trying everything. Set EVGA Precision up again with an idle fan speed of 50% and a little more aggressive as it goes up (had this set beforehand but I disabled it when removing apps to rule out conflicts there).

Bottom line is like I said not sure if the CMOS reset, power outlet change, or additional power cord running to the GPU did it but so far it is still running as I type this at 1 hour and Red Dead 2 is not exactly a forgiving game at Ultra settings and 1440p. What really trips me out is I stressed the GPU with Furmark until the onboard sensors ramped the fans up to 100% and the GPU leds came on red showing high temp, system did not crash. I also stressed the CPU with Prime for 30 minutes and temps were stable at 60C being the highest so I am wondering why stress testing them hard did not cause a reset but gaming for 30 minutes did.

 

[Vic 40]

Guess the next is a typo?

I kept it plugged into one 4 pin on the GPU

Having the second cable might really have stabilized the gpu. Had that problem before where one would just not cut it.

Since having Hwinfo if in the future you want to give system specs do the next,is not such a wall of text,

download hwinfo,
install and open it=click run,
close the top window which is the system summary,
in the main window at the left top click "save report",
at the bottom of the next window check "sSummary for Clipboard",
after that you'll see what's in the pc,
copy by clicking "copy to clipboard" and rightclick+paste in your next respons

well without that download since already having it.

 

[elderdrake]

Guess the next is a typo?



Having the second cable might really have stabilized the gpu. Had that problem before where one would just not cut it.

Since having Hwinfo if in the future you want to give system specs do the next,is not such a wall of text,


well without that download since already having it.

lol, indeed and I fixed that post quick. Still running stable and I will mark this as solved in a few days if so. Those general hardware faults can be hard to isolate and trouble shoot sometimes when everything seems to be fine. In this case the PSU is putting out proper voltage tested on a multimeter, the BIOS readings, and application logging like HWInfo but it may just have turned out to be one VGA cable in the end. Seems you literally have to isolate every part and cord of the PC to lock it really down.

Next I am getting a Line Interactive UPS from Amazon rated as high as I can to rule out power from the outlets. Seems we have had some fluctuations where I live, enough to cause a modem reboot once in the last couple weeks. Lots of variables and no single silver bullet when it comes to this, just a lot of patience and diligence mixed with looking at threads I have seen like mine to see what worked for others and how they worked through it.

 

[elderdrake]

Update:

PC is back to restarting all over again, was gaming for a bit and then blam it just reset. Anyone have any ideas based on what I've reported here? I am completely at a loss.

 

[Vic 40]

Something non hardware,but did you after the hardware swap reinstall windows?

 

[elderdrake]

Something non hardware,but did you after the hardware swap reinstall windows?

I did a Reset Windows and deleted everything on the drive via Windows itself but i think I am going to reinstall from CD and see if that makes any difference. I have a feeling this is hardware related but everything is checking out with every test I can think to run.

 

[Vic 40]

Would do that fresh install,see how it goes.

 

[elderdrake]

Further Update:

PC got stuck in a reboot cycle where I could not get to the BIOS splash screen, if I powered the PC off and let it sit for a moment I could boot into windows for a minute or two and then it would crash into a non stop reboot cycle (never hitting even as far as the BIOS screen) until I hit the physical PSU switch to kill power.

Trouble shoot steps I took were to reseat the RAM, removed the GPU and booted off the onboard video (to rule out the GPU), removed all case connections (USB ports, power, restart, leds, etc) to rule out a short from the case wiring. Still no joy yet every crash the CPU temps were below 30C.

My thoughts at this point is that it is either the motherboard or PSU hitting a thermal limit and cutting off. I base this assumption on the fact that when it was reboot looping last night, powering it off for a minute and back on let me get to Windows for 1-2 minutes. I also ruled out Windows as I let it sit in the BIOS menu and it did the same thing (crash then reboot loop cycle). Voltage on the CPU seems steady in reports from HWInfo recordings over time and in snapshots looking at the what the BIOS reported as well as other monitoring tools as well as a multimeter.

Any thoughts gents? If it is a thermal issue then I will not know until I say game for a bit and even then I do not know what is causing the instability or how to lock it down exactly. I could pull my wifes PSU and see what happens but it would be a pita as it is not a modular system and cable management was a nightmare on hers. Could something on the motherboard be slowly heating up and causing this?

Update mid-post: PC crashed/restarted just typing this thread so I hooked my wife's old TX 750W up and I am nearing 1 hour on with no sudden restart and boot loop. Worried as I have only 1 8 pin connected for the i7 9700k but I read that this is not a big issue. May salvage hers and RMA my PSU to EVGA but the voltage on the pins is fine, maybe it is just failing to keep steady voltage or is faulting someplace.

 

[elderdrake]

Final Update (unless system crashes again, knock on wood):

Don't like running an old PSU but it looks like (knock on wood again) the EVGA PSU was the culprit. I RMA'd it but again it is odd that the voltage read fine on every pin, but I know there is a lot more under the hood and it seems under load it pushed it past something that faulted out. My wifes old Corsair TX 750 Bronze is running it fine so far (final wood knock).

Final question for the community though. I had removed every case power pin from my mobo as part of troubleshooting to rule out a short, I think it is safe to reconnect it as I can't see a faulty case power/reset/led causing the PSU to do all this but worth asking and throwing it out there. The Z390 Gigabyte Master has a convenient power and reset CMOS button on the back I/O panel so right now it is just inconvenient.