[SOLVED] PC freezes and needs to be force shutdown (Games/Benchmarks)

[isuckatbuildingpcs]

Hi,

so I spent the last 2 days and nights trying to fix my custom built pc and I don't know what to test anymore. The main problem is, that the PC freezes while gaming and one time after a benchmarktest. When the PC freezes, the capslock light won't light up when I press the key, the PC won't unfreeze himself, the restart button on the PC won't work, only way is to hold the powerbutton.
These crashes start to happen after some time (1h, sometimes 2h). And sometimes they happened in a row: PC freeze, reboot, restart game, freeze withing 10-15 mins

Setup (everything is stock, no OC):
CPU: i9900k @ 3.6Ghz
CPU-Cooler: be quiet! Dark Rock 4
Thermal Paste: Arctic MX-4
Mainboard: MSI MPG Z390 GAMING EDGE AC
RAM: 64GB Corsair Vengance LPX @3200Mhz (XMP is enabled, otherwise it runs at 2133Mhz)
GPU: MSI V371-026R GeForce RTX 2080Ti Gaming X Trio
PSU: Corsair CP-9020133-EU TX-M 850W

So I tried:

• Reinstalling Windows and all drivers (Drivers from the MSI website for the MOBO + newest BIOS, Graphics only via Geforce Experience.)
• Reapplying thermal paste (was fine the three times when I changed it, good spread)
• Chkdsk: No errors
• Userbenchmark: No underperforming parts
• Tested RAM with mem86 and Windows own tool: No errors
• Switching powersocket in my room
• Logging all temperatures with OpenhardwareMonitor: Nothing too fancy while the gaming freezes happened, CPU temp was at max 80°, unlike during prime95 with 100°

Stresstests:

• Stresstesting with Valley-Benchmark: NO crashes (1h runtime), tmp for GPU maxxes at about 70°C
• Stresstesting with prime95: tmp goes up to 100°C, I stop the test after 1min at this tmp. Sometimes the tmp drops down to 80ish again, but sometimes it won't. After my last test the PC froze one second after I closed prime95. Coincidence?

Games it crashed with:
MW Warzone
Totally Reliable Delivery Driver (many physics involved?)
Overwatch (only after it froze 3 times beforehand, otherwise 3h of Overwatch have been fine without any freeze)

Game it didn't freeze with:
Borderlands 3, played for 3h+, no freeze happened

My biggest concern is the CPU temperature. As I stated above, the CPU temp was mostly fine during gaming, sometimes it shot up to 90+ only to instantly drop back to whatever it was at before (missread by the sensor?). Only prime95 let it stay at high temps for a long time, but I read that prime95 might be overdoing it sometimes.
The CPU Cooler is mounted tightly onto the CPU, the thermal paste was spread evently everytime I checked when reapplying it. One thing I am wondering: When the CPU temp was at 100°, I touched the radiator, the temperature was maybe slightly above room temperature. Is this normal?
My other thought is the mainboard: I googled my setup and had many cases pop up with the z390 (and most of them also had the Vengence Ram). I will send it back if noone has another clue what it might be, but in that case, what MOBO should I get for this setup?


Any help is greatly appreciated!

 

[kingbowcat]

Is your CPU fan cooler defective in some way? that would explain the high temperatures and why the radiator was room temperature and not as hot as the readings. Also some games worked and some not because some games are more CPU intensive?

Have you got another CPU cooler you could try?

 

[isuckatbuildingpcs]

In idle the temperature is quite low (30°), but after running the PC for 20mins just sitting in teamspeak it went up to 40°.
I don't have another cooler, but I can try this cooler with a mates PC and check if it also cools very badly.
Thanks for the hint!

 

[isuckatbuildingpcs]

So now it crashed with a CPU temperature at around 60°... I don't think that it's the main reason why it is crashing. Weirdly the Windows Event Log doesn't show anything. The latest events logged was 10 minutes before the crash.

 

[isuckatbuildingpcs]

So it got fixed with one of those three things:

1. Downgraded the BIOS version from A.70 -> A.60
2. Moved my USB Headset to another port (sometimes I got a critical error message in the event log)
3. Reset CMOS