It's been some time now I'm getting BSOD and reboots (generally message is critical process died). It seemed to appear after I moved house or after I move my computer. I kinda "resolved it" by placing my PC horizontally. Not a good resolve since I have the problem once again.
It generally appears while playing games, not youtube or netflix for example.
When it reboots, sometimes it goes directly to BIOS configuration and the boot options don't seem to recognize the SSD my OS (Windows 10) is installed on (the only disk in the computer). That's how I started guessing that maybe the problem comes from the SSD. SSD is samsung evo 970, it's a NVMe M.2 SSD.
I tried applying a slight pressure on the SSD and it reproduced the issue. Applying slight pressures on other parts of the motherboard rarely reproduce it but sometime does. The thing is, I don't know if I should change the SSD or if the motherboard is the issue. Maybe it's only the NVMe M.2 connector but there is only one on the motherboard so I can't cross check. Maybe it's another part of the motherboard. Maybe it's a software problem since it happens without me having to apply pressure on any hardware part (even if it reproduces the issue). My guess was that when the fans of the motherboards are running fast, for example in games) it created vibrations creating the problem (similar to when I touch the SSD).
I can't get any minidump when it happens, BSOD stay stuck at 0% collecting information. Having a minidump would help me though...
So here are my two questions :
- How can I have minidumps, how can I get the information collection unstuck ?
- What do you think the problem can come from ?
Thanks ! FYI the computer has less than a year old so maybe I can have some pieces exchanged still... But still I'd like some advices before asking for it.
EDIT :
Here are my performance and Startup and Recovery setting for minidumps :
Virtual Memory is auto : 34816MB (I have 32Go of RAM) System and recovery : Dump File : %SystemRoot%\MEMORY.DMP Complete memory dump.
I installed whocrashed and I have the crash test dump. Is it possible that if it comes from the SSD, the minidum can simply not be written on it ?
EDIT 2 : I ran Prime95 today for 5 hours. With an open case. After 30mn, CPU was at 76°C (generally run around 50°C) and SSD 61°C, after 5 hours CPU was at 85°C and SSD 58°C.
I'll try with a closed case tomorrow.
I'll also try a GPU stress test with closed case. SSD is just next to the GPU fans so I guess maybe the hot air from the GPU heats the SSD.
EDIT 18/08/2020
Yesterday it ran ok for hours of gaming, no problem. At one moment for a totally unrelated thing, I plugged one cable on the motherboard, not far from the SSD. When I tried booting the computer, it went directly on bios menu, SSD was unrecognized. I touched the SSD to see if it was well seated. Afterwards, PC booted but after some minutes, it crashed with BSOD Kernel Data Inpage. Tried Prime95, it crashed instant. SSD was really hot, 72°C (Samsung advises max 70°C for this model).
I shut the PC down. I booted it this morning and now Prime95 make it crash after a few tens of minutes. SSD runs around 65 to 70°C temperature real quick. My conclusions for now is that I a parts seating issue (Some of the screws aren't there...) (that's why PC start crashing after I move it or if I touch the motherboard around the SSD) and a heat problem (which seem to only happen after I touch the motherboard but maybe both are related ? ).
Maybe it's just I damaged the card because of the lack of some screws...
I just reset windows to be sure there is no software issue.