Jan 29, 2023
2
0
10
Hello and thanks for looking into this.

I have been having random BSODs, seemingly out of nowhere, for the past year or so.
Sometimes these happen during gaming, other times these happen over zoom calls and some times they happen due to having too many Chrome tabs open (I think). Other times they happen even when the system is not under any significant load. There are periods where several months pass without having any, then there are days with 3 or 4 crashes in a row.

After a BSOD, there's a 50% chance that my PC will not POST and the status light on my motherboard stays orange. A successful POST has the status light cycling from red to white.
To get around this, I typically either reseat my RAM, or change its slot. (more on this below). Sometimes this works right off the bat, other times I need to keep doing this for 3-4 times before the system starts POSTing normally.


BSODs (roughly from most encountered to least encountered):
  • Kernel security check failure
  • Kmode exception not handled
  • Memory management
  • IRQL not less or equal
  • Unexpected kernel mode trap
  • System service exception nvlddmkm.sys
  • PFN list corrupt
  • DPC watchdog violation


Things I've tried:
  • Updating BIOS
  • Updating drivers
  • Fresh windows installs
  • System file checker (sfc /scannow)
  • CHKDSK
  • Memtest. Passed but the hammer stage took a really long time. 5+ hours.
  • Tried running the system with RAM Stick 1 in slot A, then after getting another BSOD, moving it in slot B. Then tried running with RAM Stick 2 in slot A, then in slot B. Then tried running it with the sticks reversed, or in their original position. I don't have a clear conclusion for this experiment, but what I know for sure, is that if I run the system with both RAM sticks at the same time, the chance of getting a BSOD increase considerably. So that left me with sitting with half of my available RAM for the past year or so.
  • I am not overcloking anything


Specs:
  • CPU: AMD Ryzen 5 5600x
  • CPU Cooler: EKWB EK-AIO 240 D-RGB
  • Motherboard: Asus ROG Strix B550-i
  • Memory: G.Skill Trident Z Neo 32 GB
  • Storage: ADATA XPG SX8200 Pro 1 Tb
  • Power Supply: Corsair SF750 750 W 80+ Platinum
  • GPU: NVIDIA GeForce GTX 660

Yes, I am aware that my GPU is ancient. I have ordered a 3060 TI.



Please guide me towards what other tests I should be running to figure out what the issue is, before I go buying new RAM or sending my motherboard for RMA.

Thank you very much.
 
With all the different BSoD and since RAM passed memtest, i have a suspicion that it's your CPU which is acting up. 🤔

A while ago, there was another topic, where after extensive troubleshooting, the issue was solved by replacing a CPU,
link: https://forums.tomshardware.com/threads/system-bsods-and-crashes.3787658/

Might want to try those troubleshooting steps discussed there.


Thank you, I have seen that thread before posting. I'm not quite sure what troubleshooting steps to follow though.

I have run PCmark and 3Dmark and the system passed both tests without crashing.

What others tests should I run before starting to buy new components?
 
What others tests should I run before starting to buy new components?

CPU benches are CinebenchR23 and Prime95.

You could run both (one after another) and look how your CPU fares. Both the temps under load and final score.

Unigine Heaven or Superposition (if your GTX 660 supports Superposition) would tell if, perhaps, issue could be with your ancient GPU.

Memtest. Passed but the hammer stage took a really long time. 5+ hours.

One full pass (all 13 tests) or less?

Since one full pass is bare minimum, while 2 is better and 4 is considered acceptable.
Also, do note that for each 8GB of RAM you have, memtest86 takes ~1h. And if you have 16 GB stick, it takes ~2.5h per stick. So, for full 32GB (2x 16GB), one full pass takes ~5 hours or so. 2 passes would be ~10 hours and 4 full passes would be ~20 hours.