I've been getting random Windows 11 BSODs for about a month, often mentioning ntoskrnl.exe, IRQL_NOT_LESS_OR_EQUAL, and similar errors. The crashes used to happen even while simply browsing. My system has a Ryzen 5 3600, 16GB DDR4-3000 RAM, an ASRock B450 Pro4 motherboard, a Super Flower Gold 550W PSU, an RTX 3070, a 1TB main SSD, a 240GB secondary SSD, and a 2TB hard drive.
I've reinstalled Windows and updated the BIOS and drivers without fixing the problem. Removing one RAM stick made the system last longer, but it eventually crashed; switching sticks and reinstalling both produced similar results. After reseating the RAM, the system seems more stable for now.
Windows Memory Diagnostic reported no errors, but I'm wondering whether I should run MemTest86 overnight. There is also another issue: while gaming, my secondary NVMe SSD disappeared and then reappeared. Since it is installed in an M.2 slot, it could not have physically come loose. Could the disappearing drive and the BSODs point to a motherboard, PCIe, power, or memory stability problem?
4 Answers
DISM /Online /Cleanup-Image /RestoreHealth can repair corrupted Windows component files, but a successful result only means DISM did not find an unrepaired problem; it does not rule out faulty RAM, storage, power, or motherboard hardware. Since Windows and driver reinstallations did not solve this, hardware-focused testing is the better next step.
Power is another possibility. An RTX 3070 can have significant transient power spikes, so a 550W supply may be marginal depending on its age and exact model. Make sure the GPU uses two separate PCIe power cables rather than a single daisy-chained cable. If possible, testing with a known-good, higher-capacity PSU would help distinguish a power problem from a memory or motherboard issue.
The NVMe drive disappearing during a game makes this more than just a RAM question. Check the SSD’s firmware and temperatures, reseat the M.2 drive, and inspect its health information. A faulty drive, overheating controller, motherboard slot, or PCIe stability problem could cause storage dropouts and system crashes. Updating the motherboard chipset drivers may also help.
MemTest86 is worth running overnight because Windows Memory Diagnostic can miss intermittent errors. Also test the system with the memory overclocking profile disabled, running the RAM at a standard JEDEC speed such as 2133 or 2666 MT/s. DDR4-3000 may be stable, but lowering it for several days is a useful diagnostic. If the crashes stop, the issue could be memory settings or the CPU’s integrated memory controller rather than a physically defective RAM stick.
If the crashes continue even at 2666 MT/s, would that confirm the RAM is bad, or could the CPU or motherboard still be responsible? I’ll run MemTest86 as well.