My Windows 11 PC crashes roughly every 2.5 to 3 hours whether it is idle or running a game. During gaming, I have seen SYSTEM_SERVICE_EXCEPTION (0x3B), while other crashes report SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (0x7E), IRQL_NOT_LESS_OR_EQUAL (0xA), KMODE_EXCEPTION_NOT_HANDLED (0x1E), and MEMORY_MANAGEMENT (0x1A). Some failures instead produce a black screen while the computer remains powered on, requiring a hard shutdown. Minidumps are created after these events.
The system uses a new Intel Core i7-14700KF, ASRock Z790 Riptide WiFi motherboard, new DDR5 memory with XMP disabled, an RTX 4070 Ti Super, a new power supply, and an AddGame NVMe SSD. The BIOS is updated and configured with Intel default settings. GPU stress tests, Cinebench, Windows Memory Diagnostic, and MemTest86 have not reported problems.
I have also used DDU to reinstall the graphics driver, run SFC and CHKDSK, and disabled C-states and Windows NVMe/PCIe power-saving features. The crashes reference different drivers, including Intel Wi-Fi software, FLTMGR.SYS, and ntoskrnl.exe. What is the most likely cause, and what should I test next?
3 Answers
Treat the minidumps as more useful than the names shown by BlueScreenView. Kernel files such as ntoskrnl.exe are often victims rather than the original cause. Collect several dumps and analyze them together, because one isolated crash can point at the wrong module.
Since the system has already had multiple replaced parts and the failures occur at idle as well as under load, do not focus only on GPU or CPU stress tests. Test each RAM stick alone, verify the BIOS microcode, and consider testing with the NVMe drive or motherboard replaced if the memory-isolation test does not change anything.
The different stop codes and modules do not necessarily mean all of those drivers are faulty. The dump patterns look more like memory corruption. That can come from a defective DIMM, the CPU's integrated memory controller, motherboard-related memory instability, or less commonly storage. Passing MemTest86 does not completely rule out intermittent DDR5 problems.
First, confirm that the BIOS really includes Intel microcode 0x12F or a newer revision; 0x12B is not the final fix for the earlier 13th/14th-generation voltage issue. Leave all overclocks and undervolts disabled. Then test the machine normally with only one memory stick installed, repeating the test with the other stick. If only one stick causes crashes, suspect that module. If both sticks produce the problem individually, the CPU or motherboard memory path becomes more likely. The memory controller is inside the CPU, so a CPU fault can look exactly like bad RAM.
The Intel Wi-Fi driver appearing in one dump is worth investigating, but it may simply be where corrupted data happened to surface. Update or temporarily remove the Wi-Fi adapter driver, and if possible test with Wi-Fi disabled or use Ethernet. Also check the motherboard support page for chipset, LAN, Wi-Fi, and storage-controller drivers rather than relying only on Windows Update.
The CPU was replaced through an RMA after the previous one was considered damaged, and the power supply was replaced after a repair shop suspected it during earlier random restarts. The current crashes started after those changes, so I will test each memory stick separately and verify the exact microcode version.