My Skytech gaming PC has continued crashing and restarting even after a completely clean Windows 11 installation. The system has a Ryzen 7 5700, RTX 5050, 16GB of DDR4-3200 memory, and a 1TB NVMe SSD. Before reinstalling Windows, I saw PAGE_FAULT_IN_NONPAGED_AREA and KERNEL_MODE_HEAP_CORRUPTION errors, along with Kernel-Power events. Since reinstalling, the crashes have included SYSTEM_SERVICE_EXCEPTION, repeated DRIVER_POWER_STATE_FAILURE (0x9F), an EXRESOURCE TIMEOUT live dump, and CACHE_MANAGER (0x34). The 0x9F error has appeared several times, but the dump summaries do not clearly identify a particular third-party .sys file; some point to nt.sys, which may only be where the failure surfaced.
CHKDSK completed successfully with no bad sectors, DISM and SFC are clean, and I have not changed BIOS settings or opened the case because the computer is still under warranty. I currently use a TP-Link Archer TXE5100UH USB Wi‑Fi 6E adapter, while the PC also has a built-in Realtek 8821CE wireless adapter. The TP-Link driver is from September 2025, and the Realtek adapter shows an older Microsoft-provided driver.
Could the TP-Link adapter or its driver be causing the repeated DRIVER_POWER_STATE_FAILURE crashes? Could the built-in Realtek adapter be involved, especially with both devices present? Would disabling or removing one adapter at a time—or testing with Ethernet—be a sensible diagnostic step? Does the mix of bugchecks suggest a memory, motherboard, storage, or general driver problem, and is the CACHE_MANAGER crash likely related? I have already provided the dump information to the manufacturer and would prefer warranty-safe troubleshooting rather than randomly changing drivers or BIOS settings.
2 Answers
The variety of stop codes makes a single Wi‑Fi driver less certain. Driver problems can produce different symptoms, but recurring memory corruption and heap errors also justify testing the RAM. Run a bootable, preferably overnight memory test if that does not conflict with the warranty process. If memory passes, the motherboard, storage, power delivery, or another device may still be involved. The clean SFC, DISM, and CHKDSK results do not rule out faulty hardware. CACHE_MANAGER could be related to broader memory or storage corruption, but it may also be a separate crash, so the actual dump analysis matters more than the label alone.
A 0x9F usually means a device or driver failed during a power-state transition, so either wireless adapter is a reasonable suspect. Having two wireless devices installed can complicate testing, even if only one is actively connected. The safest first test is to disable one adapter in Device Manager, leave the other unchanged, and use the system normally for a while. Then reverse the test. If possible, testing with Ethernet is even cleaner because both Wi‑Fi devices can remain disabled. Since the machine is under warranty, avoid BIOS changes or aggressive driver-cleaning tools and document each test for the manufacturer.
I would also avoid treating nt.sys as the cause by itself. Windows often reports the component where the failure became visible rather than the driver that triggered it.
Testing one wireless device at a time is a good low-risk starting point, but if crashes continue with both adapters disabled, give the manufacturer the full pattern of bugchecks and request hardware diagnostics rather than continuing to swap drivers randomly.