I have an HP 255 G8 with 16 GB of RAM running Windows 11 25H2. An Amazon reseller replaced the original SSD with a Fastro MS150-512GB using a Silicon Motion SM2263XT DRAM-less controller and HMB. The BIOS and available HP drivers are up to date.
The laptop repeatedly crashes with CRITICAL_PROCESS_DIED (0xEF), but only while running on battery. It remains stable when connected to the charger, even under similar workloads. The crashes also happen while the laptop is stationary, so movement does not appear to be the cause.
The BSOD dump process freezes at 0%, and no minidump is created. Event Viewer records volmgr Event 161, Ntfs Event 98, Kernel-Power Event 41, and an unexpected-shutdown event. This suggests the boot drive becomes unavailable while Windows is trying to handle the crash.
Memory tests, SMART, storage diagnostics, DISM, SFC, and read-only CHKDSK all pass. The SSD reports 96% health with no media or data-integrity errors. The battery retains about 95% of its design capacity. Disabling disk sleep, PCIe Link State Power Management, and raising the minimum processor state did not help.
My suspicion is that a battery-related NVMe power-state transition, possibly APST, causes the DRAM-less SM2263XT drive to disappear from the PCIe bus. Is this a known failure pattern? Can APST or HMB be disabled or limited in Windows 11, or is replacing the SSD the sensible next step? If I replace it, would a drive with onboard DRAM be safer than another HMB-based model?
2 Answers
The missing dump files are probably a consequence rather than a separate configuration problem. If the boot volume drops offline during the crash, Windows cannot write the dump to C:, which explains the permanent 0% progress and volmgr Event 161. Since the system is stable on mains and the SSD was installed by the reseller, replacing the drive is a reasonable practical fix after making a full backup.
The storage device is the main suspect here. A clean SMART report does not rule out a controller, firmware, or power-state problem; SMART mainly tells you that the NAND is not reporting media failures. A drive can still disappear from the PCIe bus when entering or leaving a low-power state.
That matches what I am seeing. The drive appears healthy in normal diagnostics, but it vanishes specifically on battery power, so this looks more like a controller or power-management failure than worn-out flash.
The dump settings are already configured correctly and the Minidump folder stays empty. The drive disappearing at the same moment as the crash is why I am treating the failed dump as evidence of storage loss rather than as the original cause.