diff --git a/reports/2026-09-15-clevo-system-stability-diagnostic-notes.yaml b/reports/2026-09-15-clevo-system-stability-diagnostic-notes.yaml new file mode 100644 index 0000000..2f6ede0 --- /dev/null +++ b/reports/2026-09-15-clevo-system-stability-diagnostic-notes.yaml @@ -0,0 +1,892 @@ +id: "BEF-2026-09-15-CLEVO-DIAG-001" + +organization: "Beffroi" + +title: "System Stability Investigation — Diagnostic Test Record" +subtitle: "Chronic NVMe controller errors, PCIe device instability and hard-freeze correlation — internal technical notes" + +report_type: technical +theme: default + +status: "open" +severity: "high" + +metadata: + Date: "2026-09-15" + Device: "DESKTOP-SAU019H" + Manufacturer: "PC Specialist / Clevo" + Platform: "NH5x_NH7x_HHx_HJx_HKx" + Motherboard: "CLEVO NHxxHJHK" + BIOS: "Insyde 1.07.07TPCS1 — 2021-09-16" + OS: "Windows 11 Pro — 10.0.26200 / build 26200" + CPU: "Intel Core i7-11800H" + Memory: "16 GB Corsair — single module" + Graphics: "Intel UHD Graphics + NVIDIA GeForce GTX 1650" + SystemDrive: "128GB PCS 2.5\" SATA SSD — serial U9DI16295T" + SecondaryStorage: "1TB PCS PCIe M.2 SSD — NVMe — intermittently unusable from Windows storage stack" + NVMeController: "Silicon Motion PCI VEN_126F DEV_2263 family — Microsoft stornvme" + Wireless: "Intel(R) Wi-Fi 6 AX200 160MHz" + Author: "Maximilien Rigaut" + Purpose: "Internal technical diagnostic notes; preserve evidence and working hypotheses pending hardware isolation" + +summary: > + Investigation of DESKTOP-SAU019H has progressed from a general Windows + stability investigation to a substantially more focused hardware/platform + investigation. Windows component-store and protected system-file corruption + identified during the investigation were successfully repaired. A subsequent + SFC validation after reboot reported no integrity violations. Fast Startup + was also disabled to remove the previously demonstrated hybrid-shutdown + failure path from further testing. + + Despite establishing this cleaner Windows baseline, the underlying instability + persists. The 1 TB PCS PCIe M.2 SSD and its NVMe controller continue to exhibit + repeated stornvme Event 129 device resets and Event 11 controller errors on + \Device\RaidPort2. These errors are not confined to the September incident: + review of historical event data establishes recurrence across March, June, + September 4, September 14 and September 15 2026. + + Windows currently demonstrates an important split in device state. The + Standard NVM Express Controller and the child "1TB PCS PCIe M.2 SSD" can be + enumerated as physically present by Plug and Play, with the controller + reporting Status OK, CM_PROB_NONE and an active PCIe Gen3 x4 link. However, + Get-Disk and Get-PhysicalDisk have failed to expose the 1 TB NVMe as a normal + usable storage device while the 128 GB SATA SSD remains available. + + The 2026-09-15 evidence adds a second major dimension to the investigation. + During the hard-freeze episode reported at approximately 09:18-09:20, the + Intel Wi-Fi 6 AX200 also entered a severe failure sequence consisting of an + internal adapter failure, TX/CMD timeout, reports that the adapter was not + functioning correctly, invalid driver-returned values and miniport reset + watchdog events. A stornvme Event 11 against \Device\RaidPort2 occurred + during the same failure cluster. + + Historical review identifies comparable coupled NVMe / AX200 episodes on at + least 2026-03-19 and 2026-06-08. This recurrence materially increases concern + that the observed instability may extend beyond an isolated Windows problem + or a single storage-device failure. A failing NVMe SSD remains a credible + explanation for the persistent RaidPort2 failures, but a shared platform-level + fault involving the M.2/PCIe path, motherboard/chipset, platform power + delivery, firmware or another common dependency must now be considered a + leading diagnostic branch. + + No single hardware component has yet been conclusively proven defective. + Hardware substitution and isolation testing are required before authorizing + replacement of the SSD, motherboard or other components. + +timeline: + - time: "2026-03-17 through 2026-03-26" + event: "Historical NVMe fault activity" + notes: > + Historical System log review shows repeated stornvme Event 129 resets + and/or Event 11 controller errors against \Device\RaidPort2 across + multiple days in March. The storage fault pattern therefore predates the + September investigation by approximately six months. + + - time: "2026-03-19 ~08:17-08:20" + event: "Historical coupled NVMe and AX200 failure episode" + notes: > + A RaidPort2 reset was recorded at approximately 08:17:31. Roughly + one and a half minutes later the Intel Wi-Fi 6 AX200 entered a failure + sequence including Event 5005 internal failure, Event 5007 TX/CMD timeout, + Event 5002 adapter-not-functioning reports and Event 5032 miniport reset + watchdog activity. stornvme subsequently recorded Event 11 against + \Device\RaidPort2 at approximately 08:19:26. This is the earliest + currently identified example of storage and wireless failures clustering + in the same short time interval. + + - time: "2026-06-08 ~12:10-12:16" + event: "Second historical coupled NVMe and AX200 failure episode" + notes: > + RaidPort2 reset events occurred at approximately 12:10:31 and 12:12:09. + At approximately 12:13:11 the AX200 experienced another TX/CMD timeout + and associated adapter/reset-watchdog failures. Another NVMe reset was + issued at approximately 12:14:01 followed by stornvme Event 11 at + approximately 12:14:04. Additional storage and wireless fault activity + continued within the same incident window. The similarity to the March + and September sequences materially strengthens the correlation. + + - time: "2026-06-09" + event: "Additional historical RaidPort2 errors" + notes: > + Additional stornvme Event 129 / Event 11 activity was identified without + requiring a simultaneous AX200 failure. This demonstrates that storage + instability can occur independently of the wireless failure sequence. + + - time: "2026-09-04 before ~17:23" + event: "Hard-freeze incident" + notes: > + System experienced a hard freeze before approximately 17:23. Exact freeze + onset was not timestamped. + + - time: "2026-09-04 ~17:25" + event: "Second hard-freeze / hang" + notes: > + A second hang was observed at approximately 17:25. Normal shutdown could + not subsequently be completed and a forced power-off was required. + + - time: "2026-09-04" + event: "Firmware/NVRAM warning observed" + notes: > + Firmware displayed "A first boot or NVRAM reset condition has been + detected." This remains an independent unresolved firmware-state finding. + The warning alone does not establish failure of the RTC/CMOS battery. + + - time: "2026-09-04" + event: "Fast Startup failure identified" + notes: > + Kernel-Boot Event 29 recorded a Fast Startup failure with status + 0xC000007B. Restart and a full shutdown bypassing Fast Startup subsequently + completed successfully. + + - time: "2026-09-04" + event: "RaidPort2 errors identified" + notes: > + System logs showed repeated stornvme Event 129 resets and Event 11 + controller errors against \Device\RaidPort2 around the incident period. + At this stage the responsible physical storage device had not yet been + conclusively isolated. + + - time: "2026-09-14" + event: "Fast Startup disabled for controlled diagnosis" + notes: > + powercfg /h off completed successfully. This disables hibernation and + Windows Fast Startup, removing the previously abnormal hybrid-shutdown + path as a variable during subsequent testing. + + - time: "2026-09-14" + event: "Windows component-store repair" + notes: > + DISM /Online /Cleanup-Image /RestoreHealth completed at 100 percent with + "The restore operation completed successfully" and "The operation + completed successfully." + + - time: "2026-09-14" + event: "Windows protected-file repair" + notes: > + sfc /scannow found corrupt Windows files and successfully repaired them. + + - time: "2026-09-14 after reboot" + event: "Post-repair Windows integrity verification" + notes: > + A subsequent sfc /scannow reported "Windows Resource Protection did not + find any integrity violations." The Windows integrity repair therefore + remained clean across the reboot. + + - time: "2026-09-14" + event: "Storage inventory discrepancy identified" + notes: > + Get-Disk and Get-PhysicalDisk exposed only the 128 GB PCS 2.5-inch SATA + SSD. The expected 1 TB PCS PCIe M.2 SSD was not exposed as a normal disk. + + - time: "2026-09-14" + event: "NVMe controller positively enumerated" + notes: > + Get-PnpDevice -PresentOnly enumerated a Standard NVM Express Controller + using PCI identifier VEN_126F&DEV_2263. The controller reported Status OK + and was bound to the Microsoft stornvme driver. + + - time: "2026-09-14" + event: "1 TB NVMe child device positively enumerated" + notes: > + Get-PnpDevice -PresentOnly -Class DiskDrive enumerated + "1TB PCS PCIe M.2 SSD" with Status OK. The 128 GB SATA SSD was also + present. + + - time: "2026-09-14" + event: "NVMe PnP state inspected" + notes: > + Detailed controller properties reported Present True, HasProblem False, + CM_PROB_NONE and ConfigManagerErrorCode CM_PROB_NONE. Windows therefore + did not consider the PCI device absent or in a conventional PnP error + state. + + - time: "2026-09-14" + event: "NVMe PCIe link inspected" + notes: > + Controller properties identified PCI bus 2, device 0, function 0 with + current link speed 3 and link width 4, matching maximum link speed 3 and + width 4. The NVMe controller was therefore actively linked at PCIe Gen3 + x4 during inspection. + + - time: "2026-09-14 16:23:21" + event: "Fresh stornvme controller error" + notes: > + stornvme Event 11 reported "The driver detected a controller error on + \Device\RaidPort2." This occurred after Windows repair and while the NVMe + controller was currently enumerated. + + - time: "2026-09-15 08:54:49" + event: "Current diagnostic boot" + notes: > + Intel AX200 initialization events were recorded during the current boot. + The adapter initially entered service normally. + + - time: "2026-09-15 08:55:20" + event: "RaidPort2 reset" + notes: > + stornvme Event 129 reported "Reset to device, \Device\RaidPort2, was + issued." + + - time: "2026-09-15 08:56:23" + event: "Second RaidPort2 reset" + notes: > + A second stornvme Event 129 reset was issued against \Device\RaidPort2 + approximately one minute after the first. + + - time: "2026-09-15 08:58:17" + event: "RaidPort2 controller error" + notes: > + stornvme Event 11 reported a controller error against + \Device\RaidPort2. The storage path was therefore already producing + errors well before the later hard freeze. + + - time: "2026-09-15 ~09:18-09:20" + event: "Hard system freeze" + notes: > + User observed another hard system freeze. The event-log cluster immediately + surrounding this incident contains simultaneous severe AX200 and NVMe + storage failures. + + - time: "2026-09-15 09:19:52" + event: "AX200 severe failure begins" + notes: > + Netwtw10 Event 5005 reported that Intel(R) Wi-Fi 6 AX200 160MHz + encountered an internal error and failed. Event 5007 reported a TX/CMD + timeout with a hung TfdQueue. + + - time: "2026-09-15 09:19:53-09:19:57" + event: "AX200 failure and recovery cascade" + notes: > + The AX200 generated Event 5002 reports stating that the adapter was not + functioning properly, followed by a large burst of Event 5010 messages + reporting invalid values returned to the driver. Event 5032 subsequently + recorded Driver Miniport reset watchdog failures. This represents a + substantive adapter/driver/device communication failure rather than a + routine Wi-Fi disconnect. + + - time: "2026-09-15 09:19:56" + event: "NVMe controller error during AX200 failure cluster" + notes: > + stornvme Event 11 reported "The driver detected a controller error on + \Device\RaidPort2" during the same second as the AX200 Event 5010 storm. + The temporal coincidence is diagnostically significant, particularly + because comparable clustering is present in March and June historical + logs. + +findings: + - title: "Chronic RaidPort2 / NVMe storage-path instability" + severity: "high" + status: "confirmed" + evidence: > + Repeated stornvme Event 129 device resets and Event 11 controller errors + against \Device\RaidPort2 occur across multiple months, including March, + June, September 4, September 14 and September 15 2026. Fresh resets and + controller errors continue after Windows component-store and system-file + repair. + assessment: > + The recurrence establishes a persistent low-level storage-path problem. + This is no longer reasonably explained as a single transient Windows + incident. The evidence demonstrates failure somewhere in the NVMe device / + M.2 slot / PCIe path / associated platform stack, but does not by itself + identify which physical component is defective. + + - title: "1 TB NVMe is enumerated but not exposed as a usable disk" + severity: "high" + status: "confirmed" + evidence: > + PresentOnly PnP enumeration identifies both the Standard NVM Express + Controller and its 1TB PCS PCIe M.2 SSD child device. The controller + reports Status OK, CM_PROB_NONE and an active PCIe Gen3 x4 link. Despite + this, Get-Disk and Get-PhysicalDisk expose only the 128 GB SATA SSD. + assessment: > + PCIe enumeration alone is therefore not evidence that the NVMe storage + device is operational. The controller can establish a PCIe/PnP presence + while the storage stack is unable to expose the underlying media as a + normal disk. This strongly implicates the NVMe storage path. + + - title: "Severe Intel AX200 failure correlated with September 15 freeze" + severity: "high" + status: "confirmed" + evidence: > + During the September 15 hard-freeze window, the AX200 generated Event + 5005 internal failures, Event 5007 TX/CMD timeout, Event 5002 + adapter-not-functioning reports, a very large Event 5010 invalid-value + burst and Event 5032 miniport reset watchdog failures. + assessment: > + The wireless subsystem experienced a genuine low-level failure during + the freeze interval. This does not prove that the AX200 caused the system + freeze. Its diagnostic importance is primarily the temporal correlation + with the concurrent NVMe controller failure. + + - title: "Repeated cross-device NVMe / AX200 failure correlation" + severity: "high" + status: "confirmed-correlation" + evidence: > + Comparable short-window clustering of RaidPort2 reset/controller errors + and severe AX200 failures has now been identified on at least three + distinct dates: 2026-03-19, 2026-06-08 and 2026-09-15. + assessment: > + Repetition on three separated dates makes accidental one-time temporal + coincidence substantially less plausible. The evidence establishes a + repeatable correlation, not causation. A common platform dependency, + platform power issue, PCIe/chipset problem, firmware interaction or a + severe fault in one device disrupting the wider platform are credible + explanations requiring hardware isolation. + + - title: "RaidPort2 failures also occur without visible AX200 failure" + severity: "medium" + status: "confirmed" + evidence: > + RaidPort2 resets/controller errors are present on additional dates where + the reviewed event window does not show a corresponding severe AX200 + failure cluster. + assessment: > + The NVMe/storage problem therefore has an independent persistent + manifestation. The AX200 correlation appears to accompany some severe + incidents rather than every storage-path fault. This prevents treating + the AX200 as a necessary cause of every RaidPort2 event. + + - title: "Windows component-store corruption" + severity: "medium" + status: "repaired" + evidence: > + DISM /Online /Cleanup-Image /RestoreHealth completed successfully. + assessment: > + Windows corruption was genuine but has been repaired. Continued + stornvme failures after repair substantially weaken Windows component-store + corruption as the primary explanation for the continuing hardware + symptoms. + + - title: "Protected Windows system-file corruption" + severity: "medium" + status: "repaired-and-verified" + evidence: > + sfc /scannow initially found and repaired corrupt files. A subsequent SFC + run after reboot reported no integrity violations. + assessment: > + Windows protected-file integrity is currently clean. Earlier corruption + may have resulted from forced power-offs or storage instability, may have + contributed to instability, or may have been unrelated. Direction of + causality is not established. + + - title: "Fast Startup failure" + severity: "medium" + status: "mitigated-for-diagnosis" + evidence: > + Earlier Kernel-Boot Event 29 identified a Fast Startup failure with status + 0xC000007B. powercfg /h off was subsequently used successfully. + assessment: > + Fast Startup was a real abnormality but is now removed from the diagnostic + path. The persistence of storage errors and another hard freeze with + hibernation/Fast Startup disabled demonstrates that Fast Startup cannot + explain the entire incident. + + - title: "Firmware / NVRAM reset warning" + severity: "medium" + status: "open" + evidence: > + Firmware previously displayed "A first boot or NVRAM reset condition has + been detected." + assessment: > + This remains unresolved. The warning is insufficient by itself to diagnose + a failed RTC/CMOS battery. A retention test is still required. Given the + emerging possibility of a platform-level fault, the firmware finding + should remain visible rather than being dismissed as unrelated. + + - title: "BIOS storage visibility" + severity: "medium" + status: "inconclusive" + evidence: > + During manual BIOS inspection, the expected 1 TB NVMe was not visibly + listed in the inspected storage menus. + assessment: > + This observation is useful but not conclusive because the exact Clevo / + Insyde firmware interface may not expose NVMe inventory in the inspected + menu. Windows subsequently demonstrated physical PCIe/PnP enumeration of + the NVMe controller and child disk. BIOS-menu absence must therefore not + be treated as proof that the PCIe device failed firmware enumeration. + + - title: "128 GB SATA SSD remains normally exposed" + severity: "info" + status: "currently-operational" + evidence: > + Get-Disk and Get-PhysicalDisk consistently expose the 128GB PCS 2.5-inch + SATA SSD as Online / Healthy / OK. Reliability telemetry reported + approximately 2033 power-on hours, 40 C temperature, wear 0 and no + reported total or uncorrected read errors. + assessment: > + Available telemetry does not presently implicate the SATA system disk in + the same manner as the missing NVMe device. These counters must not be + mistakenly attributed to the 1 TB NVMe. + +diagnostics: + confidence_scale: + type: "investigator-estimate" + range: "0-100" + interpretation: > + Confidence values below are engineering diagnostic estimates derived from + the currently available evidence. They are not statistical probabilities, + failure-rate calculations or manufacturer test results. They express the + relative strength of each working hypothesis and should be revised after + hardware isolation testing. + + hypotheses: + - id: "D1" + diagnosis: "Persistent fault in the 1 TB NVMe / M.2 / PCIe storage path" + confidence: "very-high" + confidence_percent: 95 + scope: "fault-domain" + status: "strongly-supported" + evidence_for: + - "Repeated stornvme Event 129 resets across multiple months." + - "Repeated stornvme Event 11 controller errors across multiple months." + - "Fresh Event 129 and Event 11 activity after Windows repair." + - "1 TB NVMe PnP device is present while Get-Disk does not expose usable media." + - "NVMe controller can establish PCIe Gen3 x4 while storage functionality remains unavailable." + evidence_against: + - "PnP controller reports Status OK and CM_PROB_NONE." + - "No hardware substitution test has yet identified the exact failed element." + interpretation: > + Confidence is high that the fault domain includes the 1 TB NVMe storage + path. This does not mean there is 95 percent confidence that the SSD + module itself is defective; SSD, connector, slot, motherboard path, + power and firmware remain inside this fault domain. + tests_to_confirm: + - "Reseat the 1 TB NVMe and repeat enumeration/event checks." + - "Test the suspect NVMe in a known-good compatible system or enclosure." + - "Test a known-good compatible NVMe in the Clevo M.2 slot." + + - id: "D2" + diagnosis: "Failure of the 1 TB NVMe SSD module itself" + confidence: "moderate-high" + confidence_percent: 70 + scope: "component" + status: "probable-but-unconfirmed" + evidence_for: + - "Persistent controller resets/errors associated with its storage path." + - "Device can enumerate while usable disk media is unavailable." + - "Symptoms recur over several months." + evidence_against: + - "The controller establishes a normal-width PCIe Gen3 x4 link." + - "No direct SMART/health telemetry from the missing NVMe has been obtained." + - "A defective M.2 slot, motherboard path or power source can produce similar symptoms." + - "Concurrent AX200 failures raise concern about a wider platform issue." + interpretation: > + The SSD remains a strong individual-component suspect but should not yet + be condemned solely from stornvme Event 11/129. A/B substitution is + required before assigning the fault specifically to the SSD. + tests_to_confirm: + - "Test suspect SSD externally or in another known-good PCIe/NVMe host." + - "Install a known-good NVMe in the Clevo slot and monitor Event 11/129." + + - id: "D3" + diagnosis: "Motherboard / chipset / shared PCIe platform instability" + confidence: "moderate-high" + confidence_percent: 75 + scope: "platform" + status: "leading-alternative" + evidence_for: + - "At least three dates show temporally coupled NVMe and AX200 severe failures." + - "September 15 includes simultaneous NVMe Event 11 and AX200 failure storm during a hard freeze." + - "March and June historical logs show similar cross-device clustering." + - "Firmware/NVRAM anomaly exists as an additional unresolved platform-level observation." + evidence_against: + - "RaidPort2 failures also occur without a severe AX200 failure." + - "No WHEA evidence currently proves PCIe root-port or chipset failure." + - "A severely malfunctioning endpoint device can potentially destabilize software or shared platform resources without the motherboard being defective." + - "No known-good-device substitution test has yet reproduced the failure." + interpretation: > + Cross-device recurrence materially raises the motherboard/platform branch + above a speculative possibility. It is now a leading alternative to an + isolated SSD failure, but direct evidence identifying a motherboard + component is still absent. + tests_to_confirm: + - "Known-good NVMe in Clevo slot." + - "Suspect NVMe in known-good host." + - "If possible, operate Clevo temporarily without suspect NVMe and monitor stability." + - "Review PCIe root-port, WHEA and ACPI/platform events during any future incident." + - "Correlate AX200 and NVMe physical PCIe topology where possible." + + - id: "D4" + diagnosis: "Shared platform power-delivery instability" + confidence: "moderate" + confidence_percent: 60 + scope: "platform" + status: "plausible" + evidence_for: + - "Multiple PCIe-connected devices have demonstrated severe failures in the same short windows." + - "Hard freezes are compatible with severe platform-level power or bus disruption." + - "Firmware/NVRAM anomaly provides an additional reason to retain power-state/platform hypotheses." + evidence_against: + - "No voltage measurement or rail telemetry is available." + - "No direct power-controller fault has been logged." + - "The storage fault can occur without simultaneous AX200 failure." + interpretation: > + Power delivery is a credible mechanism for correlated endpoint failures + but is presently inferred rather than directly measured. + tests_to_confirm: + - "Repeat testing with known-good AC adapter if available and electrically compatible." + - "Inspect motherboard and M.2 area for physical damage or thermal evidence." + - "Compare behavior with suspect NVMe removed." + - "Escalate to board-level diagnosis if substitution testing clears the SSD." + + - id: "D5" + diagnosis: "Firmware / BIOS / EC interaction affecting PCIe or power state" + confidence: "moderate" + confidence_percent: 50 + scope: "platform-firmware" + status: "plausible-unconfirmed" + evidence_for: + - "Firmware previously reported a first-boot / NVRAM-reset condition." + - "Failures involve devices whose state is affected by platform firmware and power management." + - "System firmware dates from 2021 while current Windows and drivers are substantially newer." + evidence_against: + - "No firmware-specific error directly explains RaidPort2 failures." + - "No BIOS/EC update comparison has been performed." + - "No evidence currently proves that the NVRAM warning and PCIe failures share a cause." + interpretation: > + Firmware remains relevant but should not currently outrank the directly + evidenced NVMe fault domain. Firmware update should not be used as the + first uncontrolled intervention because doing so would alter the + diagnostic baseline. + tests_to_confirm: + - "Perform NVRAM/RTC retention test." + - "Document current BIOS configuration and exact firmware availability." + - "Consider firmware update only after evidence preservation, hardware isolation and recovery preparation." + + - id: "D6" + diagnosis: "Intel AX200 is the primary root cause of the system freezes" + confidence: "low-moderate" + confidence_percent: 30 + scope: "component" + status: "possible-but-not-leading" + evidence_for: + - "AX200 enters severe TX/CMD timeout and reset-watchdog failure during the September 15 freeze." + - "Comparable AX200 failure clusters exist in March and June." + evidence_against: + - "Chronic RaidPort2 failures occur independently on many dates." + - "Storage errors pre-exist some AX200 failure clusters." + - "A concurrent failure does not establish that AX200 initiated the incident." + - "The missing/unusable 1 TB NVMe is independently abnormal." + interpretation: > + The AX200 is demonstrably failing during some incidents, but current + evidence more strongly supports treating it as either another victim of + a shared platform fault or a secondary contributor rather than the sole + root cause. + + - id: "D7" + diagnosis: "Windows corruption / Fast Startup as primary remaining root cause" + confidence: "low" + confidence_percent: 10 + scope: "software" + status: "substantially-deprioritized" + evidence_for: + - "Windows component-store and protected-file corruption were previously confirmed." + - "Fast Startup previously failed." + evidence_against: + - "DISM repair completed successfully." + - "SFC repair completed successfully." + - "Post-reboot SFC reported no integrity violations." + - "Fast Startup was disabled." + - "Fresh RaidPort2 errors occurred after repair." + - "Another hard freeze occurred after establishing the cleaner software baseline." + - "Storage failure history spans multiple months." + interpretation: > + Windows problems were real and warranted repair, but they no longer + provide a satisfactory primary explanation for the persistent incident. + They may instead be consequences or secondary effects of storage errors + and forced shutdowns. + + - id: "D8" + diagnosis: "RTC/CMOS battery or NVRAM-retention fault" + confidence: "low-moderate" + confidence_percent: 35 + scope: "firmware-retention" + status: "open" + evidence_for: + - "Firmware displayed a first-boot / NVRAM-reset warning." + evidence_against: + - "A single NVRAM warning does not establish RTC battery failure." + - "No controlled retention test has yet demonstrated loss of clock or settings." + - "No direct link between RTC retention and RaidPort2/AX200 failures is established." + interpretation: > + This remains a separate open diagnostic branch. It may eventually + contribute to a broader motherboard diagnosis, but the current evidence + does not justify combining the two findings. + tests_to_confirm: + - "Record BIOS clock and settings." + - "Perform complete shutdown and remove external power for a controlled interval." + - "Re-enter BIOS directly and verify clock/settings retention." + - "Record whether the NVRAM warning recurs." + + current_leading_assessment: + confidence: "high" + statement: > + A persistent hardware-level or platform-level fault is now more likely + than a purely Windows/software root cause. The strongest directly evidenced + fault domain is the 1 TB NVMe / M.2 / PCIe storage path. However, repeated + temporal correlation between severe NVMe and AX200 failures makes an + upstream motherboard/chipset/power/firmware problem sufficiently plausible + that the SSD should not be replaced as a conclusively diagnosed component + without A/B hardware testing. + + not_yet_supported: + - "The 1 TB SSD is definitively defective." + - "The Clevo motherboard is definitively defective." + - "The Intel AX200 caused the hard freezes." + - "The CMOS/RTC battery caused the NVRAM warning." + - "The NVRAM warning and NVMe failures have a common cause." + - "A BIOS update will resolve the problem." + - "Windows corruption caused the original freezes." + - "The SATA SSD is responsible for the RaidPort2 events." + +sections: + - title: "Commands and results — Windows baseline" + content: + - type: bullets + items: + - > + powercfg /h off — completed successfully; hibernation and Fast + Startup disabled. + - > + DISM /Online /Cleanup-Image /RestoreHealth — completed successfully. + - > + sfc /scannow — corrupt files found and repaired. + - > + Post-reboot sfc /scannow — no integrity violations detected. + - > + Windows repair is therefore retained as completed remediation, not + as proof of root cause. + + - title: "Commands and results — storage enumeration" + content: + - type: bullets + items: + - > + Get-Disk — only the 128GB PCS 2.5-inch SATA SSD exposed as a normal + disk during the affected state. + - > + Get-PhysicalDisk — only the 128 GB SATA device exposed. + - > + Get-PnpDevice -PresentOnly -Class SCSIAdapter — Standard NVM Express + Controller present and Status OK. + - > + Get-PnpDevice -PresentOnly -Class DiskDrive — both 1TB PCS PCIe M.2 + SSD and 128GB PCS 2.5-inch SSD enumerated with Status OK. + - > + NVMe controller properties — Present True, HasProblem False, + CM_PROB_NONE, Microsoft stornvme. + - > + PCIe properties — active Gen3 x4 link, matching the controller's + reported maximum Gen3 x4 capability. + - > + The discrepancy between PnP presence and Get-Disk usability is a + central diagnostic finding. + + - title: "Commands and results — storage event history" + content: + - type: bullets + items: + - > + stornvme Event 129 repeatedly reports resets to + \Device\RaidPort2. + - > + stornvme Event 11 repeatedly reports controller errors on + \Device\RaidPort2. + - > + Events recur across March, June and September 2026. + - > + September 14 produced a fresh Event 11 after Windows repair. + - > + September 15 produced Event 129 at 08:55:20 and 08:56:23, + Event 11 at 08:58:17, and another Event 11 at 09:19:56 during + the hard-freeze/AX200 failure window. + + - title: "Commands and results — wireless correlation" + content: + - type: bullets + items: + - > + September 15 09:19:52 — AX200 Event 5005 internal failure and + Event 5007 TX/CMD timeout. + - > + September 15 09:19:53 onward — Event 5002 reports adapter not + functioning correctly. + - > + September 15 09:19:56 — large Event 5010 invalid-value burst. + - > + September 15 09:19:56 — stornvme Event 11 occurs during the same + second as the AX200 Event 5010 storm. + - > + September 15 09:19:57 — AX200 Event 5032 Driver Miniport reset + watchdog failures. + - > + Comparable short-window NVMe / AX200 failure clustering was found + on March 19 and June 8. + + - title: "Interpretation at this checkpoint" + content: + - type: callout + title: "Current state" + kind: "warning" + text: > + The investigation has crossed the threshold where a purely Windows + explanation is no longer satisfactory. Persistent RaidPort2 failures + are directly evidenced, the 1 TB NVMe is not being exposed as a usable + disk despite current PCIe/PnP enumeration, and severe AX200 failures + repeatedly correlate with some of the more serious NVMe episodes. + Hardware isolation is now the highest-value next step. + + - type: bullets + items: + - "Windows component-store corruption has been repaired." + - "Windows protected-file integrity is currently clean." + - "Fast Startup has been disabled and cannot explain the September 15 recurrence." + - "Persistent NVMe/RaidPort2 instability is confirmed." + - "The individual 1 TB SSD has not yet been conclusively proven defective." + - "A motherboard/platform fault is now a leading alternative diagnosis." + - "AX200 failure is real but is not established as the initiating cause." + - "NVRAM/RTC retention remains an independent open diagnostic branch." + - "No BIOS flash or TPM clear is currently justified as a first-line diagnostic action." + + - title: "Evidence-preservation considerations" + content: + - type: bullets + items: + - > + Avoid unnecessary heavy I/O against the 1 TB NVMe before important + data is secured. + - > + Preserve System event logs containing Event 11, 129 and the AX200 + failure clusters before major firmware or driver changes. + - > + Record exact wall-clock time of any future freeze before forcing + power off whenever possible. + - > + Change one diagnostic variable at a time. Do not combine SSD reseat, + BIOS update, driver replacement and firmware reset into a single + intervention. + - > + Photograph the 1 TB NVMe label and physical installation before + reseating or replacement. + +actions_title: "Recommended continuation" + +actions: + - title: "Preserve important data" + status: "priority" + text: > + Ensure important user data is backed up before additional storage + diagnosis. Do not deliberately stress the suspect 1 TB NVMe merely to + reproduce Event 11/129 before data preservation. + + - title: "Perform controlled NVMe physical inspection" + status: "pending" + text: > + Fully shut down the machine using shutdown /s /t 0, disconnect AC and, + if servicing procedure permits, disconnect the internal battery before + handling the M.2 device. Locate the 1 TB NVMe, photograph its label and + installation, inspect the retaining screw, seating, connector area, + thermal pad and surrounding PCB for visible abnormalities. + + - title: "Reseat the 1 TB NVMe" + status: "pending" + text: > + Remove and reinstall the NVMe carefully, changing no unrelated BIOS or + Windows setting during the same intervention. The purpose is to test the + connector/contact branch while preserving diagnostic isolation. + + - title: "Repeat post-reseat enumeration" + status: "pending" + text: > + After reseating, repeat Get-Disk, Get-PhysicalDisk and PresentOnly + DiskDrive/SCSIAdapter enumeration. Determine whether the 1 TB device + becomes a normal disk, remains PnP-only, or disappears entirely. + + - title: "Review stornvme immediately after reseat" + status: "pending" + text: > + Review stornvme events from the new boot/session for Event 129 resets and + Event 11 controller errors. Record exact timestamps and whether the system + remains responsive. + + - title: "Perform SSD / slot A-B isolation" + status: "pending" + text: > + If the problem persists, test the suspect 1 TB NVMe in a known-good + compatible host or NVMe enclosure and test a known-good compatible NVMe + in the Clevo M.2 slot. This is the highest-value test for separating an + SSD failure from a motherboard/M.2/PCIe-path failure. + + - title: "Consider temporary operation without suspect NVMe" + status: "pending" + text: > + If operationally practical and the SATA system disk contains the required + Windows installation, test system stability with the suspect NVMe + physically removed. Continued AX200/platform freezes without the NVMe + would materially strengthen the motherboard/platform branch; disappearance + of the instability would strengthen the SSD/storage-endpoint branch. + + - title: "Perform firmware/NVRAM retention test separately" + status: "pending" + text: > + Record BIOS date/time and selected harmless settings, save configuration, + perform a complete shutdown and leave the system without external power + for a controlled interval. Re-enter firmware directly and verify retention. + Do not combine this test with the NVMe reseat if avoidable. + + - title: "Do not clear TPM at this stage" + status: "deferred" + text: > + TPM clearing is not currently supported by the storage/PCIe evidence and + may introduce recovery-key or device-encryption complications. Preserve + the existing TPM state unless later evidence specifically justifies this + intervention and recovery information has been secured. + + - title: "Do not perform uncontrolled BIOS update yet" + status: "deferred" + text: > + A BIOS/EC update may eventually become relevant, but hardware isolation, + current-firmware documentation and evidence preservation should precede + firmware flashing. Do not use a firmware update as a substitute for + component isolation. + + - title: "Correlate any future hard freeze" + status: "pending" + text: > + Record the exact wall-clock freeze time. After recovery, review at least + five minutes before and after the incident for stornvme, disk, WHEA-Logger, + Netwtw10, Kernel-Power, Kernel-PnP, ACPI and related platform events. + Particular attention should be paid to whether RaidPort2 and AX200 again + fail in the same short interval. + + - title: "Prepare manufacturer escalation evidence" + status: "recommended" + text: > + Preserve a concise manufacturer-facing extract showing the repeated + RaidPort2 Event 129/11 history and the March 19, June 8 and September 15 + coupled NVMe/AX200 failure clusters. Include the observation that the + September 15 cluster coincided with a user-observed hard freeze. + +closing: > + The investigation now provides strong evidence of a persistent fault in the + NVMe/M.2/PCIe storage domain and substantially reduces the likelihood that + Windows corruption or Fast Startup alone caused the recurring hard freezes. + The 1 TB NVMe SSD remains a major suspect, but the repeated correlation of + severe NVMe and Intel AX200 failures across at least three distinct dates + raises credible concern for an upstream motherboard, chipset, PCIe, platform + power or firmware problem. + + The correct next diagnostic phase is therefore physical and comparative rather + than additional broad Windows remediation. Reseating the NVMe provides a + low-cost contact test, while A/B testing the suspect SSD and Clevo M.2 slot is + required to distinguish endpoint failure from platform failure. Until those + tests are completed, neither SSD replacement nor motherboard replacement + should be represented as a conclusive diagnosis. + + Diagnostic confidence should be revised after each controlled hardware test. + In particular, a suspect SSD that reproduces failure in a known-good host + would strongly support SSD replacement; a known-good NVMe that reproduces + RaidPort2 failures in this Clevo would strongly support motherboard/platform + escalation; and continued cross-device freezes with the suspect NVMe removed + would materially increase confidence in a shared platform-level defect. \ No newline at end of file