What could be causing rare instant power-offs across two different PC builds?

0
0
Asked By MellowCedar47 On

For roughly two years, my PC has occasionally powered off instantly under load with no blue screen or useful error message. The original system used a Gigabyte B450 I AORUS PRO WIFI, Ryzen 7 5800X, four 16GB DDR4 sticks, and an RTX 3090. Reducing the memory speed from 3600 to 3200 seemed to help temporarily, but the shutdowns eventually returned. Replacing the PSU in March 2025 did not resolve the issue.

In December 2025, I replaced the motherboard, CPU, memory, and graphics card. The current system is an MSI B650 Gaming Plus WiFi, Ryzen 7 7800X3D, 2x16GB DDR5-6000, and an RTX 5080, but it still has the same instant power-off behavior. The only parts shared between both builds are one M.2 drive, two SATA SSDs, and the Windows installation.

I also bought a second 2x16GB DDR5 kit to test memory. The new kit fails MemTest86 with thousands of errors even when installed by itself at stock JEDEC settings, while the original kit passes. All four sticks are unstable together. Windows Memory Diagnostic reports memory problems, and OCCT cannot complete testing reliably. BIOS is current, memory training settings were reset, and Memory Context Restore and Power Down Mode are disabled. The new kit is being returned.

The power supply is a Corsair RM850x. I have also tried different outlets, rooms, and power cables. Event Viewer only shows Kernel-Power Event 41 with BugcheckCode 0 and Event 6008. Monitoring from another system shows normal temperatures, voltages, and loads immediately before shutdowns; some happen while idle or under light gaming load. Could the remaining storage drives or Windows installation cause a complete instant power loss, or is there another hardware or power-related issue I should investigate?

4 Answers

Answered By RiverAsh6 On

Check firmware and memory-training behavior carefully. Update the motherboard BIOS, clear CMOS, run only a single matched memory kit, and leave all overclocking, EXPO, PBO, and undervolting disabled. Very long memory training followed by crashes is another sign that the newer RAM should not be used. If the system still shuts off with known-good memory at stock settings, test outside the case with only the essentials connected.

CopperWren29 -

The BIOS was already current, both memory kits were tested separately, and Memory Context Restore and Power Down Mode are disabled. The new kit is being returned, but the remaining power-offs also occur at idle with normal temperatures and power draw.

Answered By NovaPine8 On

The new memory kit is clearly defective or incompatible: thousands of MemTest86 errors at stock settings means it should be returned, and mixing two different kits is especially risky on DDR5. Test with one known-good matched kit using conservative settings first. That may explain the newer instability, although it probably does not explain every historical instant shutdown by itself.

QuartzLemon52 -

The older kit passes MemTest86, but Windows Memory Diagnostic and OCCT still report problems, so I would validate it with several longer tests and make sure the CPU's memory controller and motherboard socket are also checked.

Answered By BriskMarble31 On

Kernel-Power 41 with bugcheck code 0 only means Windows lost power or was shut down unexpectedly; it does not identify the failed component. Since the machine has done this with two complete platforms and a different PSU, focus on anything retained between builds: the storage devices, Windows installation, case wiring, front-panel switch, and external power path. A clean Windows installation with only one boot drive connected is a reasonable isolation test, though a faulty drive would more commonly cause freezes or errors than a literal power cut.

Answered By TidalMango64 On

An instant shutdown without a bugcheck usually points to power protection, a board-level fault, a loose connection, or a physical switch/wiring problem rather than Windows. Reseat the 24-pin, CPU EPS, and graphics power connectors, inspect for damaged pins, disconnect unnecessary SATA drives and front-panel connectors, and try a temporary case-power setup. If it persists with the known-good RAM, a minimal configuration, and a clean install, the motherboard or another shared component becomes much more suspicious.

Related Questions

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.