How should I narrow down random PC crashes and restarts?

November 20, 2026 5 answers
Asked by 7m!spoolbadge_418

My Windows 11 PC has been crashing and restarting unpredictably. Both monitors sometimes briefly go black, too. It has a Ryzen 9 7900X, an RTX 5070 connected through a PCIe 4.0 riser, and 64GB of DDR5 RAM running at 4800 MT/s.

I've tried each RAM stick separately and it still crashed with either one. I haven't run a dedicated memory test yet. What would you check next?

Answers

6d+canvasratchet_291

I'd start by removing the riser from the equation. If the case allows it, connect the GPU directly to the motherboard and see whether the crashes or display dropouts continue. A riser is a plausible suspect here, not a guaranteed culprit.

You can also try manually setting the PCIe slot to Gen 4 in the BIOS rather than leaving it on automatic. If that doesn't help, Gen 3 is another useful test. Change one thing at a time so you know which change affected the behavior.

FennelSwitch+3q:

A lower PCIe generation working would be a clue about the connection, but it wouldn't prove the riser itself is defective.

Reply to this answer
9P.lumvexo!4

The 7900X has integrated graphics, so another test is to remove the graphics card and riser, connect a monitor to the motherboard, and see if the system stays stable. That won't reproduce every gaming load, but a crash there would make the GPU and riser less likely to be the whole story.

Reply to this answer
6q!gimbalwicket_382

Run MemTest86 next. Testing each stick separately was useful, but crashing with either stick doesn't rule out memory or the rest of the memory subsystem. Note any reported errors, then use those results to guide further testing rather than immediately buying replacement RAM.

Reply to this answer
8w!brackenpiston_426

If you can borrow a known-good PSU with enough capacity, a swap is worth trying. Random restarts can come from the power supply as well as other hardware. Use only the cables supplied for the replacement PSU, not the cables from the current one.

Wall power is another possibility, though there's nothing here that establishes it as the cause. A suitably rated UPS might help with power interruptions, but I wouldn't buy one expecting it to diagnose the PC.

Reply to this answer
_x.63~Tivnaro+8

If the easier tests don't narrow it down, you could test a minimal setup outside the case to check for a case-related issue. Keep the CPU cooler installed and put the motherboard on a flat, nonconductive surface such as cardboard. Power down and disconnect the PC before moving components.

Keep a record of what changed and how long the machine ran before another crash. An intermittent problem can make a short stable period look like a fix when it isn't.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions