How to Diagnose Failing PC Hardware: A Practical Order of Checks

Hardware faults are diagnosed badly because the symptoms overlap. Failing RAM, a dying drive, an overheating CPU and a struggling power supply can all present as random crashes. Guessing costs money; testing in the right order costs an evening.

This is that order. Cheapest and most likely first.

Read the symptoms first

That last row matters. If it only crashes in one application, hardware is unlikely. You’d expect a failing component to affect everything eventually.

Order of checks

The tools worth having

• HWiNFO. Temperatures, voltages, fan speeds, with logging so you can see what happened just before a crash.
• CrystalDiskInfo. SMART data in plain language. Anything other than Good deserves attention.
• MemTest86. Bootable, so it tests memory without Windows in the way. The definitive RAM test.
• Windows Reliability Monitor. perfmon /rel, to see when it started and what changed.
• Event Viewer. Windows Logs, System, filtered to Critical and Error.

# Quick storage check without extra tools
Get-PhysicalDisk | Select-Object FriendlyName, HealthStatus, OperationalStatus

# Read-only filesystem scan
chkdsk C: /scan

# Recent unexpected shutdowns
Get-WinEvent -FilterHashtable @{LogName='System'; Id=41,1001,6008} -MaxEvents 20 |
    Select-Object TimeCreated, Id, Message

Event ID 41 is “the system rebooted without cleanly shutting down”. The fingerprint of a power or hardware fault rather than a software crash.

The symptoms that mislead

• A failing drive looks like a slow computer. Long pauses and unresponsive windows, not error messages.
• Bad RAM looks like software corruption. Programs crashing, files failing to open, Windows appearing damaged.
• A struggling PSU looks like anything. Random restarts with no blue screen is the classic pattern.
• Overheating looks like poor performance. Throttling makes a machine slow rather than crashing it.
• A dying GPU can look like a driver problem, right up until the artifacts appear in the firmware screen too.

Knowing when to stop

If you’ve tested memory clean overnight, storage reports healthy, temperatures are fine under load and it still crashes randomly, you’re into PSU or motherboard territory. Both are diagnosed by substitution, which means borrowing parts or accepting the cost of guessing.

At that point, weigh it against the machine’s age. On something five or more years old, a new PSU plus a motherboard is a substantial fraction of a new build, and you’d be putting money into a platform that’s already limiting you.

Glossary

Frequently asked questions

How long should I run MemTest86?

At least four complete passes, which usually means overnight. Intermittent memory faults can take hours to show, and a single quick pass proves very little.

Is Windows Memory Diagnostic good enough?

It’s better than nothing and takes minutes, but it’s considerably less thorough than MemTest86. Use it for a quick check; use MemTest86 before concluding memory is fine.

My PC restarts with no blue screen. What causes that?

Usually power. A failing PSU, or one being pushed past its limits by transient spikes. Check Event Viewer for ID 41, and check temperatures to rule out thermal shutdown.

Should I replace the PSU if I can’t find anything else?

It’s the reasonable next step once memory, storage and cooling are ruled out, particularly if the unit is old or was cheap. Borrow one first if you possibly can.

Can software cause the same symptoms as failing hardware?

Yes. Bad drivers especially. That’s why Reliability Monitor is worth checking early: if the crashes started exactly when something was installed, test that theory before buying parts.

Disclosure: this post may contain affiliate links. If you buy through one of them, we may earn a small commission at no extra cost to you. We only recommend products we’ve tested or genuinely rate.