No blue screen. No error. The PC just goes black, instantly, like someone pulled the plug. When that pattern follows heavy work or warm afternoons, you are usually looking at a power supply quietly giving up.
A sudden shutdown with no blue screen and no warning is the power system raising its hand. Windows crashes leave evidence — blue screens, error logs, restarts. A machine that dies instantly and completely usually lost power, not software. And inside the SMPS, the usual culprit is not a dramatic failure. It is ageing capacitors and heat, shrinking the unit's strength a little every month until ordinary work becomes too much.
Software faults and power faults die differently, and the difference saves hours:
Check cooling honestly too: a dust-choked CPU cooler also shuts PCs down under load. Clean vents and a temperature reading separate that cause in minutes — do not skip it, and do not let it excuse the SMPS from testing either.
Inside every SMPS, capacitors smooth the power into clean, steady rails. Heat and years dry them out. A dried capacitor still works — just less. The rails gain ripple, sag a little deeper under every load spike, and the unit's real capacity quietly shrinks.
That is why the pattern creeps: first a shutdown a month, then weekly, then daily. The PC did not change. The supply's strength fell below the PC's appetite, one hot afternoon at a time. Swollen or leaking capacitor tops, visible through the fan grill with a torch, are the confession — but you never open the unit to look for them.
You already own the test equipment: the workload that kills it. Note what the PC was doing at each death — game, render, five heavy tabs — then reproduce it deliberately while watching.
+12 V rail under that load. Sagging below 11.40 V under load is a failing supply, measured, not guessed.Write the times and workloads down. A page of notes beats a month of impressions when you talk to a seller or technician.
The shutdowns themselves are not the worst of it. Every hard power cut risks the file that was open, the drive that was writing, and the Windows installation itself. Worse, a supply in decline can start feeding dirty, unstable power to the motherboard and drives on its way out.
So this fault has a deadline the others do not. Test promptly, and if the pattern matches and the numbers sag, replace the unit — do not nurse it along because it still works on light days. Light days are exactly how it fools you.
The swap test is decisive here too: borrow a known-good SMPS of equal or higher wattage and run the exact workload that used to kill the PC, for twice as long.
When you buy, buy honestly: real wattage math for your parts, plus headroom for age and future upgrades. A quality unit sized with margin is what ends this pattern for years, not months.
The video walks through the SMPS testing method live — use it alongside this pattern check before you replace anything.
▶️ Watch Full VideoA restart with no blue screen can be the same fault — a rail dipping so briefly that the board resets instead of dying. Windows can also restart on errors by itself. Disable automatic restart in Windows settings and watch: if it still dies instantly with no blue screen, treat it as a power pattern and test the SMPS.
Software almost always leaves fingerprints — blue screens, error events, freezes before death. A PC that loses power mid-frame with a totally clean log is describing hardware. Check the Event Viewer: a lonely 'unexpected shutdown' entry (Event ID 41) with no bugcheck code is Windows saying the power simply vanished.
Rarely. The card raised the system's appetite past a tired or small supply's real capacity. The unit may be healthy but undersized, or old and now exposed. Add up the system's real draw, check the +12 V rail under load, and size the replacement for the PC you actually have now.