Spot a Failing PC Power Supply Before It Stops

A desktop PC power supply converts mains electricity into several regulated DC voltages for the motherboard, graphics card and drives. Failures can develop gradually, but the first visible event may be a sudden shutdown during a game. Knowing how to find out if your computer’s power supply is failing before it stops working is therefore a matter of combining symptoms, load conditions and safe physical checks.

Many symptoms are deceptive. A graphics card, faulty memory, a crashing driver or an overheated CPU can all cause a reset. The goal is to identify the failing part without replacing good components or making the power supply more dangerous.

Stop immediately if there is a burning smell, buzzing, sparking, smoke or scorch mark. Unplug the PC from the wall and do not switch it on again. A computer power supply can retain a dangerous charge after disconnection, so opening its casing is not a sensible first step.

Local conditions matter. Australian mains is nominally 230 volts at 50 hertz, and computers generally use Type I plugs. A loose plug, overloaded power board, tripped residual-current safety switch or damaged extension lead can imitate a dying supply. The checks below move from simple observations towards evidence a repairer can trust.

Recognise Early Warning Signs

Random restarts become significant when they repeatedly follow the same high-load activity, especially a graphics-heavy game. A power supply that cannot hold the required voltage may trigger protection and shut the machine down. Record whether the display goes black, the PC powers itself off, or Windows restarts without warning.

Listen and look as well. A PSU fan that stutters, stops during a known heavy load, or is much noisier than before may be nearing failure. Check for excessive heat at the casing or power plug, discoloration, bubbling paint and a persistent burnt-electronics smell. Fans may deliberately stop at idle, so judge them under load.

Coil whine, a single click or an occasional black screen are weaker clues. They can come from a graphics card, display cable, monitor or software. If one heavy application consistently triggers the problem while light browsing does not, the power and cooling systems deserve closer attention.

Rule Out Cables, Outlets and Heat

Begin outside the case. Unplug the PC, inspect the mains lead and plug for heat damage, then try a known-good wall outlet. Do not repeatedly cycle a safety switch or continue using an outlet that sparks, smells hot or trips. An electrician should investigate those signs; they are not proof that the PSU is faulty.

Environmental heat can produce load-related shutdowns. Summer afternoons in Melbourne, Sydney and Brisbane can push a compact case and its power supply towards their thermal limits, while dust in Perth can restrict airflow. Clean accessible dust from the case and confirm every internal fan spins freely, with the PC off and disconnected.

Reseat the 24-pin motherboard connection and relevant graphics or CPU power plugs while disconnected, using the manual’s guidance. Stop if a connector is discoloured or loose. Modular PSU cables are not interchangeable between units, even when the brands match, so never substitute a cable from another supply.

Create a Repeatable Test Record

Create a short test record before changing parts. In Windows, Reliability Monitor and Event Viewer can show unexpected restarts, the application involved and nearby hardware errors. A “Kernel-Power” entry often only says the system shut down unexpectedly; it does not identify the cause. Note drivers, temperatures and whether the issue follows a particular game or workload.

If there are no signs of heat, arcing or burning, reproduce the workload for 10 to 20 minutes while watching CPU and GPU temperatures and fan behaviour with reliable monitoring software. Stop at the first abnormal noise or shutdown. A controlled load test can expose a fault that remains hidden at idle, but it is not proof and can aggravate a weak unit.

An inexpensive PSU tester checks only basic output voltages and usually does not test regulation, ripple, high-load behaviour or protection circuits. It may pass a supply that fails under load. Multimeter testing inside the PSU requires opening a mains-powered, capacitor-backed device; leave that to a technician with a proper bench load test and safe isolation equipment.

For a repairer, retailer or warranty claim, send the model number, serial, event times, temperatures and clear photographs in one concise message. This WhatsApp sharing guide can help keep that evidence organised for the exchange.

Compare Symptoms Before Buying a Replacement

Verify the exact replacement before buying. Check the wattage, supported connectors, 12-volt output, model number, serial, warranty and 80 PLUS certification where claimed. Also confirm compatibility with the motherboard and graphics card. A high headline wattage does not guarantee a stronger or more reliable unit.

Compare reputable local stockists such as JB Hi-Fi or Officeworks with specialist computer stores and established online sellers. Marketplace listings can look convincing while the actual unit differs, so a page such as the Virtex WA product reference should not replace checking the physical label and serial. Ask for an invoice and written warranty terms, and keep the packaging.

For goods bought in Australia, consumer guarantees under the Australian Consumer Law may apply to a defective product even when the manufacturer’s warranty has expired; the exact remedy depends on the circumstances. Treat the retailer’s policy and the manufacturer’s support process as separate evidence. A genuine-looking replacement that behaves differently may still create a compatibility problem.

Use the pattern, not one symptom, to decide what to test next. This separates the most common PSU clues from components that often produce the same visible result.

Observation	PSU concern	Another possible cause	Sensible next check
Random reboot during a GPU-heavy game	Voltage instability or overcurrent protection	GPU or VRM overheating, driver fault, memory error	Compare logged temperatures and repeat at a lower load
Sudden power-off under load	Thermal or electrical shutdown	CPU overheating, board fault	Stop repeated testing and inspect cooling and connections
PSU fan stutters, clicks or stops	Worn fan or internal power fault	Case airflow restriction	Listen and record the fan during a known high load
Display flickers or blacks out	Weak 5-volt or standby output	Display cable, monitor or graphics card	Test with a known-good cable and display
Burning smell, scorch marks or buzzing	Imminent electrical failure	Damaged socket or external arcing	Unplug immediately and arrange professional inspection

Remember the Decisive Checks

The most trustworthy diagnosis is usually a convergence of evidence. A recurring load-linked shutdown, combined with fan or heat changes and a clean wall outlet, cables and cooling system, is more persuasive than any single warning light. A cheap tester passing does not clear the unit, and a non-specific Windows crash does not condemn it.

Remember the sequence: isolate the outlet and cables, stop at any sign of burning or arcing, record the failure pattern, rule out overheating and graphics problems, then use a compatible replacement or professional load testing. A load-linked failure after those checks is strong PSU evidence, while unexplained output instability means disconnection and professional service—not opening the supply.