← Back to Blog

How to Tell If Your Power Supply Is Failing

October 15, 2026 · Luxurious Computers

The power supply unit (PSU) converts wall outlet AC power into the DC voltages your computer components need — 12V for the CPU and GPU, 5V for USB and SATA devices, 3.3V for RAM and chipset logic. When a power supply starts failing, it does not always fail completely and obviously. A PSU can degrade gradually, delivering unstable or insufficient power that causes symptoms you might attribute to other components. Understanding the signs of PSU failure helps you diagnose the actual problem instead of replacing parts that are working fine.

Random shutdowns with no warning

The most characteristic symptom of a failing PSU is the computer shutting off instantly — no blue screen, no shutdown sequence, no warning. The power just cuts as if someone pulled the plug from the wall. This happens because the PSU momentarily drops below the voltage or current threshold needed to keep the system running:

• Under load: If shutdowns happen during gaming, rendering, or heavy tasks, the PSU cannot deliver enough power at peak demand. It is either undersized (insufficient wattage) or degrading — capacitors lose capacity over time, especially in hot environments, reducing effective wattage below the rating
• At random: If the shutdowns happen at unpredictable times including during idle use, the PSU has an intermittent internal fault — a failing capacitor, a cracked solder joint on the main PCB, or a thermal protection circuit that triggers prematurely
• Distinguishing from overheating: CPU and GPU thermal shutdowns also cause instant power-off, but these happen specifically during heavy load after the temperature reaches the thermal limit (typically 100 to 105 degrees Celsius for CPUs). Monitor temperatures with HWiNFO during the workload that causes shutdowns. If temperatures are normal (under 85C) when the shutdown occurs, overheating is not the cause

Random restarts

A computer that restarts on its own — going through the full boot sequence without you pressing anything — can indicate a PSU delivering voltage outside the acceptable tolerance. The ATX specification requires the 12V rail to stay within plus or minus 5 percent (11.4V to 12.6V). When the voltage sags below 11.4V momentarily, the motherboard may interpret it as a power cycle and restart:

• Check Event Viewer (Windows Logs, System) for Kernel-Power Event ID 41, which indicates an unexpected restart. If this event appears without a blue screen dump file, the restart was caused by a power interruption — the PSU or the wall power
• Check the wall outlet and power strip. A loose wall outlet, a failing surge protector, or a circuit shared with a high-draw appliance (space heater, air conditioner, microwave) can cause momentary voltage drops that the PSU cannot compensate for. Try a different outlet on a different circuit

Computer does not turn on at all

When you press the power button and absolutely nothing happens — no fans spin, no LEDs light up, no beep — the PSU is the most likely cause:

• Verify wall power: Check that the outlet works (plug in a lamp or phone charger), that the power cable is fully seated in both the wall and the PSU, and that the PSU's rear toggle switch (if present) is in the On position (the I side, not the O side)
• Paperclip test: Disconnect all cables from components. On the 24-pin ATX connector, use a paperclip to connect the green wire (PS_ON, pin 16) to any black wire (ground). The PSU should turn on and its fan should spin. If nothing happens, the PSU is dead. This test only confirms the PSU can power on — it does not verify voltage rails are within specification
• Smell check: If you smell a burnt or acrid odor from the back of the computer (where the PSU exhausts), a component inside the PSU has likely burned out. Do not continue using a PSU that smells burnt — replace it immediately

Burning smell or visible damage

Any burning smell coming from the PSU area is a serious warning sign. PSU failures can involve capacitor venting (a chemical smell, sometimes described as fishy), resistor burnout (acrid, sharp electrical smell), or in severe cases, melted plastic or smoke:

• If you see smoke or smell burning from the PSU, shut the computer down immediately using the PSU's rear toggle switch or unplugging it from the wall. Do not use the Windows shutdown — cut power directly
• A PSU that has vented a capacitor or burned a component should be replaced, not reused. Even if the computer seems to work after the event, internal damage may cause unstable power delivery that damages other components over time
• In rare cases, a catastrophic PSU failure can send unregulated voltage to components, damaging the motherboard, CPU, RAM, or storage. This is more common with cheap unbranded units that lack over-voltage protection. Quality PSUs include protection circuits that shut down the unit before it damages other components

Electrical noise and coil whine

Some electrical noise from a PSU is normal — a faint hum or buzz under load is common and not a sign of failure. However, certain sounds indicate problems:

• Clicking or ticking: A rhythmic clicking sound from the PSU, especially one that changes with load, can indicate a failing relay or a component arcing internally. This is different from the single click you hear when turning the PSU on (which is normal — that is the internal relay engaging)
• Loud buzzing that changes with load: While light coil whine is normal, a loud buzz that was not present before or that has gotten louder over time suggests loosening transformer windings — a sign of aging
• Fan bearing noise: PSU fans can develop grinding, rattling, or intermittent spinning sounds as their bearings wear out. A failed PSU fan does not directly cause power issues, but without cooling, the PSU's internal temperature rises, accelerating capacitor degradation and potentially triggering thermal shutdown

USB devices disconnecting

If USB devices (keyboard, mouse, external drives, webcams) randomly disconnect and reconnect — you hear the Windows disconnect and reconnect sounds — the PSU's 5V rail may be sagging. USB devices require stable 5V power. When the PSU cannot maintain 5V under load, connected devices lose power momentarily:

• This symptom is often confused with USB port problems, bad USB cables, or driver issues. The distinguishing factor is if the disconnections happen when the system is under heavy load (gaming, rendering) and affect multiple USB devices on different ports simultaneously. A bad USB port or cable would affect only one device on one port
• Front panel USB ports are more susceptible because they use extension cables from the motherboard that add resistance, making them more sensitive to voltage drops

Blue screens and system instability

Unstable power delivery can cause seemingly random blue screens, application crashes, and file corruption because the CPU, RAM, and storage devices require stable voltage to operate correctly:

• A PSU delivering slightly low voltage on the 12V rail can cause CPU errors that manifest as blue screens with stop codes like WHEA_UNCORRECTABLE_ERROR, KERNEL_DATA_INPAGE_ERROR, or UNEXPECTED_KERNEL_MODE_TRAP. These stop codes are also caused by actual CPU, RAM, and storage failures — which is why PSU problems are often misdiagnosed
• The clue that points to the PSU rather than other components: if memtest passes, if storage drive health checks are normal (use CrystalDiskInfo), if the CPU is not overclocked, and if the blue screens do not follow a pattern tied to a specific application or driver, the PSU becomes a prime suspect
• Voltage ripple — small oscillations in the DC output that should be smooth — increases as PSU capacitors age. High ripple can cause data corruption on storage drives, memory errors, and erratic behavior that defies logical diagnosis. Measuring ripple requires a multimeter or oscilloscope on the PSU's output rails, which is beyond typical home diagnosis

How long power supplies last

PSU lifespan depends heavily on quality, load, and operating temperature:

• Budget units (3 to 5 years): Entry-level power supplies using lower-quality capacitors, especially those rated at 85C rather than 105C, degrade faster. Running these units at high load (above 80 percent of rated wattage continuously) shortens their lifespan further
• Mid-range units (5 to 7 years): Power supplies from established brands with 5-year warranties typically use better capacitors and tighter voltage regulation
• High-end units (7 to 10+ years): Premium power supplies with 10-year warranties use Japanese 105C-rated capacitors and can outlast the rest of the system
• Environmental factors: Heat is the primary enemy of PSU longevity. Poor case airflow, hot rooms, and fan intake blocked by carpet accelerate degradation. Dust buildup inside the PSU reduces cooling and shortens lifespan

Testing a power supply

Beyond the basic paperclip test, you can check PSU voltage levels:

• Software monitoring: HWiNFO or HWMonitor reads voltage sensors on the motherboard. Check 12V, 5V, and 3.3V readings. Readings consistently below spec (11.8V on the 12V rail, 4.85V on the 5V rail) suggest the PSU is failing. Software readings are approximations — motherboard sensors have their own tolerance
• Multimeter testing: For more accuracy, measure the PSU's Molex or SATA connectors while running. Yellow wire to black (ground) should read 11.4 to 12.6V. Red wire to black should read 4.75 to 5.25V
• PSU tester: Dedicated PSU testers plug into all connectors and display voltage readings. They flag out-of-spec values but do not test the PSU under load

PSU diagnosis in Columbus

LuxuriousComputers tests power supply units under real load conditions — checking voltage stability, ripple, and thermal behavior to determine whether the PSU needs replacement or if the symptoms point to another component.