What are the symptoms of a failing power supply?

Failing PSUs typically cause sudden shutdowns under 3D load, random restarts without warning or blue screens, failure to power on at all (no fans, no lights), clicking noises, or a burnt smell.

How do I test if my PSU is dead using a paperclip?

You can test the PSU by unplugging it from all computer components. Take the 24-pin cable and use a bent paperclip to bridge the green wire (pin 16, PS_ON) to any black wire (ground, e.g., pin 15 or 17). Plug the PSU into the wall and flip the switch. If the fan spins, the PSU basic circuitry functions.

What are the safe voltage tolerance ranges for ATX PSUs?

According to Intel ATX design guides, a +/- 5% tolerance is allowed on all primary rails: +12V rail should be between 11.40V and 12.60V; +5V rail between 4.75V and 5.25V; +3.3V rail between 3.14V and 3.47V. Anything outside this range indicates hardware instability.

Can an insufficient PSU cause graphics card crashes?

Yes. Modern GPUs draw significant power and experience 'transient spikes' (brief demands for double their typical power). If your PSU cannot handle these brief surges, it triggers OCP (Overcurrent Protection) or OPP (Overpower Protection), shutting off the PC instantly to protect the system.

Power Supply (PSU) Troubleshooting – Complete Diagnostic Guide

A failing, faulty, or insufficient power supply unit (PSU) is one of the most difficult PC components to diagnose. It can manifest as sudden system shutdowns under 3D load, random restarts without blue screens, boot loops, or component-damaging voltage spikes. This guide walks you through diagnosing PSU problems systematically — from identifying wattage requirements to performing safety tests and reading system voltages.

⚡ Safe ATX PSU Voltage Tolerances (+/- 5%)

+12V Rail: 11.40V – 12.60V | +5V Rail: 4.75V – 5.25V | +3.3V Rail: 3.14V – 3.47V. Use HWiNFO64 to check these values under load or check the Hardware Monitor section in BIOS.

Most Common Signs of PSU Failure & Instability

Instantly shutting down under heavy load (looks like the plug was pulled)
Random PC restarts without showing a Blue Screen of Death (BSOD)
Loud clicking noises, squealing coil whine, or grinding fan bearings
PC fails to boot, but motherboard standby lights are illuminated
Noticeable burnt electrical smell coming from the rear exhaust vent
Random USB connectivity dropouts or PCIe errors due to ripple noise

How to Troubleshoot a PC Power Supply – Step by Step

Analyze crash patterns and logsIf your PC turns off instantly during 3D gaming or rendering, check Windows Event Viewer for Event ID 41 (Kernel-Power). While Event ID 41 indicates an unexpected shutdown, a shutdown without a corresponding bug check code (BSOD) points heavily to a PSU cut-off.
Calculate total system power drawLook up your GPU's typical power draw and your CPU's package power under load. Add them together, add 150W for system overhead, and add 20% safety margin. If you have a 300W GPU and 150W CPU, your system needs a high-quality 650W-750W PSU.
Perform the paperclip jumper testUnplug the PSU power cable from the wall. Disconnect all cables from the motherboard, GPU, and drives. Use a paperclip to bridge Pin 16 (Green, PS_ON) to Pin 17 or 15 (Black, Ground) on the 24-pin ATX cable. Turn the PSU back on. If the fan does not spin, the PSU is dead.
Monitor voltage rails under loadInstall HWiNFO64, run a CPU/GPU stress test (e.g. OCCT, Cinebench + Furmark), and look at the +12V, +5V, and +3.3V readings. If the +12V rail drops below 11.40V under heavy load, your PSU is failing or overloaded.
Use dedicated PCIe cables for high TDP GPUsEnsure your GPU is not powered via daisy-chained splitter cables. Use separate PCIe power cables from the PSU for each connector on the graphics card. This spreads the electrical load across separate +12V rails and prevents overcurrent shutdowns.
Swap with a known good test unitThe only 100% reliable diagnostic for an unstable power supply is swapping it with a known working test unit. If the system stops shutting down or restarting with the replacement unit, your original power supply must be replaced.

Related PSU Error Codes & Events

Code / ID	Error Name	Severity
`Event ID 41`	Kernel-Power – unexpected shutdown without BSOD	Critical
`PSU Insufficient`	PSU Insufficient – power supply wattage exceeded	Critical
`PSU Overvoltage`	PSU Overvoltage – motherboard triggers protection	Critical
`PSU Ripple Noise`	PSU Ripple Noise – unstable DC voltage delivery	High
`Power Limit Throttle`	POWER_LIMIT_THROTTLE – CPU/GPU throttles due to power limit	Medium
`PCIe Connection Fail`	PCIE_ERROR – graphics card reports power loss	High

Frequently Asked Questions

What are the signs of a failing power supply?

Common signs include sudden shutdowns under load, random restarts without warning, failure to boot, clicking noises, or a burnt smell.

What is the paperclip test for a PSU?

A test to see if a PSU powers on without being plugged into the motherboard by bridging Pin 16 (Green, PS_ON) to Pin 17 (Black, Ground). If the fan spins, the PSU basic standby rails are functional.

How do I check if my PSU wattage is sufficient?

Combine CPU TDP and GPU TDP, add 150-200W for memory, drives, fans, and headroom. E.g., a 300W GPU + 150W CPU needs at least a 650W PSU. High transient spikes on modern GPUs may require even higher wattage.

What are the safe voltage ranges for +12V, +5V, and +3.3V?

Intel ATX specifications allow a +/- 5% tolerance: +12V should be 11.4V to 12.6V; +5V should be 4.75V to 5.25V; +3.3V should be 3.14V to 3.47V. Dropping below these thresholds indicates instability.