Power Supply Failures Explained: Causes, Symptoms, and Fixes
You're halfway through a game when the monitor flickers, the graphics card fans stutter, and the room goes quiet. You press the power button. Nothing. A laptop can give the same warning in a quieter way: it closes normally, then refuses to wake, shows no charge light, or runs only while the adapter is held at a particular angle.
These failures rarely arrive without clues. A brief click, warm-electronics smell, USB devices resetting, unexplained restarts, or a machine that loops during startup can all point towards unstable power. The difficulty is that a failing power supply can imitate a motherboard, GPU, RAM, battery, or charging-board fault. The symptoms matter, but the order in which you test them matters more.
Table of Contents
The Moment Your Machine Goes Dark - Read the behaviour before opening anything
What a Power Supply Actually Does - From mains input to usable DC - Why one fault can affect everything
Common Causes and Symptoms of PSU Failure - Ageing hardware leaves a trail - External events can produce immediate failure
Diagnosing a Suspected Power Supply Failure - Begin with the five-minute checks - Escalate without guessing
Power Supply Myths That Cost You Money - More wattage isn't automatically better - A dead PSU doesn't always destroy the computer - Laptop adapters follow a different path - Wattage labels don't make brands interchangeable - Cleaning helps, but it doesn't reverse damage
Safety Rules Before You Touch Anything - These signs mean workshop, not screwdriver - What not to touch
Your Next Steps and When to Call a Workshop - Tonight - This week - Before the next upgrade
The Moment Your Machine Goes Dark
A desktop that dies during a demanding game creates a particularly misleading scene. The fans may surge just before the shutdown, the screen may lose signal, and the computer may restart once before going completely dead. Some users notice a small click from the case. Others hear a short buzz, smell heated plastic or electronics, or find that the machine works again after sitting unplugged for a while.
Those signs don't prove the PSU is at fault, but they're useful evidence. A system that fails only when the GPU is working hard suggests a different problem from one that shows no motherboard lights at all. A restart loop can mean unstable delivery, a protection circuit reacting to a short, or another component pulling a rail outside its expected range.
Workshop observation: “It was fine yesterday” often means the final failure happened yesterday. The underlying warning may have been weeks of random USB resets, sleep-wake failures, or brief black screens.
Laptop users often miss the early symptoms because the adapter and battery hide the power path. The charge indicator may blink, the battery percentage may stop increasing, the display backlight may flicker, or the charger brick may become unusually hot. A laptop that runs from its battery but refuses to charge points towards the adapter, DC jack, charging circuit, or battery path rather than automatically proving the mainboard has failed.
Read the behaviour before opening anything
Write down exactly what happens when you press the button:
No lights or sound: Check the outlet, cable, adapter, fuse, and external connections before assuming an internal failure.
Fans twitch, then stop: The PSU may be entering protection, or a shorted component may be pulling the system down.
Starts, then dies under load: Heat, ageing components, or inadequate stable output become more plausible.
Repeated clicking or looping: Stop repeated power attempts and arrange a proper diagnosis.
Burning, chemical, or ozone smell: Unplug it immediately and don't continue testing.
The rest of the diagnosis is about translating those odd moments into likely failure points, while keeping unsafe workshop work away from the kitchen table.
What a Power Supply Actually Does
Think of mains electricity as turbulent water arriving through a high-pressure pipe. Your computer doesn't want that turbulence delivered directly to its components. The power supply unit, or PSU, acts like a regulator, reducing, filtering, and distributing the input into controlled streams that the motherboard, processor, graphics card, drives, and USB devices can use.
From mains input to usable DC
The conversion follows several functional stages:
AC input and protection: Mains alternating current enters through the socket, fuse, filtering components, and surge protection. A metal-oxide varistor, or MOV, can sacrifice itself during a major surge.
Rectification: The PSU converts alternating current into a form that can be stored and processed inside the unit.
Filtering: Capacitors smooth the converted energy. Heat and age can dry or weaken them, allowing ripple and instability to reach later stages.
Switching: High-frequency switching circuitry controls how energy passes through the transformer. This makes modern PSUs smaller and more efficient than old linear designs.
Regulation: Feedback circuits keep the output within the limits expected by the computer. Failed MOSFETs, controllers, or solder joints can interrupt that control.
DC output: The unit sends regulated power along the motherboard and component connectors.
The familiar rails are 3.3V, 5V, and 12V. Modern processors and graphics cards draw heavily from the 12V supply, so a weak 12V path can cause a GPU to shut down during a game. The 5V rail supports parts of USB power and storage electronics, so instability there can appear as devices disconnecting and reconnecting. The 3.3V rail serves other motherboard and low-voltage circuits.
Why one fault can affect everything
A PSU contains switching electronics, capacitors, transformers, cooling hardware, protection circuits, and connectors. Fan bearings can wear out, dust can restrict airflow, thermal cycling can fatigue solder joints, and capacitors can deteriorate after years of heat. A unit may still light the motherboard while failing to provide clean power when the GPU or processor demands more current.
That makes the PSU a single power source for the entire system, not just another replaceable box. Testing it before replacing a graphics card or motherboard prevents expensive parts swapping and reduces the risk of running valuable hardware from a supply that's already unstable.
Common Causes and Symptoms of PSU Failure
Power supply failures usually fall into a few practical categories, but the symptoms overlap. A technician doesn't diagnose from one noise or one missing light. The useful pattern is when the failure occurs, whether the machine is under load, and whether the behaviour changes after cooling or disconnecting peripherals.
Cause Category | Desktop Symptoms | Laptop Symptoms |
|---|---|---|
Age and component wear | Intermittent shutdowns, restarts during games, fan noise, buzzing, or failure after warming up | Charger instability, battery not charging reliably, flickering backlight, or an adapter that runs hot |
Electrical stress | Sudden death, a tripped protection device, a click followed by silence, or power-on followed by immediate shutdown | No charge light, a dead adapter, intermittent charging, or failure after a surge |
Heat and contamination | Thermal shutdowns, erratic restarts, loud fan operation, and instability after dust buildup | Hot adapter brick, charging dropouts, or symptoms that worsen in enclosed spaces |
Liquid or physical damage | Corrosion, burning smell, melted connectors, or complete loss of power | No charging, shorted DC input, board damage, or battery-management faults |
Poor design or defective parts | Early failure, unexplained instability, inadequate protection, or shutdowns under normal load | Adapter overheating, unstable output, or repeated charge interruptions |
Ageing hardware leaves a trail
Capacitors lose performance, fans slow down, and solder joints experience repeated expansion and contraction as the unit heats and cools. The user often sees this as a desktop that runs normally on the web but switches off when a game, rendering task, or benchmark increases demand.
Don't confuse a louder fan with proof that the PSU is failing. It can indicate dust, a worn bearing, restricted case airflow, or a unit working hard. It becomes more meaningful when paired with shutdowns, hot exhaust, electrical odour, or a failure that appears only after sustained use.
External events can produce immediate failure
A lightning event, bad wall outlet, surge, or brownout can damage input protection or switching components. Some units stop completely. Others still start but produce unstable output, which may show up as random restarts, USB resets, or a system that loses display when the GPU begins drawing current.
UK network evidence supports looking beyond dramatic national shortage stories. Ofgem's 2001/02 quality-of-supply report recorded 87.4 minutes of interruption per customer across the year, with around 70% of interruptions and minutes lost linked to high-voltage network faults in its historical reporting. That benchmark describes network disruption, not a diagnosis for an individual PC, but it explains why local supply events can matter even when the wider grid remains dependable.
The symptom table is a starting point, not a verdict. Motherboards, GPUs, RAM, batteries, charging circuits, and wall equipment can create the same outward behaviour.
Diagnosing a Suspected Power Supply Failure
Start outside the machine. The safest checks often eliminate loose connections before you reach for a meter.
Begin with the five-minute checks
Try another wall socket: Plug the desktop directly into a known-working socket, bypassing the power strip or extension lead. Test the monitor separately so you don't mistake a display fault for a dead computer.
Inspect and reseat cables: Unplug and reconnect the mains lead, then reseat the PSU-side and component-side connections. Pay particular attention to the 24-pin motherboard connector and the CPU 8-pin connector. A connector that looks fitted can still be incompletely latched.
Remove extras: Disconnect USB hubs, external drives, lighting controllers, and other accessories. A damaged peripheral or cable can trigger protection.
Record startup behaviour: Note motherboard LEDs, fan movement, beeps, clicking, smell, and whether the machine dies instantly or after several seconds.
A brief fan twitch isn't a clean pass. It tells you that some startup activity occurred, not that the rails remain stable under load.
Escalate without guessing
A known-good PSU swap is often more informative than repeatedly pressing the power button. Use a compatible unit with equivalent or greater capacity, connect the motherboard and CPU power first, and reduce the system to a basic POST configuration. Never mix modular cables between PSU brands or models, even when the plugs fit.
A paperclip test only checks whether a start signal can make the PSU switch on. It doesn't show how the unit behaves under a real load or whether its output is clean. An ATX load tester gives a clearer screening result, while a multimeter can check individual rails when used by someone who understands connector pinouts and safe probing.
For a practical connector-by-connector walkthrough, use this guide to testing a desktop power supply. If the problem affects several essential devices at home, it can also help to understand how a VPP keeps critical appliances powered, particularly where backup and load management are being considered.
Stop testing if you smell burning, see arcing, find melted insulation, or hear repeated clicking. At that point, the useful next step is controlled bench diagnosis, not another restart.
Repair or Replace How to Decide
Repair makes sense only when the unit is worth repairing, the fault is understood, and the work can be performed safely. A warranty claim should always come before opening a relatively new PSU. Opening it can expose dangerous circuitry and may remove the manufacturer's coverage.
PSU Age | Symptom Type | Typical Action |
|---|---|---|
Under three years | Intermittent restarts, buzzing, inconsistent fan behaviour, or shutdowns | Check warranty first. Don't open the unit. |
Three to six years | Degraded output, burnt smell, repeated protection trips, or load-related failure | Replacement is usually the sensible route after diagnosis. |
Over seven years | Any instability, especially in a high-demand gaming PC | Replace rather than invest in component-level repair. |
Any age | Damaged cable, failed external fuse, or removable modular lead | Replace the affected external part if compatibility is confirmed. |
A laptop power brick may have an accessible external fuse in some designs, but many modern adapters are sealed. Desktop modular cables can sometimes be replaced without repairing the PSU electronics, provided the exact compatible cable is used. A fan replacement may restore cooling in a suitable repair, but it still requires opening a high-voltage enclosure and should be treated as workshop work.
Decision rule: A repairable fault isn't automatically an economical repair.
Take a five-year-old 650W desktop unit that switches off when a game loads. The bench time needed to confirm the fault, inspect the rails, replace ageing parts, and test it under load can exceed the value of a dependable replacement. If the diagnosis also finds heat damage or stressed switching components, replacing the unit avoids rebuilding an ageing failure point.
For a safe installation checklist and compatibility considerations, see this power supply replacement guide. The replacement must match the system's connectors and demand, not just the wattage printed on the old label.
Power Supply Myths That Cost You Money
Several PSU assumptions lead people towards unnecessary upgrades or unsafe experiments.
More wattage isn't automatically better
A 1200W supply in a system that needs far less power doesn't cure instability. It can be less appropriate for the actual load, more expensive, and capable of hiding a poor diagnosis. Match capacity to measured or credibly estimated system demand, then leave sensible room for startup peaks and a future graphics upgrade.
A dead PSU doesn't always destroy the computer
Modern ATX designs commonly include protections such as over-voltage protection, over-current protection, and short-circuit protection. Those systems can shut the PSU down before a fault spreads, although protection isn't a guarantee against every surge, component defect, or wiring mistake. A clean shutdown is often less damaging than continued operation from an unstable supply.
Laptop adapters follow a different path
A laptop adapter feeds a charging and power-management system, not a desktop motherboard's interchangeable ATX rails. That's why a laptop can run from its battery while refusing to charge, or charge intermittently while the computer itself remains functional. The adapter, DC jack, battery, charging controller, and mainboard must be tested as separate parts of the path.
Wattage labels don't make brands interchangeable
Two units with the same capacity can differ in efficiency, rail arrangement, thermal design, component quality, and protection implementation. 80 Plus Bronze, Gold, and Titanium describe efficiency certification levels, not a complete reliability guarantee. Check the exact model's independent electrical testing and connector layout rather than choosing by the badge alone.
Cleaning helps, but it doesn't reverse damage
Dust restricts cooling and can accelerate thermal stress. It won't restore a dried capacitor, repair a cracked solder joint, or undo surge damage. Clean external vents with the system disconnected, but don't treat cleaning as proof that a failing PSU has recovered.
Never open a PSU because a video makes capacitor replacement look simple. Stored energy and mains-side components make that a qualified repair task.
Safety Rules Before You Touch Anything
Power supply work needs a sharper safety boundary than ordinary cable reseating. Shut the computer down properly, switch the PSU off if it has a rear switch, unplug the mains lead, and disconnect the external adapter from a laptop. Pressing the case power button after disconnection can help remove residual system power, but it doesn't make an opened PSU safe.
Wait before handling internal parts, and keep the unit disconnected while you inspect the case. Use a clean, grounded work surface and an anti-static wrist strap where possible. Avoid carpet, wool clothing, and loose jewellery, and hold circuit boards by their edges.
These signs mean workshop, not screwdriver
Bulging or leaking capacitors: Stop immediately. Don't press, puncture, or discharge them with improvised tools.
Scorching or melted connectors: Photograph the damage and leave the wiring disconnected.
Chemical or ozone smell: Unplug the equipment and don't perform another power-on test.
Liquid ingress: Remove external power and seek board-level inspection.
Lightning or mains-surge damage: A visual check may miss damage in the input and switching stages.
Burn injury or sparking: Treat the device as unsafe and don't handle the damaged area unnecessarily.
What not to touch
Don't open a desktop PSU or sealed laptop brick unless you're trained to work around high-voltage power electronics. Avoid the primary-side capacitors, exposed mains-side circuitry, and any loom with melted insulation. Keep hands and tools away from the fan area during live bench testing, as the fan can start unexpectedly and the surrounding circuitry remains hazardous.
Capacitor banks can retain dangerous charge after disconnection. A laptop's internal battery also remains an energy source during board work, so charging faults and liquid damage belong with a qualified technician.
Your Next Steps and When to Call a Workshop
Deal with the problem in stages rather than repeatedly forcing the machine to start.
Tonight
Run the external checks first. Try a direct wall connection, reseat accessible cables, disconnect unnecessary peripherals, and record motherboard LEDs, fan movement, clicking, smells, and the exact point of failure. If instability keeps returning, stop using the computer, especially for gaming, rendering, or important file work.
This week
Check the warranty before opening anything. Write down the PSU make and model, the computer's specifications, and a short symptom timeline. Include whether the fault appears at startup, after warming up, or only under graphics load. That information helps separate a PSU fault from a GPU, RAM, motherboard, battery, or charging-circuit problem.
For symptoms that remain ambiguous, professional computer diagnostics can use controlled testing rather than repeated part replacement. A proper assessment may include a known-good PSU, multimeter measurements, a bench supply where appropriate, rail checks, and inspection of connectors or board damage.
Before the next upgrade
Estimate actual power demand with a meter or reliable component analysis instead of relying only on a nameplate. Choose a replacement with meaningful headroom for the intended graphics card and future expansion, but don't oversize it blindly. If a unit is still relatively new and shows partial symptoms, pursue warranty support before considering repair.
Call a workshop for surge damage, liquid spills, burn marks, melted cables, persistent clicking, or any unit still covered by warranty. In Sheffield, Steel City IT handles computer, laptop, and console diagnostics, including power, charging, and board-level faults, so bring the device and its original adapter or PSU details rather than guessing at a replacement.
If your PC, laptop, or console is showing random shutdowns, charging faults, clicking, burning smells, or no power, visit Steel City IT for structured diagnostics and repair advice. Bring the device, power adapter or PSU model, and a clear description of when the failure occurs so the workshop can test the fault instead of replacing parts by guesswork.
