Desktop PC Won’t Power On? A Step-by-Step Guide By ErlenTek
A desktop computer that “won’t turn on” can stop at several different points. A dead outlet, a failing power supply unit (PSU), a disconnected case switch, a shorted USB device, bad memory, a graphics problem, a motherboard fault, or a Windows startup issue can all have similar symptoms, but the repair path is different for each.
That is why the first job is not replacing parts. The first job is identifying which stage of startup is failing.
This article covers most desktop PCs, including custom-built systems and common tower-style OEM computers. All-in-one systems, very small form factor desktops, and proprietary OEM designs may use layouts or power configurations that don't follow normal ATX (the most common desktop motherboard and power-supply standard), but troubleshooting steps are similar and much of the article will still be applicable. One thing to note though is that when a model-specific manual says something different, the manual typically wins.
Safety disclaimer: Internal PC work carries electrical, static-discharge, component-damage, and data-loss risk. Unplug AC power before opening the case or changing components. Don't open a power supply. Never measure household voltage as a DIY step unless trained or qualified to do so. Stop immediately for smoke, a burning smell, liquid exposure, visible charring, repeated arcing, damaged cords, or a power supply that keeps shutting itself off to protect the system. This guide is general educational information and is not a substitute for a qualified repair evaluation.
First, identify the actual symptom
Before the guide starts, it helps to sort the failure into the right category. “Won’t turn on” is a description that covers several different failure points.
| Observed symptom | What it often means | Where the sequence goes next |
|---|---|---|
| No lights, fans, sounds, or reaction | The system may not be receiving standby power, the low-power feed that stays available while the PC is “off,” or the start signal may not be reaching the power supply | All steps are applicable |
| A motherboard LED is lit, but pressing the case button does nothing | Standby power may be present, but the start signal, power delivery, board, or a connected component may be failing | Look at Step 2, and continue from step 4 |
| Fans twitch or the system starts for less than a second, then stops | A short circuit, wiring issue, memory problem, board fault, incompatible part, or the power supply shutting itself off to protect the system is possible | Step 2, and continue from step 4 |
| Fans and lights run, but the monitor stays blank | This is often a no-display problem or a failure before POST (Power-On Self-Test), the first hardware check before Windows loads | After step 2, start at step 10 and continue until problem resolved |
| The system reaches a logo or POST screen but does not load Windows | Windows, storage, or boot-configuration trouble is more likely than a no-power fault | This guide is not applicable to current issues - read only if interested |
| A diagnostic LED or beep code appears | The motherboard or OEM system is reporting a startup checkpoint or fault code | Step 2, and the step 5 is the best to start at. |
POST is the early hardware check that happens before Windows begins loading. A computer can have power and still fail before POST. It can also complete POST and then fail later during Windows startup.
A motherboard standby LED is useful evidence, but it does not prove that the power supply is healthy or that all power rails are intact. It only suggests that one low-voltage path is present.
The complete diagnostic flow
The following sequence moves from safe external checks to progressively deeper isolation. It is one straight line of steps, some steps may be skipped depending upon the symptoms above.
Step 1: Protect people, hardware, and data first
Before the case comes open or the power button gets pressed ten more times, pause and assess the situation.
Unplug the PC from AC power before opening the case or changing components. If the power supply has a rear rocker switch, move it to O before unplugging the cord. On most supplies:
- I = on
- O = off
Never open the PSU enclosure. A PSU can hold dangerous electrical energy even after the computer is unplugged.
Stop immediately if any of the following is present:
- Smoke or visible vapor
- A burning smell
- Repeated clicking, arcing, or sparking
- Liquid exposure
- Visible charring or melted plastic
- Bulging or leaking components
- A hot, cut, frayed, or damaged power cord
- A surge or lightning event followed by unusual behavior
- A power supply that repeatedly shuts itself off
That last point matters. A surge can damage the power supply and also damage the motherboard, graphics card, storage, monitor, or other connected hardware. Replacing one damaged part does not mean the rest of the system is healthy.
If important files exist only on that computer, data preservation may change the order of work. Repeated startup attempts, operating-system repair attempts, and unnecessary drive handling can make a bad situation worse. Sometimes the drive is fine and the board is dead. Sometimes the drive is the part that needs the most protection. It helps to know which problem is being solved before more stress is added.
Step 2: Confirm and document the symptom
Press the power button once and watch closely. The exact behavior matters because several different faults look almost identical at first glance.
Record what happens:
- No response at all: no lights, no fans, no sound
- A motherboard standby LED is already on before the button press
- Pressing the case button does nothing
- Fans twitch briefly and stop
- The system starts for less than a second and shuts down
- Fans spin and LEDs stay on
- A click comes from the PSU
- Beeps occur
- Diagnostic LEDs stop at a labeled stage
- The system restarts over and over
- The system powers on but shows no video
- The system reaches a logo or POST screen but does not load Windows
These symptoms break down into a few main groups:
- No AC power or no standby power
- Standby power present, but no start signal or startup command
- Brief start, then shutdown
- Powered, but no display
- POST completes, but Windows does not load
- Intermittent startup
Also note what changed recently. That history does not prove the cause, but it often helps narrow the path:
- A storm, surge, or power outage
- Moving the computer
- Cleaning or dust removal
- Installing memory, a graphics card, a drive, a fan, or an RGB controller
- A BIOS/UEFI update (an update to the board’s built-in startup software and base settings)
- A failed overclock or memory-setting change
- A new USB device
- A cable being moved or replaced
- An unusual smell, noise, or heat event
Good notes save time. They also help keep the process honest. A repair goes sideways fast when three different changes are made and no one remembers which one happened first.
Step 3: Check the external power path in order
External power checks come before internal work because they are usually safer, faster, and frequently solve the problem without touching the hardware.
3.1 Verify building power
Plug a lamp or another known-good device into the same outlet. Start with the simplest question first: is power actually available at that spot?
A customer once called while trying to diagnose a computer. When asked to check whether the computer was plugged in, the customer said, “I would check that, but I don’t have a flashlight.” When asked why a flashlight was needed, the customer replied, “Because the power is out.”
That story has survived for a reason. Troubleshooting should always start with the basics.
3.2 Check the wall outlet
If the known-good device does not work:
- Try another known-good outlet.
- Check whether a wall switch controls the outlet.
- Check whether a breaker or GFCI outlet (if equipped) has tripped.
3.3 Temporarily bypass strips, UPS units, and extension cords
Power strips, surge protectors, UPS units, and extension cords fail more often than many people expect (a UPS, or uninterruptible power supply is a type of battery-backup box often used to prevent loss of power). As with all electronics - surge protectors and UPSes have a life expectancy and anything can fail.
For testing, connect the desktop directly to a known-good wall outlet with an undamaged, correctly rated cord (this is a temporary diagnostic bypass, not a recommendation for permanent use).
3.4 Inspect and reseat the IEC power cord
The removable desktop power cord is usually an IEC cord, meaning the standard appliance-style computer power cable used on many monitors, desktops, and other electronics.
With the PC disconnected:
- Inspect the cord for cuts, crushed insulation, exposed conductors, heat damage, or bent contacts.
- Confirm that the cord is fully seated in the PSU and in the wall outlet or tested power source.
- Substitute a known-good compatible cord when easily available or questionable.
- Replace a damaged cord rather than testing it further.
3.5 Check the PSU rear switch
Confirm that the PSU rocker switch is set to I.
(Some systems do not have a rear switch. That is normal on certain OEM and small-form-factor designs.)
3.6 Check a manual voltage selector only if one exists
Some older or specialized PSUs have a physical 115/230 V selector. Most current desktop power supplies handle both voltages automatically and do not have one.
If a manual selector is present:
- Confirm the correct setting for the region.
- Unplug AC power before changing it.
- Never move the selector while the cord is connected.
- Never change the setting as a guess-and-check experiment.
For a standard North American installation, the correct position is commonly in the 115–120 V range, but the PSU label and local electrical standard take priority.
Step 4: Perform a safe power discharge
A safe power discharge can sometimes clear a stuck power-management state, but it doesn't always prove that the hardware is good (and it does not fix a failed PSU, motherboard, or shorted component).
Use this generic desktop sequence:
- Move the PSU switch to O (if present).
- Unplug the AC power cord from the PSU.
- Disconnect external peripherals such as USB drives, hubs, printers, docks, and external storage.
- Hold the case power button for about 15-30 seconds.
- Reconnect only the AC power cord.
- Set the PSU switch to I, if present.
- Press the case power button one time and observe the result.
Most generic desktop procedures use roughly that 15-30 second hold. Some OEM systems call for longer. Dell’s current guidance for certain desktop hard-reset procedures is about 30–35 seconds. (The exact documentation for the specific model takes priority here.)
If the system starts after this step, record the result but do not treat it as a finished diagnosis. A machine that responds only after a discharge may still have a power-management fault, a PSU problem, a motherboard issue, or an intermittent hardware condition. We recommend writing down the time and date somewhere safe in order to diagnose later if a problem persists and repeats itself.
Step 5: Check model-specific diagnostics before generic testing
Before the case comes open or parts start coming out, check the manufacturer’s documentation or website for built-in diagnostics.
Possible indicators include:
- Dell PSU BIST, or Built-In Self-Test
- Motherboard standby LEDs
- CPU, DRAM, VGA, or BOOT diagnostic LEDs
- POST code displays
- Beep codes from an onboard speaker or buzzer
- Onboard power or reset buttons
- Diagnostic buttons on an OEM chassis
Dell’s PSU BIST is Dell-specific and applies only to supported Dell systems. It does not transfer to a generic ATX supply.
On supported Dell models with a BIST button, Dell’s current guidance is that a solid, non-flickering LED together with the fan spinning normally indicates that the PSU is delivering power. A dark LED or abnormal fan behavior counts as a failure. On supported Dell models without a BIST button, Dell instructs the owner to unplug the cord, wait about 15–20 seconds, reconnect it, and watch for the PSU LED to light for roughly three seconds. That brief light indicates the supply is delivering power.
Dell's system still has its limits though. It only helps answer whether the power supply is delivering power on that supported Dell model. It does not prove that the motherboard, graphics card, storage, or the rest of the system is healthy.
Many OEM HPs, Lenovos, and Acers have similar lights or ways to indicate common faults. We only highlight Dell's here due to its exact application - the power supply.
If lights are blinking, codes displayed on board, etc - checking the computer or motherboard's manufacturer site is a great place to look.
A few other limits are worth keeping in mind:
- A motherboard standby LED does not prove that the PSU can deliver stable operating power.
- Outside a supported model-specific self-test, a PSU fan not spinning is not automatically a failure. Many new PSUs use zero-RPM fan mode, meaning the fan stays off at light load to reduce noise.
- Diagnostic LED labels and beep patterns vary by motherboard model.
- Onboard power buttons are not present on every board.
The exact motherboard manual or OEM documentation should be consulted before interpreting any LED, code, button, or self-test result.
Step 6: Open the case only when safe and appropriate
If the external checks and model-specific diagnostics do not identify the problem, internal inspection may be appropriate.
Before opening the case:
- Shut the PSU switch off.
- Disconnect the AC power cord.
- Press the case power button once to discharge the remaining low-voltage standby energy.
- Place the system on a stable, clean surface.
- Use reasonable ESD precautions. ESD means electrostatic discharge, the small static spark that can damage electronics. Working on a non-carpeted surface and touching the metal chassis before handling parts is a good start.
- Keep screws, tools, and loose metal objects away from the motherboard.
Inspect first. Look before disconnecting everything. The goal is to spot obvious trouble without creating new trouble. If not sure, never assume the computer hasn't been tampered with.
Check for:
- Loose connectors
- Dust blocking fans or vents
- Liquid residue
- Scorch marks
- Melted plastic
- Damaged sockets or pins
- A graphics card partly out of its slot
- Recently disturbed memory
- Cables trapped against fans
- Extra motherboard standoffs or loose screws
- A disconnected front-panel power-switch lead
- Damaged USB headers or bent pins
Important connectors to identify
24-pin ATX connector: the main motherboard power connector. It carries several voltages and control signals to the board.
CPU EPS connector: usually a 4-pin or 8-pin connector near the CPU socket. EPS is the dedicated processor power connector. It is not the same nor compatible with a PCIe connector, which is used for graphics-card power even though the plastic shapes can look similar. PCIe stands for Peripheral Component Interconnect Express, the expansion connection used by graphics cards and other add-in devices.
GPU auxiliary power: a discrete graphics card may require one or more PCIe power connectors from the PSU. These must be connected.
SATA or Molex power: SATA power is the flat power connector commonly used for slightly older drives and accessories. Molex is the even older 4-pin peripheral power connector still found on some older fans, hubs, and accessories.
Front-panel PWR_SW connector: the small two-wire lead from the case power button to the motherboard header.
Reseat connectors only with AC power disconnected. Confirm connector shape, latch position, and labeling against the correct motherboard or PSU documentation if unsure. Connector names and layouts vary by board and power supply, so the documentation is better than a visual guess.
Modular PSU cables need special caution. They are not pinned the same way across all brands or models. Never mix modular cables between different PSU brands or models unless the manufacturer explicitly confirms compatibility. A cable that fits physically can still be wired differently, and can damage the PSU, motherboard, drives, or graphics card.
Step 7: Test the case power switch and front-panel wiring
A case power button is usually a momentary switch. That means it makes contact briefly and tells the motherboard to request startup. It does not carry the computer’s full operating power by itself.
Start with the exact motherboard manual or clear board markings. Locate the front-panel header and identify the two pins labeled:
PWR_SWPWRSWPOWER SWPW- or a similar model-specific designation
Pin layouts are not universal. One motherboard diagram is not a safe substitute for another.
Safer options, in order, are:
- Use an onboard power button if the motherboard has one.
- Temporarily connect a known-good momentary switch to the positively identified
PWR_SWpins. - Consider a brief bridge test only if the exact
PWR_SWpins are 100 percent identified from the board manual or obvious board labeling.
If the correctly identified bridge is performed, it should be momentary, controlled, and limited strictly to the documented PWR_SW pins. Do not bridge unknown pins. Do not drag a screwdriver across the header looking for the right pair. A successful start does not always prove that the motherboard or PSU is fully healthy, but unless other factors are present, a bad power switch is likely at this point.
If the system starts from the onboard button, a known-good momentary switch, or a properly identified PWR_SW trigger, the case switch or front-panel wiring becomes the primary suspect. If nothing changes, continue with minimum-configuration isolation.
Step 8: Evaluate the PSU safely
The PSU converts household AC power into the lower-voltage DC power used by the motherboard, CPU, drives, fans, and graphics card. It also provides standby power before startup and responds to the motherboard’s power-on signal.
A PSU can fail in several ways:
- No output at all
- Standby power present but no full startup
- Startup followed by immediate shutdown
- Unstable voltage under load
- Failure after a surge
- Intermittent operation
- Adequate idle readings but failure during real system demand
Use the following order.
8.1 Use a built-in self-test when the model supports it
For Dell systems with PSU BIST, follow the exact Dell procedure for the model. That procedure is only for supported Dell systems and is not a generic ATX test.
8.2 Use a properly rated known-good compatible PSU
A replacement PSU must match the system’s:
- Form factor
- Motherboard connectors
- 24-pin and CPU EPS requirements
- GPU auxiliary power requirements
- Output capacity
- OEM compatibility requirements
For proprietary OEM systems, a standard retail ATX PSU may not be electrically or mechanically compatible even if the wattage looks sufficient.
A known-good compatible PSU is one of the best diagnostic controls in this part of the process. Even then, the result still needs interpretation. A supply that runs a bare-minimum setup may still fail under the real load of a graphics card and/or multiple drives.
8.3 Paperclip Test: A VERY Optional PSU Check
The paperclip test can help determine whether a desktop PC power supply (PSU) will turn on. It does not confirm that the PSU is safe or functioning correctly under load and if done incorrectly can cause more problems than it solves and is not applicable to proprietary PSU/connectors simply because they resemble an ATX system. Proceed with caution here:
⚠️ Safety warnings
- Turn the PSU off and unplug it from the wall before connecting anything.
- Never open the PSU housing—dangerous voltages can remain inside.
- Use the test only with a 24-pin ATX connector and follow your PSU manufacturer's documentation.
- Keep the paperclip insulated except for the portions making contact, and make sure it cannot touch other pins or metal parts.
- Stop immediately if you notice sparks, smoke, unusual noise, or a burning smell.
Basic procedure
- Disconnect the PSU from the PC and switch it off.
- On the 24-pin ATX connector, identify the PS_ON pin 16 (normally green) and ground pin 15 (normally black).
- Bridge those two pins with an appropriately insulated jumper.
- Ensure the jumper is secure and clear of other contacts.
- Plug the PSU in and switch it on. If the PSU starts, the test indicates that its basic startup circuit is responding.
Some modern PSUs have zero-RPM/fan-stop modes, so the fan may not spin even when the PSU has started. Consult the manufacturer's documentation rather than relying solely on fan movement.
8.4 Use a PSU tester or meter within safe limits
A PSU tester can identify certain missing or abnormal outputs, but it has limits. It may not detect:
- Voltage instability under load
- Excessive ripple, meaning poor voltage smoothness or noisy output
- Protection behavior
- Intermittent failure
- A failure that appears only when the unit is warm
- A fault in a cable or connected component
- Full-system compatibility under real operating conditions
A standby LED, a spinning fan, a paperclip test, a simple PSU tester, or an idle voltage check can all provide clues. None of them proves that the supply is stable under real load.
A quick note on terminology: a power supply has multiple rails, meaning separate voltage outputs such as 12-volt, 5-volt, and 3.3-volt lines used by different parts of the system. Seeing one rail present does not guarantee that the others are healthy under load.
Detailed rail testing should be performed only with appropriate equipment, safe procedure, and a clear understanding of the platform being tested. Never open the PSU, never probe live mains without proper procedure and expertise, and never reuse modular cables from another PSU.
Step 9: Reduce the system to a minimum configuration
Record cable locations and take photos before disconnecting anything if not accustomed to computer repair, and always double-check that the power cord is unplugged.
Minimum configuration means stripping the system down to the smallest set of parts required for startup. The goal is to remove any extra variables.
For a standard ATX desktop, the minimum configuration usually includes:
- Motherboard
- CPU
- CPU cooler
- PSU (Power Supply)
- One compatible RAM module in the correct slot (Often slot A2 - but can vary by model)
- An appropriate graphics path (*always use onboard here if available and processor is compatible)
- A known method of starting the system, such as the correct
PWR_SWconnection or an onboard power button
*The graphics path varies by platform. If the installed processor and motherboard support integrated graphics, the motherboard video output may be usable. If not, a discrete GPU may still be required even during minimum-configuration testing.
Disconnect nonessential devices:
- USB devices
- External hubs
- Extra internal drives
- Optical drives
- Front USB headers
- Front audio headers
- Extra expansion cards
- RGB controllers
- Fan hubs
- Additional case fans
- Capture cards
- Wi-Fi or Bluetooth expansion cards
- Internal accessories and lighting controllers
Minimum configuration does not automatically mean moving the board out of the case. In many situations, careful in-case isolation is the safer first step. Full bench testing, meaning running the essential parts on a safe work surface outside the case, belongs later when the evidence points in that direction.
Step 10: Check graphics and display
A no-display problem is not the same as a no-power problem, and keeping those two ideas separate prevents a lot of wasted effort. If all fans are spinning, all system lights are on, and any indication LEDs are correct for normal operation, this is often the place to start.
Check the display path in this order:
- Confirm that the monitor is powered on.
- Confirm that the monitor input is set to the correct HDMI, DisplayPort, DVI, or other source.
- Reseat the display cable at both ends.
- Test a known-good display cable.
- Test a known-good monitor or television when available.
- Confirm whether the cable is connected to the motherboard video output or the discrete GPU output.
- If a discrete GPU is installed, confirm that it is fully seated in the PCIe slot.
- Confirm that all required GPU auxiliary power connectors are plugged in (at both ends if cables are modular).
- If the platform supports integrated graphics, remove or bypass the discrete GPU and test the motherboard video output. (Confirm processor and motherboard support for integrated graphics before expecting motherboard video ports to work.)
Integrated graphics are the display features built into the processor or, on some older systems, the chipset. Not every CPU has them. A motherboard video port does not become active merely because the port exists on the back of the case.
A system can also power on and stay blank because of memory faults, CPU compatibility issues, firmware settings, GPU failure, monitor failure, or motherboard failure.
Step 11: Test memory methodically
RAM (random-access memory) is the working memory that the processor uses during startup and normal operation. Memory faults can cause no POST, restart loops, no display, beep codes, DRAM warning lights, crashes shortly after startup, and intermittent behavior. Random access memory causes some of the most random errors.
Memory testing is applicable in many situations: Bad memory can stop a system before POST just as easily as it can stop a system from showing a picture.
With AC power removed:
- Remove all but one memory module.
- Install the remaining module in the motherboard’s recommended single-module slot.
- Use the motherboard manual for the correct slot. It is often, but not always, labeled
"A2". - Confirm that the module is fully seated and that the latches engage as designed.
- Reconnect AC power and test.
- If the result is unchanged, remove AC power again.
- Test another individual module in the slot recommended by the manufacturer.
- If appropriate, test the original module in another supported slot as sometimes the slot itself is bad.
Testing one module at a time isolates variables. It is one of the most useful habits in startup troubleshooting. Reseating a part can help confirm placement, but it does not prove that the part works. A memory module can look perfect and still be defective, incompatible, unstable, or affected by a bad slot or memory controller.
Document or remember any results. Fast repeated swapping can create confusion when in a hurry.
Step 12: Reset CMOS/UEFI
BIOS/UEFI and Clearing CMOS
BIOS and UEFI are the motherboard's firmware - the built-in software that initializes and controls hardware and begins the boot process.
CMOS is commonly used to refer to the motherboard's stored firmware configuration settings and the process of resetting them. On modern systems, the settings may use different memory technologies, but “Clear CMOS” means resetting those firmware settings to their defaults.
The coin-cell battery primarily maintains the real-time clock and, depending on the motherboard design, helps preserve firmware settings when the system is powered off.
These are two different procedures:
- Clearing CMOS resets stored BIOS/UEFI settings to their defaults.
- Replacing the coin-cell battery addresses problems such as a lost system clock or settings that are not being retained correctly.
A weak battery is not normally the reason a desktop is completely dead.
When Should You Clear CMOS?
Clear CMOS when there is a sensible reason, such as:
- A failed overclock
- An incompatible memory setting
- A failed BIOS/UEFI configuration change
- A firmware configuration that prevents POST
- A hardware change followed by no display or unexpected startup behavior
⚠️ Important: Encryption and BitLocker
Do not clear CMOS casually on a system using BitLocker or other drive encryption unless known access to the Bitlocker key is established. Resetting firmware settings can change settings related to the TPM, Secure Boot, boot mode, or other security configuration. Windows may then detect a change in the system's trusted boot environment and request a BitLocker recovery key. This is why we put this as a later step - it used to be near the beginning of standard procedures.
Before clearing CMOS:
- Make sure you have your BitLocker recovery key available.
- If the computer uses another form of full-disk encryption, make sure you have the necessary recovery credentials.
- If you are unsure whether encryption is enabled and important data is solely on the device - it is not recommended to proceed without obtaining and verifying key.
Do not proceed without a recovery key if you cannot afford to be locked out of the encrypted drive.
How to Clear CMOS
Follow the motherboard manufacturer's documented procedure:
- Disconnect AC power.
- Turn the PSU switch to O, if present.
- Press the case power button once.
- Use the motherboard's documented Clear CMOS pins or button.
- If the manual permits battery removal, remove the coin-cell battery according to the documented procedure.
- Before proceeding, make sure you have any required BitLocker or other drive-encryption recovery credentials available.
- Restore the jumper, button state, or battery.
- Reconnect AC power.
- Test the system and reconfigure any required firmware settings.
Clearing CMOS can remove custom boot settings, fan profiles, memory profiles such as XMP/EXPO, and overclocking values. It does not fix a dead PSU, damaged motherboard, short circuit, or failed processor.
Step 13: Isolate board, CPU, and short-circuit faults
If the system still does not respond after external checks, correct cabling, graphics and display checks, safe PSU evaluation, minimum configuration, and memory testing, the remaining possibilities include:
- motherboard in need of repair
- Shorted motherboard standoff or loose screw
- Shorted front-panel or USB header
- Failed graphics card
- Failed drive or expansion card that wasn't isolated
- CPU power-delivery fault
- Processor failure
- Socket or pin damage
- Firmware incompatibility
- Proprietary OEM board or PSU fault
Processor failure is possible, but it should not be the default conclusion (although we've seen more Ryzen failures in the last couple years than we'd like to see). In practice, power-path faults, shorts, wiring issues, memory problems, graphics issues, and motherboard faults often appear higher on the list.
A qualified repair evaluation may include bench isolation outside the case using:
- A nonconductive work surface
- Motherboard
- CPU and cooler
- One compatible memory module
- Compatible PSU
- Required graphics path
- Correct power and start connections
- Board diagnostic indicators
- Known-good compatible components
This is bench testing in the stricter sense: running only the essential parts on a safe work surface outside the case to rule out shorts from standoffs, screws, headers, expansion cards, or proprietary case hardware. It can also help separate a bad board from a bad case or accessory.
Done well, it is useful. Done randomly, it can become an expensive nightmare. Each substitution should have a reason, confirmed compatibility, and a clear interpretation.
Step 14: Know when to stop
Stop DIY work and seek a repair evaluation when any of the following applies:
- Smoke, a burning smell, liquid exposure, or visible charring
- A PSU repeatedly trips or clicks off even with minimal configuration
- A surge or lightning event preceded the failure (many components may be affected)
- The power cord is hot or melted
- No known-good compatible PSU is available
- The system is proprietary or model-specific
- Data on the storage device is at risk
- The motherboard shows visible damage
- The CPU socket or board pins appear damaged
- The system remains dead after a proper minimum-configuration test
- The repair requires live electrical measurements
- The procedure is uncomfortable or the correct pins cannot be positively identified
A PSU that has retired without submitting a support ticket is inconvenient. A PSU that damages other components is a larger problem, and unfortunately - power supplies do not fill out exit paperwork on the way out. That is why a PSU replacement should be followed by careful testing of the motherboard, graphics card, drives, and other connected hardware before the repair is considered complete.
What not to do
A short list helps here because the same mistakes come up often:
- Do not open the PSU enclosure.
- Do not measure household mains voltage as a DIY step unless qualified.
- Do not keep power-cycling a system that shows smoke, liquid damage, or a burning smell.
- Do not bridge unknown motherboard pins.
- Do not assume every two-pin header is a power switch.
- Do not mix modular cables between PSU brands or models unless the manufacturer explicitly confirms compatibility.
- Do not assume a spinning PSU fan proves the PSU is healthy.
- Do not assume a silent PSU fan proves failure.
- Do not assume a motherboard standby LED proves the PSU is healthy.
- Do not treat a paperclip test as a complete PSU diagnosis.
- Do not assume reseated RAM or a reseated GPU is therefore functional.
- Do not start clearing CMOS, updating BIOS, or attempting operating-system repair when important files may be at risk.
- Do not declare the repair final after replacing only the obvious failed part following a surge.
Frequently asked questions
Can a dead CMOS battery stop a desktop from powering on?
Not very often. A weak coin-cell battery is more likely to cause clock loss or lost settings after power has been removed. A desktop with no response at all is more likely to have trouble with a power path, the PSU, the power switch, the motherboard, or there is a shorted component.
Does a motherboard standby LED prove that the PSU works?
No. It suggests that standby power is present. It does not prove that the PSU can deliver stable operating power on all rails under load or that all rails are healthy.
Should a PSU fan spin as soon as the computer is plugged in?
Not necessarily. Many modern PSUs use zero-RPM fan mode and stay silent at light load. Fan behavior by itself is not a reliable diagnosis.
Is no display the same as no power?
No. If fans, LEDs, or other activity are present, the system may have power but fail during POST or during the graphics-output stage. Those are different problems, and telling them apart early saves a lot of wasted time.
Can Windows cause a completely dead desktop?
Under normal conditions, no. Windows does not stop the PSU from responding to the physical power button before POST. Once the system reaches a logo screen, firmware screen, automatic repair screen, or boot error, the problem moves away from no-power and toward boot troubleshooting.
Is a standard ATX power supply safe for an OEM desktop?
Not always. Some OEM systems use proprietary connectors, wiring, form factors, or power-control designs. Compatibility should be confirmed before substitution.
ErlenTek’s role in this type of problem
ErlenTek is a locally and family-owned computer repair and IT services company based in Auburn, Washington - serving Auburn, Kent, Federal Way, Covington, Maple Valley, Enumclaw, and the surrounding South King County area.
Relevant services include desktop and gaming PC repair, advanced computer diagnostics, hardware testing, power-supply and component troubleshooting, data recovery, electronics and board-level repair. Helpful starting points on the site include computer diagnostics and troubleshooting service and computer repair services.
When important files are involved, data preservation can take priority over repeated startup attempts. When a system has been through a surge, liquid exposure, or a suspected board fault, careful evaluation is usually safer than guessing with parts.