You hit the power button and nothing happens: no fans, no lights, no beep, no display. Or maybe the fans spin up and the RGB comes alive, but the monitor stays black forever. Both scenarios fall under the same frustrating umbrella of “my PC won’t turn on,” and both have a defined, repeatable diagnostic path that doesn’t require guessing or replacing parts at random. This is one of the highest-volume PC troubleshooting searches on the web, and for good reason: it can be caused by more than a dozen different failure points, from a $2 case switch cable to a dead power supply.
This guide walks through a structured, 12-step process for isolating exactly where a “won’t turn on” failure lives, using the same techniques PC builders and repair technicians rely on: the paperclip PSU self-test, motherboard debug LEDs and Q-codes, breadboarding outside the case, and a methodical process of elimination. By the end, you’ll also have a small interactive diagnostic script that walks you through the same decision tree automatically, so you can re-run it any time a system acts up again.
None of the steps below require specialized repair equipment. A screwdriver, a paperclip, and an inexpensive multimeter cover the vast majority of cases, and the order matters: working through cheap, quick checks before expensive ones (a loose cable before a motherboard replacement) is how experienced builders avoid throwing money at parts that were never actually broken. Expect the full diagnostic pass to take 30-60 minutes depending on how quickly you find the culprit.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
“Won’t Turn On” vs “Won’t POST” vs “Won’t Boot”: Know Which Problem You Have
These three phrases get used interchangeably, but they point to completely different parts of the system, and mixing them up wastes time chasing the wrong component.
- Won’t turn on: No fans spin, no lights illuminate anywhere in the case, nothing happens at all when you press the power button. This points to the power supply, the wall outlet/power strip, the front panel power switch, or the motherboard’s power circuitry.
- Won’t POST (Power-On Self-Test): Fans spin, lights come on, but there’s no video output and no beep confirming a successful self-test. This points to RAM, GPU, CPU seating, or a motherboard fault — the system has power but can’t complete its startup checks.
- Won’t boot: The system POSTs successfully (you may see a manufacturer logo or BIOS splash screen) but then fails to load an operating system. This is a storage, boot order, or Windows/OS-level problem, not a hardware power issue, and is outside the scope of this guide — though it’s worth knowing this category has gotten noisier lately: Windows 11 update KB5083769, released April 14, 2026, triggered repeated BSOD boot loops on an estimated “thousands” of HP and Dell PCs according to Notebookcheck, with PCWorld separately clocking the failure rate at roughly 5-10% of updated systems by that same month, and at least one Windows 11 Pro 24H2 user reported a kernel_security_check_failure BSOD on every boot afterward that required a full WinRE repair and reinstall to resolve. Enterprise fleets running customized Windows 11 24H2 builds newer than 26100.1742 or 25H2 GA build 26200.6584+ have reported a related but distinct failure mode as of August 2026, where the OS boots but the endpoint security agent itself won’t come up, per Broadcom.
The steps below are organized to help you quickly figure out which of the first two categories you’re actually dealing with, since that determines which half of the system to investigate first. It’s worth noting these categories aren’t always perfectly clean in practice — a marginal PSU rail can sometimes produce symptoms that look like a POST failure (fans spin, no display) purely because the motherboard isn’t receiving stable enough power to complete its self-test, even though the underlying fault traces back to the power stage. That’s exactly why Steps 4 through 6 (testing the PSU directly) come before the motherboard-focused steps that follow — ruling power out first prevents you from chasing a RAM or GPU seating issue that was never actually the problem.
Prerequisites: Tools and Safety Precautions
| Item | Purpose |
|---|---|
| Phillips screwdriver (magnetic tip) | Opens the case, removes and reseats components |
| Paperclip or short piece of wire | Performs the PSU self-test by bridging the 24-pin connector |
| Digital multimeter | Measures actual voltages on PSU rails beyond a simple on/off test |
| Anti-static wrist strap | Prevents electrostatic discharge while handling the motherboard, RAM, and GPU |
| Known-good spare parts (if available) | PSU, RAM stick, or GPU from another system for swap-testing |
| A different power outlet and cable | Rules out wall power or a faulty PSU cable as the root cause |
| Flashlight or phone light | Reading tiny debug LED labels and Q-code displays inside a dark case |
| USB drive (empty, FAT32-formatted) | Required for USB BIOS Flashback / Q-Flash Plus recovery if the BIOS itself is corrupted |
Before you touch anything inside the case, unplug the system from the wall completely and hold the power button for 5-10 seconds to drain residual capacitance from the motherboard. Put on your anti-static strap and clip it to an unpainted metal part of the case chassis before handling the motherboard, RAM, or GPU. If you don’t own a multimeter or spare parts to swap-test with, most of this guide still works using just the paperclip test, the debug LED readout, and careful reseating — those three checks alone resolve a large share of real-world cases without needing to borrow or buy anything.
Step 1-3: Isolate the Problem to Power, POST, or Boot
Step 1: Check the absolute basics first. Confirm the wall outlet works by plugging in another device, check that the PSU’s rear power switch is set to “I” (on), and make sure the power cable is fully seated at both ends. It sounds obvious, but a loose IEC power cable or a power strip that got bumped off is one of the single most common causes of a “dead” PC.
Step 2: Listen and look for any sign of life. Press the power button and watch closely for even a brief flicker of a fan blade, a chassis LED strobing for a fraction of a second, or a click from the PSU. Any sign of life at all — even momentary — usually rules out a completely dead PSU and points instead toward a short circuit that’s tripping the PSU’s protection circuitry, or a front-panel connector issue.
Step 3: Disconnect everything except the essentials. Unplug all USB peripherals, extra drives, case fans beyond the CPU cooler, and any add-in cards besides the GPU. A failing USB device or a shorted case fan header can prevent power-on entirely, and eliminating variables early saves time later.
Step 4-6: Test the Power Supply Directly
Step 4: Run the paperclip PSU self-test. With the PSU disconnected from the motherboard, locate the green PS_ON# wire on the main 24-pin connector (pin 16 in Intel’s ATX 3.0 design guide) and bridge it to any black ground/COM pin using a paperclip or short wire. If the PSU’s fan spins up, the unit is capable of powering on — though this test only confirms the rails are present, not that they hold correct voltage under load.
# ATX 24-pin PS_ON# self-test reference (do this with the PSU
# fully disconnected from the motherboard, cable unplugged from wall
# before making the bridge, then plug in AC power to test)
#
# Pin 16: PS_ON# (green wire) -- bridge this...
# Pin 15/17/18/etc: COM (ground) -- ...to any black wire
#
# Result: PSU fan spins up -> PSU can power on (basic pass)
# Result: nothing happens -> PSU is likely dead or tripped
# Result: fan spins then stops -> PSU protection circuit is trippingStep 5: Verify actual voltages with a multimeter. The paperclip test only proves the PSU turns on, not that it delivers correct voltage. With the PSU running from the paperclip bridge, probe the 24-pin connector’s pins with a multimeter set to DC volts. Current ATX12V/ATX 3.x specifications, documented in the ATX standard itself and detailed further in Intel’s motherboard support documentation, call for +3.3V, +5V, +12V, -12V, and +5V standby rails, with the gray PWR_OK signal rising to a logic-high level (roughly +5V) only once all rails have stabilized.
| Rail / Signal | Wire Color | Expected Voltage | Acceptable Tolerance |
|---|---|---|---|
| +3.3V | Orange | 3.3V | ±5% |
| +5V | Red | 5.0V | ±5% |
| +12V | Yellow | 12.0V | ±5% |
| -12V | Blue | -12.0V | ±10% |
| +5VSB (standby) | Purple | 5.0V | ±5% |
| PS_ON# | Green | 0V when active (pulled low) | Active-low control signal |
| PWR_OK | Gray | ~5V once rails stabilize | Asserted 100-500ms after PS_ON# |
Step 6: Swap in a known-good PSU if available. If you have access to a spare power supply, even briefly connecting it in place of the suspect unit is the most reliable way to confirm or rule out a PSU fault. If the system springs to life with a different PSU, you’ve found your culprit without needing to interpret a single voltage reading.
Step 7-9: Read the Motherboard’s Diagnostic Signals
Step 7: Check for debug LEDs or a Q-code display. Most current motherboards from ASUS, MSI, Gigabyte, and ASRock report POST failures through onboard status indicators rather than relying solely on beep codes. These typically appear as either a two-digit hexadecimal “Q-code” style display, or a set of four labeled LEDs marking CPU, DRAM, VGA, and BOOT stages — often branded as EZ Debug LEDs or similar. Whichever LED stays lit (or which Q-code appears) tells you which stage of POST is failing.
Step 8: Interpret basic beep codes if your board has a speaker. Classic AMI/Award-style beep codes are still supported at the firmware level on most boards, but many current motherboards ship without an onboard piezo speaker by default, so you may need to plug in a spare case speaker to hear them at all.
| Symptom | Common Meaning |
|---|---|
| One short beep | POST completed successfully — the beep itself is not an error |
| Two beeps | Minor POST error, often related to configuration or video initialization |
| Continuous/repeating beep | Serious hardware fault, frequently tied to power delivery, RAM, or overheating |
| No beep at all (speaker installed) | CPU or RAM may not be initializing — check seating and CPU power connector first |
| Debug LED stuck on “CPU” | CPU not detected/seated, or CPU power connector (EPS 4+4/8-pin) not connected |
| Debug LED stuck on “DRAM” | RAM not seated, incompatible, or in the wrong slots for dual-channel |
| Debug LED stuck on “VGA” | GPU not seated, missing PCIe power, or GPU fault |
| Debug LED stuck on “BOOT” | System POSTed successfully but can’t find a bootable drive (an OS-level issue, not power) |
Step 9: Check for critical hardware errors logged from a prior successful boot. If the system boots intermittently, Windows’ Event Viewer can reveal a pattern of unexpected shutdowns before the failure became permanent — though it’s worth ruling out a software cause first, since Microsoft has confirmed that a December 2025 update broke boot entirely on physical Windows 11 24H2/25H2 commercial devices, a bug it says its KB5077181 fix (shipped February 10, 2026) resolved for 100% of affected UNMOUNTABLE_BOOT_VOLUME cases; if a repair does prove necessary, running SFC from WinRE typically takes 15-20 minutes to fix corrupted boot files, per guidance documented June 6, 2026. This PowerShell query pulls unexpected power-loss events (Kernel-Power Event ID 41) from the system log, which often precede a PSU or motherboard fault going fully dark:
# PowerShell: find unexpected shutdown/power-loss events (run as Administrator)
Get-WinEvent -FilterHashtable @{LogName='System'; Id=41} -MaxEvents 20 |
Select-Object TimeCreated, Id, LevelDisplayName, Message |
Format-Table -AutoSizeStep 10-12: Reseat Components and Breadboard Test
Step 10: Reseat RAM, GPU, and all power connectors. Remove and firmly reinstall each RAM stick until the retention clips click fully closed, reseat the GPU until it’s flush in the PCIe slot with the latch engaged, and disconnect/reconnect the CPU’s EPS power connector (the separate 4+4 or 8-pin connector near the top of the board, distinct from the 24-pin main connector) along with all PCIe power cables to the GPU. A connector that looks seated but isn’t fully clicked in is one of the most common causes of a no-POST condition after any case work.
Step 11: Check the front panel connector and case standoffs. A power switch cable connected to the wrong pins, or reversed polarity on some boards, can prevent the power button from doing anything. Consult your motherboard manual for the exact front-panel header pinout, and while you’re in there, confirm the board isn’t shorting against a case standoff in the wrong position — a single stray standoff under the PCB is a classic cause of a board that won’t power on at all despite everything being connected correctly.
Step 12: Breadboard the system outside the case. Remove the motherboard entirely and rest it on its cardboard box or an anti-static bag on a non-conductive surface, connecting only the PSU, CPU, one stick of RAM, and the GPU (or use the CPU’s integrated graphics if available). Short the front-panel power pins directly with a screwdriver tip instead of using the case’s switch. If the system posts outside the case but not inside it, the case itself — most likely a short against the chassis — is the actual problem, not any individual component.
GPU Power Connectors: A Growing Cause of “No Display” on High-End Builds
High-wattage 2026 GPUs add a wrinkle that older troubleshooting guides don’t cover: the 12V-2×6 (formerly 12VHPWR) power connector used on cards like the RTX 5090, which draws up to 575W through a single connector rated for that load only when every pin is fully seated. A connector that’s inserted at an angle, not pushed in until it clicks, or bent slightly during a tight case build can cause anything from a card that won’t power on at all to intermittent shutdowns under load that look identical to a PSU fault. If you’re troubleshooting a “no display, fans spin” scenario on a system with a high-end GPU, unplug and firmly reseat this connector at both the PSU and GPU ends before assuming the fault lies elsewhere, and inspect the connector visually for any discoloration, which would indicate it was already running hot.
On Linux-based recovery or live-boot environments, you can confirm whether the GPU is even being detected at the PCIe level (separate from whether it’s receiving adequate power) with a quick command:
$ lspci -vvv | grep -A 2 -i "VGA\|3D controller"
01:00.0 VGA compatible controller: NVIDIA Corporation Device 2c02 (rev a1)
Subsystem: NVIDIA Corporation Device 2c02
Flags: bus master, fast devsel, latency 0, IRQ 145If the GPU doesn’t appear in this output at all, the card isn’t being detected at the hardware level — reseat it in the PCIe slot and confirm all required power connectors are attached before looking anywhere else. If it does appear but the system still shows no display, the problem is more likely the monitor cable, the GPU’s output port selection, or a driver-level issue rather than a power delivery fault.
7 Common Mistakes When Diagnosing a Dead PC
- Replacing the PSU first without testing it. A paperclip test and a multimeter check take ten minutes and can save the cost of an unnecessary new power supply.
- Skipping the wall outlet and cable check. A dead outlet, a tripped surge protector, or a bad IEC cable produces symptoms identical to a dead PSU.
- Ignoring the CPU’s separate EPS power connector. Many builders confirm the 24-pin is connected and assume that’s sufficient, forgetting the CPU also needs its own 4+4 or 8-pin connector seated.
- Assuming any fan movement means the PSU is fully healthy. A PSU can spin its fan briefly via the paperclip test and still fail under real load — a swap test or multimeter check is more conclusive.
- Not checking for a case standoff short. An extra standoff installed under a spot on the motherboard with no mounting hole can short the board the moment it’s screwed into the case.
- Overlooking front panel connector orientation. A reversed or mis-pinned power switch header can look correctly connected while doing nothing electrically.
- Chasing a beep code that doesn’t apply. Beep code charts vary between BIOS vendors and even board generations — always check your specific motherboard’s manual rather than a generic chart if a debug LED or Q-code display is also available.
Troubleshooting Guide: Specific Symptom Patterns
- No lights, no fans, absolutely nothing: Start with Steps 1 and 4 — this pattern points most strongly at wall power, the PSU’s rear switch, the power cable, or a fully dead PSU.
- Power light on, fans spin, but no display: This is a “won’t POST” pattern — focus on Steps 7, 10, and 12, checking RAM, GPU seating, and the debug LED before assuming a dead board.
- System powers on for a split second, then shuts off immediately: Classic sign of a PSU protection circuit tripping due to a short circuit somewhere downstream — breadboard the system (Step 12) to isolate whether the case is causing it.
- Continuous reboot loop, never reaching POST: Often caused by an unstable RAM configuration or a failing CPU power connector — reseat both and try a single RAM stick in the slot closest to the CPU.
- Fans spin at full speed immediately and stay there: Frequently indicates the motherboard isn’t receiving a valid PWR_OK signal and is running fans in a fail-safe state — check PSU voltages with a multimeter (Step 5).
- Debug LED stuck on “CPU” even after reseating: Try a different RAM configuration first (a CPU LED can sometimes be a red herring for a RAM training failure), then verify the CPU is fully seated with no bent socket pins.
- Works intermittently, sometimes boots fine: This pattern often points to a marginal PSU rail, a loose connector that shifts with vibration, or a motherboard capacitor beginning to fail — check Event Viewer logs (Step 9) for a pattern. Pay attention to whether the failures cluster around specific conditions, such as only happening on cold mornings (suggesting a connector that contracts slightly with temperature) or only after the system has been under heavy load (suggesting a thermal or VRM-related cause rather than a purely mechanical one).
- Brand new build has never posted once: Double-check the CPU socket for bent pins before doing anything else, confirm RAM is on the motherboard’s tested compatibility list, and verify the CPU power connector (not just the 24-pin) is seated.
- No signal to the monitor specifically: Confirm the monitor cable is plugged into the GPU’s output, not the motherboard’s integrated graphics port, if a discrete GPU is installed — this mix-up produces a “no signal” message that looks identical to a genuine POST failure.
- PC was recently moved and now won’t turn on: Suspect a connector that shook loose in transit or a static discharge event — reseat every internal cable before assuming component failure.
Advanced Tips: BIOS Recovery, CMOS Reset, and Capacitor Inspection
Recovering From a Corrupted BIOS
If a failed BIOS update or a bad overclock is the suspected cause, most current motherboards include a hardware-level recovery method that works without a CPU or RAM installed: ASUS calls its version USB BIOS Flashback, Gigabyte offers Q-Flash Plus, MSI provides a dedicated Flash BIOS Button alongside its in-BIOS M-Flash utility, and ASRock uses BIOS Flashback alongside Instant Flash. Each works similarly: rename the correct BIOS file to the name specified in your manual, place it on a FAT32 USB drive plugged into the designated port, and hold the flashback button until the indicator LED stops blinking.
Clearing CMOS
An unstable overclock or a corrupted BIOS setting can prevent POST even when every component is healthy. Clearing CMOS resets BIOS settings to factory defaults, either via a dedicated button on the rear I/O, a jumper on the motherboard, or by removing the CR2032 coin cell battery for a minute with the system fully unplugged — a firmware-level fix worth trying given that HP’s April 2026 firmware bug was found to force repeated 48-digit BitLocker recovery key prompts that effectively block Secure Boot on affected Windows 11 23H2-25H2 PCs, and Microsoft has separately confirmed that its own April 14, 2026 update, KB5083769, which targets Windows 11 24H2/25H2 builds 26200.8246 and 26100.8246, can trigger that same BitLocker recovery screen under specific Group Policy configurations even when the underlying hardware is completely healthy.
After clearing CMOS and powering back on, a successful reset is usually confirmed by the system booting straight to a “BIOS settings have been reset” or “overclocking failed, load defaults” notification screen instead of your usual boot logo. If you’re on a Linux live environment and want to double check the board recognizes a fresh POST cycle, dmesg right after boot will show the kernel’s own timeline of hardware initialization, which can help confirm whether devices like storage controllers and USB hubs came up cleanly this time around:
$ dmesg | grep -i "ACPI\|pci.*enabl" | head -10
[ 0.123456] ACPI: Core revision 20240827
[ 0.234567] ACPI: PM: (supports S0 S3 S4 S5)
[ 1.045678] pci 0000:01:00.0: enabling device (0000 -> 0003)
[ 1.045912] pci 0000:00:1c.0: PCIe Bus Error: severity=CorrectedA “PCIe Bus Error” entry like the example above, especially if it repeats every boot, is worth taking seriously — it often correlates with exactly the kind of marginal connector or slot contact issue described in Steps 10-12, even on a system that otherwise appears to boot normally.
Visual Capacitor Inspection
On an older motherboard or PSU, a visual check for bulging, leaking, or discolored electrolytic capacitors can reveal age-related failure that no amount of reseating will fix. Look specifically at capacitors near the CPU socket’s VRM and around the PSU’s primary/secondary stages — a domed top instead of a flat one is a reliable sign of a capacitor at end of life. Independent hardware labs such as igor’sLAB, which routinely tears down PSUs to inspect component quality, treat visibly aged capacitors as one of the more reliable predictors of an approaching failure, well before a unit stops working entirely.
Complete Working Project: An Interactive Diagnostic Script
Rather than re-reading this guide from the top every time a system fails to power on, this small Python script encodes the same decision tree as an interactive command-line questionnaire. Run it from a laptop or a second working PC and answer its yes/no prompts to get a targeted next step instead of working through every section manually. Save it as pc_diagnose.py:
def ask(question):
return input(question + " (y/n): ").strip().lower().startswith("y")
def main():
print("PC Won't Turn On -- Interactive Diagnostic\n")
if not ask("Do any lights or fans move at all when you press power?"):
if ask("Have you tried a different wall outlet and power cable?"):
print("-> Run the paperclip PSU self-test (Step 4). "
"If the PSU fan doesn't spin, the PSU is likely dead.")
else:
print("-> Try a different outlet/cable first -- this is the "
"single most common cause of a totally dead system.")
return
if ask("Do fans spin and lights turn on, but there's no display?"):
if ask("Does your motherboard have a debug LED or Q-code display?"):
print("-> Read which stage it's stuck on (CPU/DRAM/VGA/BOOT) "
"and jump to the matching row in the beep/LED table "
"(Step 7-8).")
else:
print("-> Reseat RAM (one stick at a time) and the GPU "
"(Step 10), then breadboard outside the case (Step 12).")
return
if ask("Does the system power off again within a second or two?"):
print("-> Suspect a short circuit. Breadboard the motherboard "
"outside the case (Step 12) to rule out a case standoff "
"or chassis short.")
return
print("-> Symptom doesn't match a common pattern -- work through "
"Steps 1 through 12 in order, starting with the PSU paperclip "
"test.")
if __name__ == "__main__":
main()Run it with python3 pc_diagnose.py on any working computer. It won’t replace careful hands-on testing, but it collapses this entire guide into a two-minute conversation that points you at the right section immediately, which is especially useful if you’re troubleshooting over the phone for a friend or family member who can’t easily follow a full written guide.
A Sample Diagnostic Walkthrough
Here’s how the process typically plays out on a real “won’t turn on” case, to show how quickly this narrows down in practice. A builder reports fans spinning for about half a second, then the system going completely dark, with no display and no further activity.
- Wall outlet and cable check (Step 1) come back fine — a lamp plugged into the same outlet works normally.
- The brief fan spin followed by a full shutdown (Step 2) points strongly toward a short circuit tripping the PSU’s protective circuitry, not a fully dead PSU.
- Unplugging all non-essential peripherals and extra drives (Step 3) doesn’t change the symptom.
- The paperclip test (Step 4), performed with the PSU disconnected from the motherboard entirely, shows the PSU fan spinning steadily and staying on — so the PSU itself passes a basic check.
- Reseating RAM, GPU, and both power connectors (Step 10) doesn’t resolve it either.
- Breadboarding the motherboard outside the case (Step 12) — the system posts normally and stays on.
The conclusion: something inside the case, not any individual component, is causing a short the moment the board is mounted. Re-checking the standoffs reveals one extra brass standoff sitting under a spot on the PCB with no corresponding screw hole — removing it resolves the issue completely. This is a genuinely common outcome, and it’s exactly the kind of fault that a full component swap-out would never have found, since every part involved was actually fine.
When to Stop Troubleshooting and Consider a Warranty Claim
If you’ve breadboarded the system, confirmed PSU voltages with a multimeter, swapped in known-good RAM and a spare GPU, and the board still shows no signs of life or gets permanently stuck on the same debug LED, the motherboard or CPU itself is the most likely remaining suspect — though it’s worth double-checking your model isn’t caught up in a known firmware recall first, since HP warned in April 2026 that a BIOS bug could brick “all HP Commercial Notebooks, Desktops, Workstations,” a defect the company said affected more than 1 million devices worldwide. At that point, stop opening the case further — continuing to reseat and reinstall components on a genuinely dead board risks damaging pins or connectors and can complicate a warranty claim. Most motherboard manufacturers offer 3-year warranties and CPU manufacturers typically cover 3 years as well; check your specific purchase date and retailer return policy before assuming a component is unrecoverable.
Keep a record of every test you’ve already run before contacting support — most manufacturer RMA processes ask specifically whether you’ve tried a different PSU, cleared CMOS, and reseated components, and having concrete answers (rather than “it just doesn’t turn on”) speeds up the claim considerably. If the board has a Q-code display, note down the exact code shown when the failure is at its most consistent, since support technicians can often look up that code against a much more detailed internal reference than what’s publicly documented in a beep code chart.
Preventing Future “Won’t Turn On” Failures
A large share of these failures trace back to routine maintenance gaps rather than sudden component death. Keeping cables tidy and away from fan blades during a cable management pass reduces the chance of a pinched or stressed connector working loose over time. Periodically confirming your power supply is still within spec before it fails completely catches a degrading unit early, especially on a PSU that’s several years old and has never been load-tested. If you’re building a new system from scratch, following a careful CPU installation process avoids the bent-pin scenario that causes a disproportionate number of first-boot failures, and keeping your motherboard’s BIOS current via a safe BIOS update procedure avoids firmware-related boot failures on brand-new hardware.
It’s also worth periodically running a full GPU and CPU stress test on a healthy system specifically to catch a marginal PSU or a failing VRM before it degrades into a full no-power failure — a unit that’s struggling to hold voltage under sustained load will often show early warning signs (unexpected reboots during gaming, Event Viewer Kernel-Power entries) weeks or months before it fails outright. If your last upgrade involved swapping in a new graphics card, revisit the connector seating described in the graphics card installation guide any time a high-end GPU is involved in a “won’t turn on” or “no display” case, since that single connector is responsible for a growing share of these tickets on 2026-generation hardware.
Frequently Asked Questions
My PC won’t turn on but the motherboard light is on — what does that mean?
A lit standby LED on the motherboard means the PSU is delivering +5V standby power correctly, which narrows the problem down to the power button/front-panel connector, the CPU, or the motherboard’s power-on circuitry rather than a completely dead PSU.
Is the paperclip PSU test safe to do?
Yes, when done correctly: disconnect the 24-pin connector from the motherboard first, make the bridge with the PSU unplugged from the wall, and only then plug the PSU into AC power. Never perform this test with the connector still attached to the motherboard.
Why does my PC turn on for a second and then shut off?
This pattern almost always indicates the PSU’s protection circuitry is detecting a short circuit or an out-of-spec condition and shutting itself down defensively. Breadboarding the system outside the case is the fastest way to determine whether the case itself is the source of the short.
Do all motherboards still have beep codes?
The BIOS firmware on most boards still supports beep code output, but many current motherboards don’t include an onboard speaker by default, so you may hear nothing even during a genuine hardware fault unless you connect a spare case speaker to the header.
Can a bad CMOS battery stop a PC from turning on?
Rarely on its own, since the CMOS battery mainly maintains BIOS settings and the system clock when unpowered. However, a dead CMOS battery combined with corrupted BIOS settings can occasionally prevent a successful POST, and clearing CMOS or replacing the battery is a quick, low-risk step to rule this out.
How do I know if it’s the PSU or the motherboard?
The paperclip test isolates the PSU from the rest of the system entirely — if the PSU fan spins during that test but the full system still won’t power on with everything connected, the fault most likely lies with the motherboard, a connected component, or a short in the case rather than the PSU itself.
Should I try clearing CMOS before more invasive troubleshooting?
Yes — it’s quick, reversible, and resolves a meaningful share of no-POST cases caused by an unstable overclock or a corrupted setting, so it’s worth doing early rather than as a last resort.
What’s the single most common cause of a PC that won’t turn on?
Loose or incompletely seated connectors — the 24-pin main connector, the CPU’s separate EPS power connector, and the front-panel power switch header — account for a disproportionate share of these failures relative to genuinely dead hardware, according to general POST diagnostic documentation and repair community consensus tracked by outlets such as Gamers Nexus.
Is it worth buying a dedicated PSU tester instead of using the paperclip method?
A dedicated PSU tester is a worthwhile purchase if you troubleshoot PCs regularly — it reads voltages across every rail automatically and is safer than probing live pins with a multimeter by hand. For a one-off repair, the paperclip test combined with careful multimeter probing gets you the same diagnostic information for the cost of a paperclip.
Can overheating alone prevent a PC from turning on at all?
Overheating typically causes a system to shut down mid-session or throttle performance rather than preventing power-on in the first place, since thermal protection engages after the system is already running. If a PC won’t turn on at all, even after cooling down for an extended period, the cause is more likely electrical (PSU, connectors, short circuit) than thermal.
Related Coverage
- How to Upgrade a Graphics Card: 12 Steps, 90 Min [2026]
- PC Cable Management: 12 Steps, 60 Min [2026]
- ROG Ally Armoury Crate SE Setup: 12 Steps, 80 Min [2026]
- Clone a Hard Drive to an SSD: 12 Steps, 45 Min [2026]
- How to Install a PS5 Disc Drive: 12 Steps [2026]
