Computer Won't Turn On? Complete Troubleshooting

You press the power button and get nothing. Not a flicker, not a whirr from the fan, just silence and a dark screen. I’ve been there — actually, I’ve been on the other end of the phone call about two thousand four hundred times at this point. Seven years of remote PC repair at RebootDoctor and these “completely dead” machines are probably our single most common ticket category.

Here’s what I wish I could tell every person who calls us panicking about a dead motherboard: around 73% of the time, it turns out to be something almost laughably simple. Bad power outlet. Cable that’s not seated all the way. The rocker switch on the back of the PSU got bumped to off. Embarrassing stuff that takes 30 seconds to fix once you know where to look.

The remaining cases were almost always a dead power supply, which is like a $45-70 part and maybe 20 minutes with a Phillips head screwdriver. So before you start shopping for a new PC, let’s go through this properly.

Quick Diagnostic

Press the power button

▼

Anything happen at all?

NO — totally dead

1. Test the outlet with a lamp

2. Check PSU switch on back

3. Reseat power cable both ends

4. Try different power cable

→ Still dead? PSU is likely failed

YES — fans spin, no display

1. Reseat RAM sticks

2. Try one stick at a time

3. Reseat GPU / try onboard

4. Check monitor cable + input

→ Still blank? GPU or board issue

Check the Obvious Stuff (Yes, Even If You Think You Already Did)

I’m not trying to be condescending here. But let me tell you about a call I took back in March — guy was absolutely certain his ASUS ROG board was toast. He’d already ordered a replacement motherboard on Amazon. $289. I asked him to plug his phone charger into the same outlet his PC was using. Dead. Tripped breaker in the garage. His PC was fine. He returned the motherboard.

So grab literally anything with a plug — phone charger, desk lamp, hair dryer, whatever — and plug it into the exact same outlet your PC uses. Not the one next to it. The same one. Different outlets on the same wall can be on different circuits. If your test device doesn’t get power, go find your breaker panel and reset it.

Now here’s a thing people don’t think about with power strips. I had a Belkin surge protector that lasted maybe three and a half years before it just… stopped actually protecting anything. The little green “protected” LED was still on. All the outlets looked fine. But the MOV inside had burned out silently and the strip was basically an expensive extension cord at that point. If your PC is plugged into any kind of strip or UPS and it’s more than three years old, take the strip out of the equation entirely. Plug straight into the wall.

While you’re back behind the PC, check two things. First: that thick power cable going into the power supply. It’s called an IEC C13 if you want to get technical, looks exactly like the cord on an electric kettle. Yank it out and push it firmly back in. I probably see one loose power cable for every eight or nine service calls, mostly from setups where the PC lives under a desk and gets kicked around.

Second thing — there’s a rocker switch right near where the cable plugs in. One side says “I” and the other says “O.” If somebody bumped it to O, your power supply is literally switched off and nothing you do will make it start. I’m embarrassed to say how many times this has been the answer. We had a customer last month whose cat knocked it while jumping on the desk. The dude had been troubleshooting for two days before he called us.

The Paper Clip Trick for Testing Your Power Supply

Alright so the wall power is definitely fine and you’ve reseated the cable and the switch is on I. PC still dead — no fan spinning, no click, no brief LED flash, nothing at all.

Nine times out of ten that’s a dead PSU.

Quick background on PSU lifespan because it matters here: a solid power supply from Corsair or Seasonic or EVGA will go 7 to 10 years easy before it starts getting unreliable. But the cheap no-name units that come bundled with $40 Amazon cases? I’ve seen those die after 18 months. I’m not even joking. One brand in particular — I won’t name them but their cases are all over Amazon in the $35-50 range — we’ve replaced probably a dozen of their included PSUs in the last year alone.

Anyway, you can test a PSU with nothing more than a metal paper clip. I’ve done this so many times I could probably do it blindfolded.

Unplug the PC from the wall first — this part is non-negotiable, you’re about to mess with power delivery components. Now open the side panel and find the 24-pin motherboard connector, the big wide rectangular plug from the PSU. Press the latch and disconnect it from the board. Take your paper clip, bend it into a U, and you’re going to bridge two pins on the connector you just pulled out. One leg goes into the green wire hole — that’s PS_ON, pin 16 on the ATX spec — and the other leg goes into any black wire hole, which is ground. Usually there’s a black one sitting right next to the green.

Once you’ve got the paper clip bridging those two pins, plug the PSU back into the wall and flip the switch to I. Watch the PSU fan. Spinning? Good news, the supply can at least turn on. Dead silent? That’s your answer — the PSU is toast, order a replacement.

But wait — here’s where it gets annoying. A PSU that passes this test can still be bad. The paper clip just proves it can switch on with zero load. I’ve personally seen power supplies pass this test perfectly and then fail the second a graphics card tries to draw 200 watts because the +12V rail droops down to like 10.8 volts. If you’ve got a multimeter lying around, measure the voltage rails while the PSU is connected and the PC is trying to boot:

Rail	Normal Reading	Bad Reading (replace it)	What It Powers
+12V	11.40–12.60V	Anything outside that range	CPU, GPU, fans
+5V	4.75–5.25V	Outside range	USB, SATA drives
+3.3V	3.14–3.47V	Outside range	RAM, chipset logic

Don’t own a multimeter? And the PSU is five-plus years old? Honestly just replace it and skip the guessing game. A Corsair RM550x runs about $65 on Amazon right now, 10-year warranty, modular cables, quiet fan. I recommend it to basically everyone.

And please — never open the metal casing of a power supply. A friend of mine who does bench repairs got zapped by a capacitor in a PSU he thought was fully discharged. Those caps hold 400V+ and they can retain that charge for days after you unplug the unit. He was fine, but only because he pulled his hand away fast enough. Not worth it.

Five Things to Reseat Inside the Case

PCs vibrate. People drag them across carpet. Someone shoves a USB drive in too hard and the whole tower scoots back on the desk. Connectors come loose over time and it’s weirdly unpredictable — I’ve worked on machines where a RAM stick was visibly not fully seated and the thing ran fine for a year and a half until the owner moved apartments and the car ride finally jostled it free.

Standard safety ritual before you dig in: yank the power cable out of the wall, mash the power button and hold it for a full 30 seconds (bleeds residual charge off the motherboard capacitors), and touch bare metal on the case frame so you don’t zap anything with static. The static thing is honestly overblown unless your house has carpet and zero humidity, but it takes half a second so just do it.

Start with the 24-pin connector — that wide rectangular plug running from the PSU to the motherboard. Disconnect it, flip it over, and look closely at the pins. You’re checking for any sign of heat damage — melted plastic, blackened or discolored pins. I pulled a 24-pin off a Thermaltake PSU last year and the 12V pin had literally fused to the plastic housing. Scary looking but also a clear-cut answer for why the PC wasn’t booting. Anyway, if yours looks clean, snap it back in until you hear it click.

Now here’s the one that catches people off guard more than anything else: the 8-pin CPU power connector. It sits in the top-left area of most motherboards, labeled EPS12V or ATX12V depending on the board manufacturer. Without this cable, the processor gets no power whatsoever. And unlike a missing GPU cable or loose RAM which might give you some sign of life, a disconnected CPU power cable gives you absolute zero — no beep, no LED, no fan twitch, total silence that looks identical to a dead motherboard. I once spent 40 minutes on a video call with a first-time builder who’d put together his whole Ryzen 5 rig and forgotten this one cable. He was days away from returning the board to Amazon as defective.

Next up — and this is genuinely my favorite fix because the success rate is wild — check your RAM. Reseating memory sticks solves roughly 1 out of every 5 dead-PC cases we handle where the power supply tests fine. Not one in twenty. One in five. Pop both side tabs on each DIMM slot and pull every stick out. Then put a single stick back in, preferably in slot A2 (second from the CPU — that’s usually the primary slot), and try to boot. Works? Add the next one. DDR5 modules are particularly obnoxious because you can push them in and the tabs kind of close but the stick is actually sitting a millimeter above where it needs to be.

Your graphics card — assuming you’ve got a dedicated one — has its own power cable from the PSU. Could be a 6-pin on older cards, 8-pin on current-gen stuff, or the new 16-pin 12VHPWR connector on RTX 40/50 series. What most people don’t realize is that on certain boards, a GPU without power doesn’t just mean “no video output.” The board straight up refuses to POST because it sees a card in the PCIe slot that isn’t powered. So even if you only care about getting the system to turn on, replug that GPU power.

Last thing, and this is weirdly common: those skinny little wires running from the front panel of your case to a row of tiny header pins on the board. Power switch, reset switch, HDD activity LED, power LED — all of them are just friction-fit onto those pins with zero latching. If the power switch header slipped off, the button on the front of your case isn’t connected to anything and pressing it is physically meaningless.

Here’s what I tell people to test: pull the PWR_SW connector off the header entirely, grab a small flathead screwdriver, and briefly tap it across the two PWR_SW pins to short them. If the machine fires right up, your motherboard is fine and the problem is either the case button itself or the wire running to it.

That Watch Battery on Your Motherboard

There’s a CR2032 coin cell sitting in a little clip on your board. Same battery you’d find in a watch or a car key fob. It costs maybe two bucks at CVS.

This battery keeps your BIOS settings stored — boot order, XMP memory profiles, date and time, overclocking settings — while the PC is unplugged. When it dies, different boards react in different ways. Some just reset the clock to January 2010 every time you unplug. Others won’t boot period. I ran into an ASUS Z690 Prime last October that flat-out refused to POST because of a dead CMOS battery. No error LED, no beep code, no indication whatsoever. Just dead silence. Popped in a fresh CR2032, thirty seconds later it booted fine. The customer had been quoted $200 for a board swap by a repair shop in his town — I felt bad for the guy. Two dollars and fifty cents at CVS and thirty seconds of his time versus two hundred bucks in labor. That’s the kind of thing that makes me want to write articles like this one.

Bad BIOS flash? Power went out while you were updating? Maybe you cranked up the overclock too far and now the machine won’t start? Any of those situations — pull the CMOS battery, count to thirty slowly, drop in a fresh one. Everything goes back to factory. Boot order, memory timings, XMP profiles, fan curves, all of it wiped clean.

Debug LEDs and Beep Codes — Your Motherboard Is Talking

This section only applies if your PC does something when you press the power button. A brief LED flash. Fans spin for half a second and stop. A quick click. Anything.

If that’s happening, your motherboard is actually trying to boot. It’s failing somewhere in the process and it’s usually telling you where if you know how to read the signals.

Most boards from ASUS, MSI, Gigabyte, and ASRock — anything above the absolute budget tier from roughly the last 6 years — have four small LEDs near the top-right edge labeled CPU, DRAM, VGA, BOOT. The board lights them up one by one during startup. Whichever one stays lit is the stage where the boot process got stuck.

If it’s the CPU light that stays on, go back and double-check that 8-pin EPS cable I mentioned earlier. Also worth looking at the CPU socket itself with a flashlight — bent pins are a thing, especially on AMD’s AM5 platform where the pins are on the board, not the chip. I wasted 45 minutes on a Ryzen 7 7700X build last year going through every other possible cause before I finally grabbed a magnifying glass and spotted a single pin in the socket that was leaning maybe two degrees off vertical. Couldn’t see it without magnification.

DRAM light means the board can’t detect your memory — pull everything out, put one stick in at a time, try different slots. If it’s the VGA LED, your graphics card isn’t registering — reseat it, check its power cable, and if your CPU happens to have onboard graphics (Intel Core chips all do, AMD only if the model name ends in G), try bypassing the dedicated card entirely by plugging your monitor cable into the motherboard’s own video output. The BOOT LED is actually good news in disguise — it means the board completed POST successfully but doesn’t know which drive to boot from. SATA cable loose? NVMe drive not clicked in all the way? That kind of thing.

And if your board is old enough that it doesn’t have debug LEDs, listen for beeps from the piezo speaker. Three short beeps on an AMI BIOS means RAM. Five means CPU. One long followed by two short is almost always a GPU problem. Complete silence — no LED, no beep, no click — means you’re not even reaching the POST sequence, which loops back around to power delivery: the PSU, the 24-pin, or that 8-pin CPU connector.

What If It’s a Laptop?

Everything above was mostly about desktops. Laptops are their own headache because there’s a battery, a charging circuit, and an embedded controller all sitting between the wall outlet and the actual motherboard. More layers of stuff that can fail.

But the first diagnostic is stupid easy and I’m surprised how many people skip it: look at your charger. The power brick that comes with your laptop — Dell, HP, Lenovo, doesn’t matter — almost always has a tiny LED somewhere on it, green or white. Is that light on? If it’s dark and the charger is plugged into a working outlet, you don’t have a laptop problem. You have a charger problem. Go buy a new one, spend the $30-40 on something from Anker or the manufacturer’s own website, and do not — I am begging you — buy one of those $12-15 knockoffs from random Amazon marketplace sellers. We had a customer whose $15 generic Dell charger killed the charging IC on his laptop’s motherboard. What would have been a free fix became a $180 board-level repair.

If the charger light is fine but the laptop is dead: unplug the charger, remove the battery if you can (most laptops made after 2019-2020 have sealed batteries you can’t easily remove, but if yours pops out, take it out). Now hold the power button for a full 30 seconds. What this does is drain the embedded controller’s capacitors and force it to reinitialize from scratch. On sealed-battery laptops the 30-second hold still works — you’re resetting the power management controller, not physically draining the battery.

Microsoft has a startup repair tool built into Windows Recovery that catches some boot failures, but you can only get to it if the laptop powers on enough to show the recovery screen. If you’re getting zero signs of life, that won’t help.

Here’s a thing I always ask about with dead laptops: has it been running hot? Like really hot, fans constantly screaming, bottom too hot to put on your actual lap? If a laptop’s been overheating for months and then suddenly dies one day, there’s a decent chance a voltage regulator on the motherboard failed from thermal stress. Heat doesn’t kill electronics instantly — it degrades them slowly over thousands of hours until something finally gives.

”My Computer Turns On But the Screen Is Black” — Wait, That’s Different

Real quick — because I get this confused with the “won’t turn on” problem all the time on service calls. If you press the power button and you hear fans spinning, see LED lights, maybe the PC even makes the Windows startup sound through the speakers, but the screen is just black?

Your computer is turning on. The power system works. You have a display problem, which is a completely separate issue from everything else in this article.

Try plugging an HDMI cable into a TV or a different monitor. If you see your desktop on the TV, your original monitor or its cable is the problem. If you still see nothing even on a different display, pull your graphics card out, take a pencil eraser — yeah, a regular pink pencil eraser — and gently rub the gold contact edge at the bottom of the card. The eraser strips off oxidation from the contacts. It sounds ridiculous but I’ve gotten cards working again this way more times than I can count. If your CPU has integrated graphics (any Intel Core, any AMD with a G suffix), try pulling the dedicated GPU out entirely and plugging your display cable into the motherboard’s video port instead.

After getting a machine running again following any kind of sudden power loss, do a filesystem health check because abrupt shutdowns can corrupt Windows system files. And watch for blue screen crashes on the next few boots — especially if a Windows Update got interrupted mid-install when the power died.

Okay Nothing Worked — Where Does That Leave You?

So you’ve tested the outlet. Reseated the cable. Confirmed the PSU switch is on. Done the paper clip test. Reseated every connector inside the case. Swapped the CMOS battery. Read the debug LEDs. And the PC is still completely dead.

At this point you’re looking at one of two things.

The PSU is failing under load. It passed the paper clip test because there was zero load on it, but when the CPU and GPU actually try to draw power, the voltage rails collapse. This is super common in power supplies that are 5+ years old, especially units that have been running in hot environments or powering high-end GPUs. Replace it. A Corsair RM650x is about $70 right now with a 10-year warranty. Swap takes maybe 20 minutes if you’re being careful about cable routing.

Or the motherboard itself is dead. Look at those small cylindrical capacitors scattered across the board — if any of them have bulging tops or brownish crust leaking from the base, the board is done. VRM failure is another common one but harder to spot visually. A dead motherboard basically means rebuilding the PC, and depending on the age of the system, that might not be worth it financially.

Here’s my honest advice once you’ve hit this wall: you’ve already sunk 45 minutes to an hour into troubleshooting and you’re going in circles. A $9.90 remote diagnosis can save you from buying parts you don’t need. Our techs do a video call, watch your board’s behavior in real time, walk you through voltage measurements, and tell you exactly what’s dead. If it turns out to be software — corrupted bootloader, trashed Windows Update, boot-sector malware — we fix it in the same session. If it’s hardware, we tell you the specific part to order.

We don’t charge if we can’t figure it out. That’s not a marketing gimmick, that’s just the policy.

---

Last verified May 2026 against Windows 10 22H2, Windows 11 24H2, and current-gen boards from ASUS, MSI, Gigabyte, ASRock. Got a weird edge case I didn’t cover? Hit us up on WhatsApp — we’ve probably seen it before.