PC Setup Basics

PC won’t boot: quick diagnostic flowchart (RAM, GPU, storage, PSU)

Índice rápido:

Disclaimer: This guide is for basic diagnostics only. If you smell burning, or see smoke/sparks, or hear anything like loud popping, or if suspecting liquid damage, stop! Consult repair techs. Never open a power supply (PSU). Capacitors inside might harbor dangerous charge.

TL;DR
Starting puzzles works like this: (A) total dead, (B) powers up but no POST/no display, (C) POSTs but can’t find boot device, (D) Windows repair screen / boot loops. Go for a “minimal boot”: motherboard + CPU + cooler + 1 RAM stick + (GPU if absolutely necessary) + PSU. Anything that’s causing no-boot in DIY desktops will most likely be: loose power connectors, RAM not seated well, GPU power/display connectors not properly connected, or a storage/boot-order problem. Craigslist-sfolers: change one thing and re-test to narrow it down. At least’real lazy claimers can just use notes so don’t get lost in the process. Wanna dive deeper? Try:

0) Determine the exact failure type (30 seconds)

Before yanking parts, first ask yourself these two questions:
Does it appear that the PC is doing anything? (are any case or motherboard lights lighting? do any fans spin briefly, or hear the PSU click? are cool lights lighting up? RGB powers?)
Do you see a motherboard splash screen/BIOS screen aka POST, or is it black the entire time?

1) Preparation: make sure it’s safe, jfcon miracle and you’ll won’t false-negative to boot-up! #Switch

  • Power-down everything: switch that PSU to Off (O), ac power off, and hold down the actual power button of the chassis for about 10 seconds to rid of residual power. Give it a quick second to marinate. Reduce variables: disconnect all USB devices except a keyboard (preferably no keyboard for first POST test), and unhook drives and accessories.
  • Open the side panel and use bright lighting. A surprising number of “mystery” failed boots turn out to be a half-seated connector or RAM stick not fully latched.
  • Now find the diagnostic LEDs on your motherboard (if it has CPU/DRAM/VGA/BOOT), and be aware they’ll tell you at least what stage failed —e.g., RAM vs GPU vs boot device.
The first steps will be safer if you touch unpainted metal on the case before you handle RAM or GPU; there’s no need for a wrist strap but do avoid working on carpet. Hold parts by the edges.

2) If it’s not DOA, follow the fast diagnostic flowchart (also in order-for-best-results) and do “fast checks” then isolate the part

[NOTE: The chart below starts in the first column for ‘totally dead’. If that doesn’t apply go to the symptom of the next column over, etc.]

The fast flowchart: calculate your symptom, run the fast checks, then isolate the part
Your symptom Most likely area Fast checks: follow this list (in order) Isolation test / next action
Totally dead: No LEDs, no fan twitch, no click PSU / AC power / front-panel wiring 1) Try plug into a different outlet, no power strip. 2) See power rocker switch on back of PSU is On (I) 3) Reseat 24-pin ATX cable and CPU EPS 4+4/8-pin. 4) Ensure case power-button connector is on correct header pins. If still totally dead, test PSU with a good PSU tester, or (advanced) correct 24-pin jump-start method, in that case replace. Test. If PSU tests good, suspect motherboard or front-panel switch.
Powers on (fans/RGB) but no display and no BIOS screen RAM or GPU (or display path) 1) Make sure the monitor cable is plugged into the GPU (not motherboard) if you have a dedicated GPU. 2) Try another cable/port (HDMI/DP) and another monitor input. 3) Watch motherboard diagnostic LEDs (DRAM/VGA). 4) Do a minimal boot: 1 RAM stick in the recommended slot; reseat GPU; confirm GPU PCIe power cables connected. If you have integrated graphics, remove the GPU and test video from the motherboard port.
Power cycles / restarts repeatedly before BIOS screen RAM (training) or PSU instability 1) Clear BIOS/CMOS . 2) Remove all but 1 RAM stick. 3) Remove XMP/overclocks (after CMOS clear). 4) Check cooler mount and CPU fan header. If it stabilizes with 1 stick, add sticks one at a time. If it only fails under load,/or randomly, suspect PSU or motherboard.
You can enter BIOS, but Windows won’t start / automatic repair loops Storage (boot files) or OS corruption 1) In BIOS, confirm your boot drive is detected. 2) Check boot order. 3) Reseat SATA data + SATA power, or reseat NVMe SSD. 4) Disconnect other drives temporarily to avoid boot confusion. Use Windows Recovery Environment tools like Startup Repair. If drive isn’t detected, troubleshoot the drive/port/cable.
BIOS shows “No boot device” / BOOT LED stays lit Storage / boot configuration 1) Confirm drive is detected in BIOS. 2) Reseat storage. 3) Try another SATA port/cable (SATA drives). 4) For NVMe, reseat and verify it’s in a compatible M.2 slot. If detected but still won’t boot, repair boot records or reinstall OS. If not detected, suspect drive failure or motherboard storage controller.

3) Minimal-boot baseline (the fastest way to stop guessing)

A minimal boot removes “noise” from extra hardware. The goal is not to boot Windows—it’s to reach BIOS/POST reliably.

  1. Disconnect everything non-essential: extra SSD/HDDs, RGB controllers, fan hubs (keep CPU fan), PCIe cards, USB devices, front-panel USB headers if you suspect a short.
  2. Install only: CPU + cooler, 1 RAM stick, PSU, and (only if required) GPU. Connect monitor + power.
  3. Double-check power: 24-pin ATX to motherboard, CPU EPS 4+4/8-pin near CPU socket, and PCIe power to GPU if using a dedicated GPU.
  4. Attempt POST: power on and watch motherboard debug LEDs/beeps, and observe whether you reach BIOS.
  5. If you get BIOS now, add one component back at a time (one drive, then the second RAM stick, then peripherals), testing after each add. The last added part is your prime suspect.

4) RAM: the #1 “powers on but won’t POST” culprit

Mis-seated RAM can look “installed” but still fail electrically. A DRAM debug light or memory beep code usually points here. Power off and unplug. Remove all RAM.

  1. Inspect and reseat: dust will sometimes be visible in the slot, so that’s easy enough to check—then just seat 1 stick and press down confidently, firmly until both latches click—note: some boards only latch on 1 side.
  2. Make sure you’re booting from the correct slot for 1-stick use (that will often be labeled A2 in the motherboard manual).
  3. Try each stick solo (in the same slot)—if one stick always fails that way, it’s good bet it’s bad. If you test all sticks that way and they all fail in that slot, test in another slot.
  4. Once you’ve got POST, keep XMP / EXPO switched off until you have a stable system—that’s when once you can turn it on and retest. Unstable memory timings can lead to “random” boot failures.
If you’re running ASUS and you’ve got Q-LED indicators, the DRAM LED usually denotes you either missing, or executing a wrong, memory attachment. Reseating and testing 1 stick at a time is the route to go next.

Common mistakes trying to fix RAM that waste time: trying to use an even specs “matching” kit, mixing that into an XMP / EXPO-compatible kit. Actually using RGB as an indicator of installation; it’s not unheard of with RGBs that they get power even while the data pins aren’t properly making contact with the motherboard. Testing 2 changes at once, after trying both to move a slot, make sure you test XMP, while at the same time.

5) GPU / no display: Is it worth proving the display path before blaming the card?

Well, a “no signal” screen can just mean cable/port mix-up, and not a dead GPU. Your end goal is answering “is it failing to POST, or is it POSTing, only you can’t see it?

  1. Confirm where your monitor is plugged in: do you have a dedicated GPU? If so you need to be plugged into the GPU outputs (not the motherboard).
  2. Try a known-good cable and try swapping port types (DisplayPort ↔ HDMI vice versa).
  3. Reseat the GPU. Pull it out, check the PCIe slot latch, plop it back in with some fervor, and make sure it seats properly (can’t rock it), and reattach it to the case.
  4. Check PCIe power: many GPUs need one or more 6/8-pin (or 12VHPWR/12V-2×6) connectors. Not only make sure they’re fully inserted, but if this is your first build, do not throw older power supply cables onto your new graphics card. They’re not interchangeable between power supplies from different brands/models.
  5. If your CPU has integrated graphics and your motherboard video outputs support it, pull the GPU, and test using the motherboard video port. If integrated graphics work, the GPU or its power is your likely culprit.
  6. If you have another PCIe slot, try that (even if it runs at reduced speed). This helps determine a “bad slot” from “bad GPU”.
If your motherboard has any kind of a VGA debug LED (or similar) and if its on, the GPU (or graphics output stage) was probably not detected/not initialized. Treat this like a strong clue, not a 100% slam dunk.

6) Storage: when BIOS works but PC won’t boot into Windows

If you’re reaching BIOS, your CPU/RAM baseline likely is okay. You’re in “boot device / boot files / OS” territory now.

  1. For SATA drives: reseat both the SATA data cable (motherboard to drive) and SATA power (PSU to drive). Try a different SATA port and a different SATA cable.
  2. For NVMe M.2 drives: power off, reseat the drive, and confirm it’s in a compatible M.2 slot (some M.2 slots share lanes or disable certain SATA ports depending on the motherboard).
  3. Temporarily disconnect all other drives. Multiple drives can confuse boot order or make the PC boot from the wrong EFI/boot partition.
  4. If the drive is detected but Windows won’t start, use Windows Recovery Environment tools like Startup Repair.
  5. If Startup Repair fails repeatedly, you may be looking at a corrupted OS install or a failing SSD/HDD. Back up data (if possible) before heavy repair attempts.
If your drive is detected in BIOS but Windows still won’t boot, use Windows Recovery Environment’s Startup Repair as a first-line option before reinstalling.

7) PSU: the fastest checks (and the safest limits)

A weak or failing PSU can cause anything from “totally dead” to random boot loops. But many PSU “failures” are actually loose connectors or a tripped protection state.

  1. Start simple: confirm AC power is good, PSU switch is On (I), and the 24-pin ATX + CPU EPS connectors are firmly seated.
  2. Reset the PSU protection state: switch Off (O), unplug, wait 30–60 seconds, then try again with a minimal boot configuration.
  3. If you it’s not long since you replaced the PSU and now the PC won’t start, consider that the issue may lie with the PSU you just installed. Clear CMOS / BIOS settings just in case, as motherboard settings from the previous PSU (or a corrupted state) can sometimes prevent a clean boot.
  4. If you have a PSU tester, then test the PSU with that (that’s the best option). A tester is safer and provides more useful information than improvised methods of checking the PSU.
  5. If you’re going to check the PSU using a manufacturer documented “jump start” / paperclip method (especially if they themselves provide guidance on this) then go ahead do that, and be aware that it’s a basic check (“Yes, the PSU turns on,” not “Yes, the PSU is definitely stable under load”). If unsure of what you’re doing, don’t do it – either use the tester or swap in a good PSU that you know works.
Don’t open the PSU enclosure. If you think it’s damaged (burn smell, visible damage, repeated clicking, or tripping breakers) then either replace it or RMA it.

8) When you have debug LEDs on your motherboard (CPU / DRAM / VGA / BOOT): how to use them

If your board gives you staged indicators, you can think of them as a “map” of the different steps of the POST sequence. The LED that remains lit or the one that crosses when the board fails tells you what to be testing next.

Interpretation guide (typical behaviour; always check your motherboard manual for its exact meaning of each LED being lit)
Indicator that stays on What it points to What to check next (fast)
DRAM Memory not detected/unable to stable Reseat RAM, try 1 stick, try different slot, clear CMOS, disable XMP/EXPO.
VGA / GPU GPU not detected/initialized or display path issue Reseat the GPU, confirm PCIe power leads are connected, try a different output and cable, try a different monitor/cable, test with iGPU if available.
BOOT No boot device detected or boot failure Confirm drives are detected in BIOS, reseat cables and NVMe drive, disconnect other drives, client may have wrong boot order configured, try Startup Repair.
CPU CPU not detected or critical platform issue Recheck CPU EPS power lead connection, cooler pressure on CPU, CPU seating (expert advance step), clear the CMOS. If issue persists suspect motherboard and/or CPU problem.

Troubleshooting Checklist (Printable)

  • Wall outlet verified (not a surge strip), PSU switch On (I)
  • 24-pin ATX and CPU EPS 4+4/8-pin fully seated
  • Minimal boot attempted (CPU + cooler + 1 RAM + PSU + GPU if needed)
  • RAM tested one stick at a time in correct slot; XMP/EXPO off for baseline
  • Monitor connected to correct output (GPU vs motherboard); cable/port swapped
  • GPU reseated; PCIe power confirmed; test of alternate slot or an iGPU
  • Boot drive detection confirmed in BIOS; other drives disconnected; boot order checked
  • Windows Recovery options attempted (Startup Repair) if BIOS works but OS won’t boot
  • CMOS cleared after major hardware change (PSU swap especially)

FAQ

My PC turns on (fans spin) but I get no screen. Is that always a dead GPU?

No! Make sure to first prove your display path: Is your monitor plugged into the correct port (GPU vs motherboard)? Try another cable/port/monitor input. Also, watch the motherboard’s VGA and DRAM debug indicators if the board has them! Nine times out of ten what appears to be a dead GPU is actually a cable/port mix (or ram instability preventing POST).

What’s the single fastest test for RAM problems?

Boot with one RAM stick only (in the recommended single-stick slot, if applicable A2) with XMP/EXPO Disabled. You will need to clear CMOS for that.

If it POSTs reliably with one stick but not another, I’ve likely found the bad stick or ___?

Slot issue.

If the PSU fan spins during a PSU test, does that prove the PSU is good?

It only proves the PSU can turn on. It does not allow you to rule out other components or guarantee stable voltages under load. If symptoms include random restarts, boot loops, or shut down under load, swapping in a known-good PSU (or testing with proper equipment) is more conclusive.

I replaced my PSU and now the PC won’t boot—what should I try first?

After confirming all connectors go to the right places and are firmly seated, clear CMOS/BIOS Settings and see if it boots in minimal bootstrap configuration. That’s a common low-effort step after that kind of work.

BIOS detects my SSD, but Windows won’t boot. What’s the least destructive next step?

Use Windows Recovery Environment tools (especially Startup Repair). If it can’t fix it, try backing up data to another drive and then repairing boot configuration or reinstalling Windows. Avoid making repeated random “trial-and-error” commands if you have no clue what they do, because they can make the situation worse.

When should I stop DIY troubleshooting and get help?

Stop if you see/smell evidence of electrical overheating/damage, if the PC trips breakers, if you suspect liquid damage, or if you’ve isolated the issue down to motherboard/CPU-level faults (CPU debug LED persists after power and CMOS checks). At that point, bench testing with known-good parts is often the fastest route.