I recently dealt with one of the strangest PC problems I’ve run into. What looked like a dead computer turned out to be an intermittent contact issue that took a fair bit of trial and error to track down. Writing it up here in case someone else hits the same symptoms.
The Symptoms
The computer was running fine when the screen suddenly filled with graphical corruption. The whole system froze, including keyboard and mouse, so this wasn’t just a display glitch.
I forced a shutdown and powered it back on. It made it through the motherboard splash screen and GRUB, then corrupted again and froze.
The next reboot was worse. The system started power cycling on its own, and eventually stopped displaying anything at all. No BIOS screen, monitor reporting no signal.
Hardware:
- ASRock H81 motherboard
- Intel Core i3 4th Gen with Intel HD integrated graphics
- DDR3 memory
- No dedicated GPU
Initial Suspects
The graphical corruption pointed at RAM or the graphics side first, since this system uses integrated graphics and shares system memory with the GPU.
Ruled out a bad VGA cable early by swapping it. Considered a failing PSU too, since the boot loop behavior fit that pattern.
What I Tried
- Verified the monitor and VGA cable
- Cleared the CMOS
- Reseated the RAM
- Tried the RAM in different slots
- Booted with only the bare minimum hardware connected
- Checked that the CPU fan, PSU fan, and hard drive were all getting power
- Did the paperclip test on the PSU to confirm it powers on
None of it fixed anything. Pulling the RAM out entirely just gave me spinning fans and no display, which is normal on a board without diagnostic speakers or LEDs. Didn’t tell me much either way.
The Fix
The actual fix came from pulling the CPU. Just reseating it didn’t do anything at first. But after taking it out, dropping it back into the socket, and letting it settle for a second before locking the retention arm, the system suddenly POSTed.
From there it loaded into BIOS, sat stable for several minutes, then booted into Linux without a hitch. No corruption, no freezing.
Looking Back
This also explains an earlier incident that never made sense at the time. Months back, the machine wouldn’t boot until I pulled one of the RAM sticks. I assumed that stick was bad. It’s been running fine in the system since, so it was never actually the problem.
What I Think Happened
I can’t be 100% certain, but everything points to an intermittent contact issue between the CPU and the socket. LGA1150 CPUs don’t have pins themselves — the motherboard has hundreds of tiny spring contacts that make contact with pads on the CPU. If even a few of those aren’t seating properly, you can get exactly this mix of symptoms: corruption, freezes, failed POST, boot loops, and what looks like a memory problem but isn’t.
Since the integrated GPU pulls from system RAM, a flaky socket connection ends up looking a lot like a RAM failure, which is what sent me down the wrong path for a while.
For now the machine’s been stable. If it comes back, next step is pulling the CPU cooler off and checking the socket itself under bright light for bent or dirty pins before touching anything else.