The PC Broke Down Again
The next PC I build will go for stability because of this
April 29, 2022
Photo by ThisIsEngineering from Pexels
A year ago, I wrote an article about how I debugged CPU debug light on POST and BSOD when idle that was happening on my PC. Back then, I tore down and reassembled the machine doing essentially nothing but somehow fixed the issue. I attributed the fix to just a matter of reseating components. I did notice that the debug lights cycle faster after a reseat. But whether that meant anything or not, like detecting and initializing parts faster and without error, I don't know.
This time around, I'm getting a slightly different issue: a CPU or VGA debug light on POST that happens at random, with a higher chance of happening if the PC was powered down for a while. I have a pretty good guess it's the board that's failing. But PC parts are expensive. If I am to buy a new part, I should be absolutely certain that I'm replacing the bad part. And if my guess was wrong, this new part must be usable elsewhere and not be an expensive paperweight. Very little room for error.
And so, I spent a day testing.
The first step is to rule out possible causes. I started stripping parts down to the bare minimum of what you could call a functioning PC. I swapped the PSU and GPU from another PC. I even tried booting off a single stick of RAM. Storage was implicitly ruled out because it usually shows its failure long after BIOS has loaded. Despite all that, the issue persisted but I've narrowed down the culprits to either the CPU or the board, with more emphasis on the board but not definitive.
Now if I buy a replacement for one of them and find out later it wasn't the bad part, I will be forced to buy the other to fix the issue. To increase the chance of getting it right the first time, the next step is to do research, lots of research. I can't possibly be the only one with this issue. And indeed I wasn't, except what resolved the issue for everyone else is all over the place. Some even RMA'ed everything they thought was bad only to find out the part they least suspected was the cause.
Investigation is still ongoing, I have not decided what to do next. I was hoping this build would be hands off and last the next five to ten years, but it looks like that's been cut short. It also feels like the older build that I did was better, the one that is now my parents' PC. It's also in these situations that I remember that sometimes, having a powerful rig is pointless if I spend half the time trying to make it work. There's a reason why enterprise workstations exist. 😉
PS: I also find that forums and boards are terrible for support. They're really just long-form chat, a middle ground between email and instant messaging. Threads are long-winded and chatty, while important details become lost in the threads. There's also very little incentive to come back and report findings and solutions, so you almost never find your answers from people who have resolved their issues.