Escalating Failure Leads to Death

My computer began having some odd problems in the last few weeks. On bootup the screen would return a useless flicker that would go away upon reset. Then during use the screen would go black, the ethernet card and hard drive would start functioning wildly, and then the system would either go into a low power state (only the cpu and graphics card fans would be running) or reboot. This failure used to take 10 to 30 minutes before it would happen, but at the end the system would shut itself down after only 2 or 3. Now it won't start at all.

So what have I done?

Before the total failure I ran HD and memory check utilities (both 3rd party and those from the hardware vendors), and received clean results from both. I checked my error logs and found only a few minor errors--mostly related to software and not unusual compared to what was returned on the other systems in my house. I had cleaned the registry with CCleaner, but most of the issues were with lost extensions...nothing unusual. I swapped out my graphics card, but that had no effect (I took the flickering to be a graphics issue at first). I've changed the power plug from my APC power strip to the wall, but no effect. Now when I start the system the fans on the power supply, cpu and graphics board are the ONLY signs of life--no FDD whirring, no HDD startup, no lights on the CD. Power LEDs on the front bezel, motherboard and ethernet card are lit.

What did I try today?

Using a multitester I checked the voltages to the drives and got the 12 and 5 V I expected. I pulled the MB connector and got readings on two pins (5 and 3.3). I stripped the system of its ethernet, sound and memory chips and tried to bootup, but same result.

I thought it was a heat issue at first because the problem escalated the longer the computer was on, so I had the door off for the final few days and the room's ambient temperature was 58F. Dust was blown out of all the nooks and crannies. I even pulled the MB battery to test that it was delivering the correct voltage.

I'm either guessing this is either a motherboard that's fried, or there's a voltage issue from the PSU that's not apparent until the system starts to boot. If I had the equipment to benchtest the MB in another system, I'd do that (same with the PSU), but it just might be time to stop throwing money at an old system.

Specs: SuperMicro 370SSE, Pentium 3, 516MB in a relatively new Allied case (AL-c400 ATX PSU). The computer was built using parts cobbled together from a few older systems and some parts from online sources, and has been completely stable for at least two years. I was running WindowXP Home, and since it was my 10 year old daughter's computer it didn't handle anything more strenuous than The Sims and Webkinz World.

If there are other tests I should run please let me know. Oddly enough the motherboard's speaker connector is not compatible with the case's connection, so I don't have that valuable piece of information available...but I never got hung up when booting up. Funny thing, though (and I'm not thinking this has anything to do with anything)...the system finally died for the last time when I started it in Safe Mode; a few minutes later it crashed and couldn't be revived.
2 answers Last reply
More about escalating failure leads death
  1. Unplug everything including the HD, optical drives, case fans, PCI cards, etc. Leave only the CPU/HSF, 1 DIMM RAM and video card plugged in. If the system fails to POST to BIOS you have it narrowed down to MB, CPU, video card or PSU. Maybe your local computer shop has a MB speaker like this. With a working speaker, remove ALL the RAM and boot the system. If the MB does not beep signaling no RAM in the system, you have a bad MB or processor.
  2. Death by a junk PSU?
Ask a new question

Read More