It’s a catastrophe! (or is it?)

I love it when what at first appears to be a catastrophe turns out to be no big deal.

Earlier my power went out, just for a minute. Once it came back on I discovered an issue with my file server. It would power up briefly then shut down. It wasn’t even making it into POST, no beeps, no nuthin’. Immediately panic starts to set in as I imagine all of the expensive, time consuming, pain in the ass possibilities. I try it several more times, each with the same result. Then, after I got ahold of myself, I decided maybe it was the power supply, as that seemed the most likely culprit.

So… I retrieved my power supply tester, turned off the power supply, opened up the case and disconnected the 20-pin connector from the motherboard. I plug it into the tester, power it up, and it’s fine. Hmm…
must be the 4 pin connector for the CPU, that’s the only other line I can think of that might keep the computer from POST-ing. Again it tests as good.

Puzzled, I plug the connectors back in and give it another try. Just as before the power LED lights and the fans spin up, but this time, one short beep! Hallelujah! I punch it up on the KVM switch and see this message: CMOS Checksum error. Hmm… could the battery be dead? I pull the battery, grab my multi-meter, and sure enough, it’s dead. Completely, 100%, graveyard dead. Happily, I have a spare battery on hand. I replace the battery and it boots right up, happy as a clam.

Hooray for successful trouble-shooting (and a bit of luck). The “experts” on the internet seem to be split as to whether or not a dead CMOS battery can stop a computer from powering up. Some say yes, some say absolutely not. Well, I now know for a fact that it can. So, if you have a computer that powers up then right back down and isn’t giving any beep codes, test the power supply and the CMOS battery before letting visions of fried CPUs, dead motherboards, or bad RAM dance through your head.

I’m Handy!

This is another one of those “hopefully someone having the same problem comes across it and finds it useful” posts.

So my video card has taken to acting wonky, of late. Whenever I’d play a game (COD4 mainly) after a few minutes it would freeze up. My guess was that it was heat related, probably a bad fan on my SAPPHIRE Radeon X700PRO 256MB 128-bit GDDR3 PCI Express x16 Video Card. This was a great “bang for the buck” card when I bought it in August of 2005 and still fulfills my gaming needs just fine today, so I’d really rather not replace it, though I feared that would be necessary. Opening the case I find the fan barely turning at all and I know that I have my culprit.

The stock fan is a proprietary piece that I’m extremely unlikely to find a replacement for (and trust me, I looked), so I’m thinking I’m going to have to buy an third party cooling solution to replace the whole factory heatsink/fan assembly, which seems ridiculous, given the depreciated value of the card today. Before I start throwing money at this thing, I decide to see if it’s fixable. I pull the card out and find that the fan is secured to the heatsink by three very small Phillips-head screws. I get the screws out, remove the fan (which is very stiff and “draggy” and remove the sticker from the bottom covering its insides. Searching for suitable lube I decide upon Boeshield T-9. Developed by Boeing for Aerospace use, sold for cycling use, seems like just the ticket. A few drips later, the fan is spinning quite freely, so I put it all back together, reinstall it and BINGO! The temp never rises near the danger range the fan spins freely and all is (for now) well. And I can get back to the important business of “shooting melon fevers” in COD4.

The moral of this story: don’t give up on “bad” cooling fans. They can almost always be replaced, resurrected (cleaned and lubed) or “upgraded” to third party cooling solutions.

Posted by Greg Evans in computer / internet, electronics, repair, hardware
Comments: (3)