Power supply was going

Post #65,232 by JayMehaffey 11/26/02 6:20:13 PM Reply	Power supply was going The power supply went dead. I swapped it out with another one and now the system works. That doesn't directly solve the heating problem, because it is still overheating. Apparently, as long as it gets enough juice it can work over heated, but get unstable if it's overheated and the power supply is marginal. Jay
Post #65,426 by deSitter 11/27/02 10:43:04 AM Reply	Re: Power supply was going A digital system has two states - alive and dead. If you run hot, you will enter the second realm eventually. Get it cooled down some way. -drl
Post #65,532 by JayMehaffey 11/27/02 5:44:09 PM Reply	Horrors Turns out it wasn't the power supply after all, about 2 hours after I put the good power supply in it up and died again. And this time I could see that it was neither heat nor power that where the problem. Nor would it boot even after cooling down. So today, I pulled everything apart piece by piece to find the problem. And at the core I found a horror. Some thermal paste had leaked down from the top of the CPU into the CPU socket. Apparently some had worked it's way dowm from the top before it had set up into a thick enough state, no telling what kind of damage this did to the processor or the motherboard but they both seem not to work properly. I rebuilt the system with my old motherboard and 1.1 Ghz processor that I had been planning to use for my second machine. Which is how I am posting now. I tried using the 1.8 Ghz processor in the old motherboard and the old 1.1 Ghz processor in the new motherboard, neither combination would work. The motherboard seems to be fried, unable to even try and boot correctly. Putting the new chip in the old motherboard worked a little but was very unstable. I think I may be able to salvage the old processor if I can clean it off properly, though how I am going to get thermal paste of the pins I don't know. Jay
Post #65,540 by Ashton 11/27/02 6:35:21 PM Reply	Paste If it's Si-based (and maybe some other too) - vinegar + toothbrush might do, followed by sprinkled bicarb to buffer back to neutral pH, rinse, maybe isopropyl to dry moisture around pins. Vinegar is EZ clean up from bathtub sealants too. (Don't think I'd try acetone for that last..) This might be a PITA to also try on the sockets, with board held upside down! but whatcha got to lose? Squeeze bottle with dist. H20 can ease the rinse. Still and all - at the temps you were recording.. NO e- likes to run there - MTBF probably suffered some unknowable amount. Lastly, for insurance and future intermittent-freedom: at local e- tech shop (not Radio Shack) you could look for Cramolin, or one of the Caig nostrums - or .. go to a high-end audio place and buy a sm. bottle of Tweek\ufffd (packaged by Sumiko in Berkeley CA). These do more than 'clean' - the molecule actually conducts under local hi-pressure.. ie It sounds as if your next worry would be that ALL these pins/sockets can conduct despite any residual goop you prolly can't get out of all the holes. Luck, Ashton
Post #65,746 by deSitter 11/29/02 10:56:53 PM Reply	Re: Horrors No doubt the processor overheated as the source of all this. The paste then liquefied and streamed out. I've noticed that thermal paste is often glopped on as if more were better. All that is needed is a dab rubbed on like skin lotion. You can't believe how hot a processor can get. I touched one once - exactly like touching a hot exhaust pipe (ssss). -drl
Post #65,747 by Andrew Grygus 11/29/02 11:06:27 PM Reply	Re: Horrors For CPU chips that have enough real estate, we've always put a white paper date/source sticker on the underside. Had an AMD K6 system come in for upgrade that had a dead fan. The sticker was a dark chocolate brown, but it didn't actually catch on fire. CPU still worked fine. [link\|http://www.aaxnet.com\|AAx]

Welcome to IWETHEY!