So I finally got around to upgrading my 3yo home desktop system, and wound up swapping out pretty much all the components because the world of bus architectures and interfaces has apparently moved on. The new rig has:
The holdovers from the old system were 4 Seagate HDDs (2 SATAs running in RAID 0 as the system drive), a couple of DVD-RW drives, and a 500W PSU. I got XP pro installed on the new system last night without too much trouble, got most of my drivers and basic programs installed, then went to bed while Windows Update was applying patches to the OS and 200GB of backed up data was transferring from a USB drive. Everything seemed perfectly fine.
Woke up this morning and the system seemed to be frozen (no video signal or response to KB/mouse). I restarted without thinking much of it (figuring Windows Update was the culprit) and that's where the problems started. Right from the POST screen, there were weird video glitches (scrambled characters etc) and Windows seemed to be having all kinds of problems, with random noise flickering across the screen regularly, lots of missing or scrambled text and graphics, and what look like intermittent resets of the video card. It was exactly the kinds of issues you'd see in an overclocked system that was overheating (things got worse and eventually the system froze again after about 10 minutes) but everything was running at stock and the core temps were 40 and 35C. I powered down and started poking around inside the case; everything seemed to be at a reasonable temperature but the motherboard chipset's heatsink, which actually hurt to touch.
I googled around a little and nobody else seems to be having this problem, so did I get a bad motherboard or somehow fry the chipset during the night? I've got a pretty good 4-fan case cooling setup and airflow has never been an issue, so I'm kind of at a loss. My only other ideas are that there's something up with the video card (seems unlikely) or the power supply is being flaky (also seems unlikely). Anyone have any insights into this one? I have the sneaking suspicion that I'll be RMAing the mobo but I hope I'm wrong. Thoughts?
It's an Apevia ATX-AS520W, rated for 520W and about 35A on the +5V and +12V rails. It's about 18 months old and I've never had a problem with it, but I've also never run it with a PCIe card before. Everything was fine until Windows was installed and had been running for a few hours, is it possible that the power supply just couldn't handle the high currents for long?
6pack: tried upgrading the BIOS, apparently the i/o is so flaky that the upgrade utility couldn't do a positive check on the new file.
cfvh: The power cable is plugged in, I even tried plugging a 4-pin plug into the socket on the motherboard that the manual says provides extra current for SLI mode. No change from either of those things.
After reading some nasty things about my PSU on Newegg I tried swapping it with a brand new Thermaltake 500W unit tonight, which made absolutely no difference at all. Judging from the fact that the glitches start right from the POST screen and that they get gradually worse (and eventually hang the system) the longer it's on, I'm thinking that the problem's in the motherboard and that it's most likely heat-related. I guess the video card is also a possibility but I don't have a spare PCIe card around to rule one or the other out.
I'm going to try a few more things tonight and then RMA the mobo tomorrow unless anybody has any better ideas. Thanks for all the help, everybody who's replied so far.
Yeah, putting a fan over it helped a little bit (read: managed to delay the freak-out by a couple of minutes). The chipset has a big fat heatsink on it though, it shouldn't be overheating to the point where it needs extra cooling when I'm running everything at stock speed. I sent it back to Newegg and they either confirmed it was bad or didn't check, so I've got a new one coming for what that's worth. Thanks again for the help.
Yep, got the new one, the problem went away and it seemed to be basically behaving itself until I fired up half-life 2 to see what the video card could do. After a few minutes of staring at some smoke effects everything went to hell again; the textures glitched and I started seeing artifacts. Shut the machine down and checked the chipset heatsink, same deal as last time (very very hot).
Out of curiosity, have you checked the temp of your chipset heatsink while the thing's running? Maybe it just gets really hot normally and the problem is something totally different. I know running an Nvidia chipset with an ATI video card is kind of dumb, but I thought it was more "sub-optimal performance" dumb, not "this will make your motherboard explode" dumb. I'll play with it some more after work today but maybe I should just exchange it for the ATI chipset version.
I spent a lot of time playing with card positions, different airflow configs (using cardboard "ducts"), and just running the thing with the case open, nothing was helping. I eventually figured out the problem when I accidentally jiggled the chipset heatsink a little bit; it was loose and not well-attached. I took it off, removed the crappy stock thermal tape (which had the consistency of dry paint), and put it back on with some Arctic Silver 5 paste to improve the thermal contact. That did the trick, it's been glitch-free since I did that and some manual GPU fan speed boosting fixed the few remaining hanging problems I was having while running games.
It really shouldn't have taken that much effort to make everything work at stock speeds but it's nice to have a computer again. I got a good deal on a much nicer case (Antec 900) over the weekend so hopefully switching from my no-namer to that will improve the airflow a little too.
I have a GA-M57SLI, too, and I have the same problem. Seems that all these boards have the same overheating issue. I put a 120mm fan close to it (29m³/h), and it runs at about 45 - 50°C. Not the best solution, yet, because I liked the board because of its passive cooling solution. So I read about the Thermalright HR-05 IFX SLI. Do you have any idea if this might fit on the board?
I already read about changing the thermal tape - but does this have mayor effects? Would I be able to use the board without a fan on it?