Weird weird problem, totally stumped - Help please!

[taunches]

Hi everyone. I'm having a computer problem that's totally stumping me. I think my mobo is starting to go, but I haven't really got a way to check this.

Here's the computer:

Everything running stock:

ASRock Dual Sata 939 Bios V2.10 (flashed to most current)
3700+ San Diego
2gb PQI (4x512 - for troubleshooting I'm only using one stick)
160gb Seagate Barracuda SATA
Geforce 7800GT
Samsung DVD-RW Drive
Antec SmartPower 450w PSU

I have other peripherals but they're all disconnected. The hard drive has passed CHKDSK /F and showed no bad sectors. RAM sticks have been gone through one pass of memtest and all seem fine.


Here's the problem:

Computer worked fine for several months. I built this back in December / January and had no problem with this system. I was gaming yesterday when I was hit with a generic BSOD (STOP message). Rebooting the computer brought it to Windows almost every time, where within 1 - 30 minutes the system would do one of the following: freeze, reboot, or crash to random BSOD message. The BSOD messages were varied - I've seen six or seven different messages, most regarding memory management (page file and etc, but others driver related like DRIVER_CORRUPTED_MMPOOL or IRQ_LESS_NOT_EQUAL).

Thinking this was probably a driver problem, I managed to do a system restore to a few days previous. This fixed the problem for a couple hours. I went back to gaming, when after a couple hours the system froze again and the whole cycle of crashing restarted.

So, I flashed my bios. No help. Time to reinstall windows. No help. Computer still crashes, and the problem is even more strange: Now, whenever I power down the computer, I have to wait several minutes (15-30) to get it to boot. Otherwise it powers on (fans and drives spin up) but the computer won't post. It just sits there powered on but nothing happens. That seems like it might be a temperature issue but the BIOS and Everest both indicate that everything is running at a good temperature (40C CPU, 50C GPU).

I've reseated all my essential hardware, popped the cmos battery for a bit (after the bios update), removed all non-essential hardware, reinstalled everything that I can think of.

The PSU seems to be running a bit hot but it always does (the Antec smartpower series is designed to be quiet so the fan isn't super efficient, as far as I can tell). It also makes a very very very faint hissing noise when plugged in, even when the power is off.

I'm totally stumped at this point. Any suggestions or tests to run would be appreciated. Unfortunately I don't have a spare pci-e card to see if it's my video card, but I don't know why my vid card would be causing all these problems either.

 

[surelock]

My first suggestion would be to replace your PSU with a better one (more air flow/ more watts, like >500) if possible.

 

[waylander]

His PSU is fine for that comp... over powered even. It does sound like a heat problem and i never really trust those temp sensors. If you can get inside your case right after it shuts down with a digital thermometer and stick it inside the heatsink of the cpu.

The kicker is that you can't restart for some time after shutting down... that definitely sounds like heat.

What case and how many fans do you have... also what is the ambient temperature where you live... remember you built this in the winter with cooler temperatures and probably didn't consider cooling a priority.

 

[The_Abyss]

This is heat, but it sounds like the PSU is overheating.

 

[frozenlead]

gotta be heat, or power lapses<BR>
cheack all the molexes on the the PSU first-make sure its putting out good power and its not damaged or arcing<BR>
add some fans and such, tighten up cables.<Br>

 

[antichrysler]

I agree. It is definitely heat related. I'm not entirely sure why but I also get the strange feeling it might be the motherboard. If the traces get too hot or the northbridge overheats it could also be kicking out.

My first step would be try a different PSU. Or just try moving the PSU to the outside of the case where ventilation is no longer an issue. See if that fixes it.

What are your voltages looking like? That is a good indicator. Keeping in mind that the ATX spec only allows a 5% tolerance. So your measured values shouldn't be more than 5% off. If they are it could just be a faulty PSU.

Generally when power supplies are overheating the voltages slowly start dropping.

 

[parlee]

put a window fan next to ur comp with the side panel off, if the problem continues upgrade cooling system

 

[sailer]

The PSU seems to be running a bit hot but it always does (the Antec smartpower series is designed to be quiet so the fan isn't super efficient, as far as I can tell). It also makes a very very very faint hissing noise when plugged in, even when the power is off.

I'm totally stumped at this point. Any suggestions or tests to run would be appreciated. Unfortunately I don't have a spare pci-e card to see if it's my video card, but I don't know why my vid card would be causing all these problems either.

I agree with everyone that heat seems to be a factor and the PSU might well be bad. The thing I wonder about is the "hissing noise". That makes me wonder if a capacitator or something is bad. The motherboard may be the culprit here. If it overheated and something gave way, it might start being erratic as to when it quits working. Also, it would get progressively worse.

For sure check the tempts closely and check out the PSU, but look toward the mobo as well. Just a thought.

 

[taunches]

Thanks guys. I'm going to pick up a new PSU; if that doesn't work I'll get a new motherboard. If that doesn't work I'm going to put my box out in traffic. I'll let you all know what happens.

 

[cptdread]

I would recommend the power supply as the easiest thing to test. I had similar issues with a power supply. It slowly went bad on you. It will affect the fans, Processor, Video Card, etc.. If it was the MB that went bad on you, typically, if you follow the same steps, it would crash when you duplicated the steps (give or take).

With this setup, I would recommend a 550+ W Power supply just to be safe down the road.

 

[antichrysler]

One thing I've noticed is power ratings mean basically nothing with PSU's. Well they're important. But so many companies cheat in their rating system that I would go as far as saying a generic 550W is about the equivalent of an Enermax 350W... Maybe even 250W.

With his set up I would say a GOOD 450W is more than enough. If he was buying a cheap Okia (just one I saw at a local computer store) or some other cheapo you probably need atleast an "800W".

 

[taunches]

Went out to Circuit City and paid a little bit too much for a 480W Antec Truepower PSU. I have an Antec Sonata 2 case and I knew it would fit and all.

This seems to help. The reboot cooldown issue is no longer around... the PSU is also a lot cooler to the touch than the old one.

CPU is running a few degrees hotter with the new PSU. 43-44C idle isn't awful but it's not great either. I have some arctic silver lying around that I should apply but I need some cleaning solution to get the old crap off the heatsink.

Anyway, thanks again for the suggestions. Hopefully this will do the trick!