Help! Computer keeps hanging!

Hmm, getting annoyed now. My computer keeps grinding to a halt for no reason apparent to me. It doesn't seem to be caused by me doing any one particular thing, although I've only noticed it having problems when the computer has been on for a while. Main indication that it's about to freeze is the IDE channel access light on the front of the case coming on for an extended period of time.

Specification:

OS - Windows XP Pro SP2
Motherboard - Shuttle SN45g v1
PSU - 200W Enermax unit that came with the Shuttle
CPU - Athlon XP 2500+ (usually running at normal clock, but underclocked to 200Mhz FSB since the problem arose)
1Gb Ram - 2x 512Mb sticks, one of TwinMos value PC3200 and one of Corsair Value PC3200
HDD - Seagate Barracuda 7200.7 120Gb
Graphics - Gigabyte Radeon 9600 Pro 128Mb
Optical Drives - NEC ND-4551, Samsung DVD/CDRW Combi
Other stuff - KCorp Wireless PCI card

CPU temperatures are within the usual range of my setup, 44-46C idle, 48C under normal usage, 52-58C for heavy loads. I've been running the computer with the case open to achieve these temperatures.

Once the computer does hang, it doesn't want to reboot for a while. The system speaker makes a "chirrup"-ing noise and something spins up (probably the hard drive, but it seems to do it in the BIOS as well which I wouldn't expect). It then hangs at progressively further stages throught the boot process. So to start with, it won't POST. Then with each reboot, it gets a little further through the Windows boot process.

My first thought was to suspect the hard drive, but the SMART reading seems to be fine. I'm just about to run a scan of it, but wanted to get this posted first before the computer dies on me.

Cheers for any help you may give, especially if you've read through all that!

Sounds like an overheat to me.....PSU, gfx and CPU fans spinning OK ? Exhaust fan blowing hot air out the case successfully ?

As soon as I reboot, I try and get into the BIOS. The temp. readings in there are usually:
CPU 48C
System 41C
PWM 44C (I'm assuming this is the PSU)

The CPU/System fan is going full tilt, as is the PSU fan (although that's only a 40mm job).
I just can't see why these temperatures are causing it to have problems when it's been working fine all this time.

OK, may be worth trying memtest (http://www.memtest86.com/) too.....it's obviously hardware related, just a case of identifying which component(s).

PSU struggling, HDD dying or the IDE cable. They seem the most likely to me.

I've had this happen when the PSU simply can't provide enough power to the system - it essentially shuts down a hard drive, though it would normally work fine after a reboot, hard or soft.

Hmm... I was gonna let this thread die, cos I thought it was all sorted. However... it's done it again.

It looks to be a heat problem, as so far the only thing that's been the same each time it hangs is the huge pile of paper stacked on top of the computer (hey, it's revision season - that's how my "filing system" works!). This really shouldn't be a problem though, because the air inlet it's blocking would normally not be open if the case was on.

But... Why would it overheating cause the computer to still hang on booting up to 20 minutes later? I'd expect any excess heat to dissipate long before this (although 20 mins is a guess at how long it was before my patience gave out!).

So, I've turned off automatic restart in Windows, so hopefully I'll get a BSOD next time it happens. I've put the case back on (for the first time in months with the computer actually running), and now it's a case of sit back and let it crash. That is, if it's a temperature related problem.

Edit: BTW - The hard drive scan came up with no problems, and memtest didn't come up with any problems on a single run. Oh, and my CPU temp has already jumped 3C since I put the case on!

Shuttles are known for being hot. Artic Silver 5 and some little mods and better fans/cooling is probably your best bet mate.

despite the shuttle psu's being pretty decent (I think?) for their ratings you might find your 200W unit is struggling to cope. To rule out cooling issues you could run it with it's case off and point a great big desk fan at it?

What are your psu voltages reporting at

Doesn't seem to be the heat, cos it's been under load for the past hour @ 60C (not deliberately, it seems Media Player decided to recode a whole load of stuff!).

The only BSOD I've had since putting the case back on was a complaint from the system that it couldn't write a memory dump to C: because the page file size was too small. It then proceeded to reboot without a hitch.

So, to answer the last question, voltages as stated by MBM 5:
Core: 1.65V (although come to think of it, this was only achieved by asking the BIOS to supply 1.7V)
3.3: 3.3V
5: 4.94-5.00V
12: 11.46-11.66V
-12: -12.11V
Memory: 2.5V (and this being my 2nd SN45G, this is DO NOT ADJUST!)

So, yeah, that 12V rail looks a little off, but I'm not the best judge of power supplies. Thoughts people?

Using software to measure voltages is bad, they never come out right.

yea 12v line seems low....a good cause of the instability is poor voltages

At scan we class a psu faulty if the voltage is 0.4 or more above or below

p.s If your bios is showing voltages, its best to take them from that as noni says, software isnt really usually all that accurate

do the voltages change much between idle and under load (preferrably doing something that pushes all the components ... re-encoding an avi whild copying a bunch of files from a DVD to hard drive or something.

I think its a memory problem run mem test it shall tell you.

For a proper test, how long should I leave Memtest to run?

We leave it for a minimum of 3 passes, but we have time constraints

I would recommend 6 hours minimum