Random crashing

[Martyuk39]

I have an issue - it's an a system almost identical to mine:
Antec Sonata with Truepower 380W
Asus K8V SE Deluxe
Athlon 64 3400
Geforce 6200
Crucial 2 x 512mb CT6464Z40B
250 GB Maxtor Maxline Plus2 SATA
Sony Combi 52x32x52+16xDVD black
XP Pro SP2

Every now and then the screen goes blank, but the machine is still switched on. I can't predict exactly when it does it, but it often does it if I'm in the middle of installing something, say a windows update. Or if I launch a game. The problem surfaced a few months ago and I did a complete re-install and it seemed to go away for a while. But then I didn't put it through any rigorous tests at the time. So I've thought of hardware. I've ran memtest86+ and there were no issues there. I've run MBProbe and I don't think there's anything sinister there. The -5V is 5.11, the -12V is 0.72 and the Vbat is 0.00 so they like to beep but it's the same on mine and I don't have a problem. The motherboard temperature is 37 and the CPU is 54. I haven't managed to observe the settings just a second before a crash. The Bios is set to factory default.

So...is it a software issue? Or hardware? I assume if it's hardware then it's PSU, motherboard or RAM. Or CPU in theory. I'd appreciate some help
Thanks

 

[Slammy1]

A few things about temperature. The maximum temp for those processors is 70C, and if you're idling at 54C that can be the issue. From what I've read those chips do report temperature high (oftem incorrext temps are corrected through BIOS updates), but it would seem to be the thing to stand out.

Do you see anything reported in Event Viewer (control panel, admin. tools)?

 

[Martyuk39]

Thanks - I'm going to upgrade the bios, when it's a little cooler - it started in the mid-30s but has stabilised around 50-51. There's not much in the event log, apart from a few save dumps - the date of the first one doesn't correspond to any software installation to speak of. It's been running about 4 hours now and of course I can't make it crash now. More later...

 

[Slammy1]

I would turn off reboot on errors (so you can record and Google error codes) and system restore (can generate disk errors) for now. You might try running with the side removed, see if that helps the temps but this definitely seems CPU related based on what you're saying.

 

[Martyuk39]

Well I don't know. I had it running for about 8 hours yesterday, and then a couple of hours this morning. I upgraded the BIOS and had the side panel off for a few hours yesterday. It's back on this morning and I've had no crashes, and the CPU temperature is stabilised at 45 degrees. The problem hasn't happened again. Is there a piece of software I can get that can really tax the CPU, to see if I can get the temperature up again? It seems OK now

 

[Slammy1]

In the overclockers' forum you'll find a link and instructions for Prime95. Be careful about running too long, and watch your temps closely to make sure you don't get thermal runaway. It may have simply been a BIOS issue, but I like your CPU temp a lot more now. 54C isn't bad in itself, it's continuously running at that temp with higher spikes during periods of load that will tax CPU and other components (HDDs, for example, are very prone to heat failure).

Hopefully it was simply a BIOS issue.

EDIT: Link to testing software.

 

[Martyuk39]

Right, started Prime95 test at 12.40. What do you reckon to the results?

12.40 46.5 degrees
12.45 51.5
12.50 52.5
12.55 53.0
1.00 53.0
1.05 53.0
1.10 53.5
1.15 53.0
1.20 52.5
1.25 53.0
1.30 53.0
1.35 53.0
1.40 53.0
1.45 53.5
1.50 53.5

 

[Slammy1]

Looks good, it doesn't appear you have a thermal issue. You are climing a bit at the end, perhaps another case fan would help. I'd take the side off and let it run for, say, 6 hours and monitor temps occasionally. Remember, these are load temps. Six hours does not guarantee CPU stability, it's merely an indication though you could argue synthetic stres testing vs real world apps. But memtest stable + prime stable is good evidence that those components are operating properly. Watch your voltage rails also, look to see if it runs OOS (could be a heat related issue in that sense)

The hope is that the instability was caused by conflict (driver) which was resolved through updating the BIOS. It sounds like you're much more stable now, is that correct?

 

[Martyuk39]

Yes it is more stable. The system hasn't crashed basically since I first posted on here. All I've done is upgrade the BIOS. Oh, and I updated to the latest NVIDIA graphics drivers. Forgot to mention that. And I've cleaned out that fancy grille at the front that these Sonata cases have. Looks like that wasn't done for a while.

The last test is to run some of the installed games. From the readme files Titan Quest appears to have the highest system requirements. I assume that won't tax the CPU any more than Prime. So I'll try that and then if that's OK then hopefully that will be the end of it. At the start of this message I stopped prime and the temperature has dropped to the mid-40s again

Thanks very much for all the advice.