Determining Motherboard or CPU

Ralck:

what are you running when the system gets buggy

is it under load when it acts up ???? or at idle


I would remove the side off the computer and blow a house fan into the box then see if you get any more crashing


I am thinking the PSU is not up to the job of running your rig or marginally so

by using the fan you will be keeping the PSU max cooled, therfore it will put out as much stable juice as possible, I am thinking this will improve your voltages / amperages enough to clean things up. if the situtation & duration improves with the fan blowing, than you know you have the culprit


I have read where multi meters are not sensitive enough to pick up fast acting spikes and droops in the voltage outputs, however the motherboard regulator is, therefore triggering the shutdowns and bsod's

In my experiance when a cpu works it works providing its cool witch you already know it is, theres always the outside chance it is bad with these complex things but I would be more inclined to look elsewere.

I would run prime's torture test and put it on "max power" (or something like that) and check out the psu with a volt meter while its running just to see if any of the rails fall off. When you check it with the meter do it at random times (as long as prime don't fail) to see if things need to really heat up to cause the failure. I would also go as far as swapping another one in if you have a spare just in case its flaky and droping one of tha rails at random times because that is ooh so hard to pinpoint.

Yep and also run memtest on each individual stick for 6 hours or so just to give that a good workout.

These situations can be very hard to pinpoint so I always try and narrow it down so I remove all non essential cards,drives,network hubs,usb devices and the like.What bothers me is the lack of bsods suggesting to me that you are right and its something major like the board or cpu and my educated guess after ruling out the above (at least the best you can) would be the motherboard.

Keep me posted I am very interested in how long prime will run without error

Just an update:

Inspected the motherboard and I don't think I saw anything that looked abnormal. I'll be taking a closer look tomorrow.

As far as when it happens, it's all over the place. I could be playing flash games in a web browser, using Microsoft Word, etc. However, I don't ever remember it happening under full load: Playing games such as Counterstrike, Battle for Middle Earth 2, Starcraft, FEAR Multiplayer, or encoding video.

Haven't found my disks with Memtest on them yet (Sorry guys, my room is a bit messy here). I'll also run Prime tomorrow, but for now I'm leaving it off to give it a nice long rest (although... it normally get's sleep at night while I'm sleeping, so I don't think it's over tired :laugh: ).

As far as the power supply, I want to believe it is working properly... I mean I had to send this unit back to get fixed... if it's still what's causing the problem, then maybe I'll just have to put a claim in with the BBB against FSP or something. I'll try scrounging up a fan, though, to see if that helps, but I don't expect it to: I have very good cooling in my case and blow out the dust bunnies regularly. Unless there is something wrong with the PSU, I can't see why it's not up to snuff. I should have more than enough power from it: I'm part of the team so I certainly know about the PSU thread, and researched to make sure I found a supply that would give me enough wattage/amperage.

I had one thought that it might actually be the video card. When I opened it up earlier tonight, it seemed rather hot, and Everest doesn't actually list a temperature for it, so I'm not sure what it's actually running at. Though if it was a case of overheating graphics I would expect to see this happen in 3D gaming, which it doesn't seem to be, or at least artifacts every now and then.

Okay, back to homework now and I'll run Prime Friday (not enough time Thursday as I have 10 hours of class/work).

I'll keep you guys posted and thanks for the replies!

Ralck:


I think you will find the house fan will provide super cooling for the PSU, which may then identify it as the culprit when its gets warmer.

I have seen this type problem with a Enermax 530 watt unit I had, as long as it was kept super cool it ran stable, when I closed the case the slightest bit of added heat from inside made it flaky again.

I know you are well versed in all this stuff but sometimes we get to close to the problem and cant see the light thru the forest.

I have all the confidence you will find the culprit, I am just as anxious to see the guilty component. I dont see many boards of your model on the net as troublesome, so I am inclined to give the mobo the benfit of the doubt, and 450 watts is cutting the line thin, especially for a 70% efficient unit.

if you remember back when you and I discussed this unit, we were looking for a way to save $$$$$ at that time the antec SP-500 was $80.00
turns out in hind sight, the SP-500 was a real dog.

they used cheap caps and cheap cooling fans, that's the recipe for disaster and disaster they got!

Well, here is a long overdue update on this situation...

I ram Memtest and it picked up errors on the first pass through. So I was kinda bumbed out because this would be the second time I would have to RMA my RAM, in which case I'd really question if Patriot was right about it being compatible. Then I thought about one other aspect of my computer that could be causing the problem... BIOS!

So I had a thought that my BIOS, since it was from somewhere around 2004 could use an update. I thought maybe it might have new features that would help with stability, so I grabbed the latest one of ASRock's website and loaded it on.
Rebooted like it asked and blue screen every time, as the Windows XP start screen (the one with the bar that goes across the screen) comes up. So I tried resetting my BIOS to the 'failsafe' settings. Everything booted up perfectly after that. But I had that annoying little problem of the motherboard not detecting my RAM properly. It tries to make it run at [email protected] instead of [email protected] like it should.
So I booted back into BIOS and set the RAM back to it's correct settings. Blue screen again. So I reset back to failsafe and once again back to DDR333.
My thinking at this point was, it's better to have DDR400 than fast timings, so I booted into BIOS and set it to DDR400 and left all timings auto, hoping it would just pick some random ones like 3-3-3-8 and I could have the correct speed of RAM. Surprisingly, though, when I booted back up, it had auto detected the RAM at 2-3-2-5!
I'm really confused about it... why I couldn't directly tell it 2-3-2-5 in the BIOS, but hey, it's working and seems to be more stable (no random restarts or blue screens in 2 days, knock on wood!).

I haven't done anything with blowing a huge fan into the side of it yet, or much actual testing of this configuration, but I plan to do so tonight and tomorrow. It just occured to me that my roommate has a window fan so our room isn't 100 degrees in the summer, so I'll try using that if I see any other issues. Anywho, I'll still have to let Prime95 and Memtest run again, and do some 'user testing' in Windows to see if I can figure out if the problem is still there.

Oh, and also one blue screen I got last week was a clear ATi error, so I updated those drivers (even though the previous ones were only 2 months old or so).

If all goes well, the problem won't show it's ugly little head again, but only time and tests will tell right now.