/ Sign-up
Your question

Random crashes - ideas anyone?

  • CPUs
Last response: in CPUs
June 12, 2004 10:27:48 PM

I am having random crash problems. Long story short, it all began with frequent crashes in Thief III where the computer would lock totally and play a scratchy sound loop. This would necessitate a hard restart. I have also experienced lockups when playing video files in Windows. Finally, I have experienced lockups when running SiSoft Multimedia test and Prime95 torture test. I have run Memtest86 and my memory is fine.

Does anyone have thoughts on where to go from here? I figure replacing my motherboard, CPU and power supply would probably fix the problem but I'd rather go at this with a rifle than a shotgun. Is my power supply adequate? Is there a way other than those 2 programs to test the CPU? Is it even likely that the CPU would have such an intermittent problem all by itself?

Here's my system description:

Windows XP SP1
1024MB DDR 2700 RAM @333
Athlon XP 2000+, non-overclocked
Gigabyte GA-7VRXP 2.0 board
Radeon 9800 Pro
Audigy 2 ZS
(1) WD 120GB 7200 RPM HDD
(1) IBM 45GXP Deskstar HDD (both on IDE controller)
(1) HP CD Writer 9500 Series (don't remember the speeds)
Antec 380W True Power PS
1 case fan, one CPU fan

Any help would really be appreciated. Thanks.

More about : random crashes ideas

June 13, 2004 6:17:24 AM

The two remaining most common causes are psu and over temp. If you download mbm5, it should tell you if it's one of those.
June 13, 2004 6:44:46 AM

Assuming its not an overheating problem like Endyen suggested -I'm beginning to sound like a broken record- but check your motherboard for bulging caps. Have a look here:
<A HREF="" target="_new">;/A>. Its by far the most likely cause IMHO.

If that is not it, it could be a dying PSU as well. Check voltages in your bios. Keep in mind low or fluctuating voltages can be caused both by a defective PSU or the badcaps motherboard syndrome though. Also note the capacitors in the PSU can die in the same way as those on the motherboard. IF its badcaps, you could try and replace them if you're good with a soldering iron. If not, replace the motherboard (and/or PSU).

= The views stated herein are my personal views, and not necessarily the views of my wife. =
Related resources
June 13, 2004 1:03:40 PM

Thanks both of you. Can either of you point me to a good resource (there are many resources but who knows what's good) for temp/voltage threshold numbers?
I think my temp is OK: it usually hovers around 40 degrees C. Plus this problem happens even with my case open.
Voltages seem OK too:

Vcore @ 1.712
Vtt @ 1.232
3.30 @ 3.216
5.0 @ 4.919
12 @ 11.712
5VSB @ 4.972

But even so, isn't it possible that with just the right combo of things drawing together the PSU could dip at just the wrong time and I'd never really know apart from the crashes?

Also, I checked my capacitors -- don't worry about sounding like a broken record, as I've never heard that before. Got a question on what is meant by "bulging." What's the threshold of normal? Is "bulging" anything beyond that really flat top? What if the four pie slice looking things are slightly pooched out but the center is still down? I know...very scientific question :-)
June 13, 2004 2:37:12 PM

>What's the threshold of normal? Is "bulging" anything
>beyond that really flat top?

Yes. the "silver" part should be completely horizontal, and the center of the caps should definately be lower than the "ring" on top. If they look anything like the caps on the link I posted above, you've got your problem right there.

Some boards/in some cases it may continue working until the top (or bottoms) literally burst and the electrolyte drops out, in other cases your MB will start acting up without even much visual "damage". Also check if the caps are still perpendicular to the MB, as they should. If not, unless you twisted them yourselve, search no further.

Here are some other pics that may help you figuring out what they should like, and how they definately shouldn't: <A HREF="" target="_new"> link </A>. Interesting to see btw, that even low current electronic appliances seem to suffer from this problem. Next time your amplifier or scope dies, you know what to check for ;) 

BTW, your voltages seem more or less okay, though are a tad low (especially the 3.3v line), your temps are definately not the problem. I'm still fairly certain its a case of badcaps. Could you take some pics and post them perhaps ?

= The views stated herein are my personal views, and not necessarily the views of my wife. =
June 13, 2004 8:35:02 PM

Well here's an update. Slightly swollen capacitors notwithstanding, it appears that an Antec 550W PSU may have alleviated the problem. Now, though, my temps are about 15 degrees F higher than before. My CPU appears to be at about 113 degrees F when idle and goes up to the 130s in the throes of a Prime95 torture test. What gives? Are these temps too high? Again, there's a lot of info out there on what temp is right, but who knows whom to trust?
June 13, 2004 8:51:08 PM

>lightly swollen capacitors notwithstanding, it appears that
>an Antec 550W PSU may have alleviated the problem

Or rather: it may have hidden it temporarely. I had the same problem as you a year or two ago, unstable voltages, crashing computer. I blamed my PSU as well (even though it was a 350W Enermax), replaced it by a 430W Enermax. problem solved... for about six weeks. Than it began *again*. That's when someone pointed me to, I checked, and indeed at that point some caps where leaking electrolyte all over the motherboard. Replaced my motherboard, and tried my old 350W PSU again, and obviously that worked fine.

A more powerfull PSU may reduce the symptons for a while, but once these capacitors are going south, there is no going back. It will die soon enough anyway. And there is a good chance your old PSU was completely ok.

As for your temps, they are good enough. Especially if its a Palomino you have, don't worry too much, they are well within spec, and not a problem if you do not intend to overclock. My old pally 1800+ remained stable all the way up to 75°C (~170°F) and beyond, but my nerves became unstable, so I reconnected the fan :) 

= The views stated herein are my personal views, and not necessarily the views of my wife. =
June 13, 2004 9:00:36 PM

Oh, and in case I didn't make my point clear: if indeed the caps are visibly swollen, you DO HAVE a problem.

Your motherboard *will* die, and it won't take long. Be fair, don't sell it to someone unwarry, either toss it in the bin, or replace the caps (or have them replaced) if you are emotionally attached to the mb. If you wish, you can keep using it until it fails with your current PSU (my guess: within a month if its up 24/7), at which point you could try with a 750W PSU that may work for a few more days ;) , but don't throw away the old PSU, and prepare to replace your mb. I'm taking bets on this one.

And if you purchased the new 550W PSU, consider returning it, and use the money for a fresh motherboard. A 550W PSU is totally overkill for such a setup, and a waste of money since the mb will be dead soon enough anyway.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
June 13, 2004 9:15:16 PM


I have an Antec 380watt PSU as well, running a system with only a few less components

heres what i got:

Athlon XP-M , running up to 2.5ghz 1.75v (sometimes i got it lower, if its very hot outside)
768megs/ram @ 400mhz
80gig HDD
2 nics

so other than the fact that i have one less hard drive and a less demanding optical drive, my system isnt that much differen than yours and my voltages are very similar to yours , and i have had no problems. but, the NF7-S is very stable because it has 4 stage voltage regulation for the cpu, whereas most other boards have 2 or 3 stage.

<A HREF="" target="_new">please dont click here! </A>
<A HREF="" target="_new">This is you, interweb junky</A>
June 13, 2004 10:29:52 PM

Like I said, the Antec 380W is most likely not the problem. The 550W PSU he has now, is just temporarly hiding the real issue. Its like upgrading your shocks because your car doesnt handle so well anymore, when the real issue is one of your tires is punctured and slowly emptying.

= The views stated herein are my personal views, and not necessarily the views of my wife. =
June 14, 2004 6:54:28 AM

I've had the same problem with my system, but lucky for me I still have system warranty. So I brought it back, they replaced the mobo, which allowed me an upgrade from my XP1800 to a 2500 Barton. They said they'd solved it but I still had the same problem, brought it back again (mildly displeased) and they found that the memory was causing the prob. They tried to blame it on me saying that the mobo was set to memory overclocking but I had not even yet entered the bios on my new mobo, so maybe they messed that up or it was the memory from the start. My guess is they first solved the prob with replacing the mobo, then upgraded the cpu (as I only ordered the upgrade once it became apparent that they'd replaced/upgraded the mobo) and made a mistake in the bios setting.

Either way, I now have a working system again with a new cpu, replaced mobo, replaced memory, and it's stable again.


<A HREF="" target="_new">New World Order</A>
June 14, 2004 8:39:54 AM

I got back one of my builds today. The psu was toast, the caps were starting to leak, and the chip was fried. I suspect it was the bad caps that did it all in. Replace the mobo before it's too late.