Random freezes or shut downs

[forumsviewer]

How does one properly diagnose what seems to be random freeze or shut down?

I have a friend who's custom build system (year old) just started to freeze or shut down. He told me that it would freeze where he couldnt use his mouse or keyboard and the application he was in stopped. He wasn't gaming, he was doing some "busy work" 2D in XP. He then told me the next day it just turned off and he couldn't turn it back on without first unplugging the power cable and plugging it back in.

I told him that most likely it is the power supply. He replaced the power supply and same thing. It shut down where there was no power and the only way he could get it to turn back on was to unplug the power cable and plug it back in.

He has been running speedfan/coretemp/prime95 for several h ours and it hasnt said anything. There are no noticeable things in the event viewer IMMEDIATELY before the shutdown or freeze. There were some ERRORS in system in event viewer, but they were a good 10 minutes time stamp before it shut down.

So again, how does one diagnose something like this? What CAN cause a system to just power off? I assumed it has to do with heat or electrical current which is why I recommended he change the power supply (he did, same result) and run prime95 incase it was a heat issue, but he can run prime95 for hours and hours without it shutting down.

 

[blkhogan]

What are the temps looking like after running a while? What are the spec's of his machine?

 

[forumsviewer]

What are the temps looking like after running a while? What are the spec's of his machine?

A bit older machine running AMD X2 5600+ and a legacy model ATI Radeon PCI Express graphics card. I believe 2x1GB DDR2 memory. Thermaltake 550W Power Supply (new). I would have to ask him about the other specs of his machine.

The temps have increased a bit while running prime95. It is showing a low of 24 degrees C and a high of 56 degrees C when he ran prime95 for 4 hours.

On Speedfan it does say "Temp3" is 127 degrees C, but once running prime95 it has gone to 84 degrees C. Maybe this Temp3 is a misreading or maybe it is the culprit? If it is the culprit, why in the world would is lower temp after hard 100% cpu usage? Anyone know what temp 3 of Speedfan is and if it can be a sensor that simply doesn't exist so it just shows "whatever?" Googling around shows that many people see "127 degrees C"

 

[blkhogan]

I dont personally use speed fan, not sure what the temp3 is, someone else should be able to tell you what your seeing.
Are all the drivers up to date? There are no BSOD to catch a code? What is your temp setting in the bios for the CPU cut off? Maybe a bios safety shutting her down.

 

[Thrackan]

Try CoreTemp or HWMonitor. They are generally better tuned to read motherboard sensors correctly.

 

[forumsviewer]

Try CoreTemp or HWMonitor. They are generally better tuned to read motherboard sensors correctly.

I'll have him do this and report back.

But generally speaking, how does one properly diagnose what seems to be random freeze or shut down?

 

[DavidV]

Most of what's being post here is Technical. Has your friend look in the Possibility of some virus. I realize this is probably not the problem but it would be good to run some software cleaner.

crap cleaner
Malwarebytes
spybot.
virus/software

IT'S just a thought if he has not already done this.

 

[Fourstaff]

Maybe there are corrupt system files somewhere?

 

[phanbuey]

it could be a virus, it could be a dying mobo, ramstick, gfx card... it could be dust short-circuiting the mobo... it could be alot of things, could be a degrading processor that crashes when cool'n'quiet lowers the vcore and freq (although unlikely). You've ruled out the PSU... so thats not it.

Usually the best way to start disgnosing these things is to try the cheapest solutions first (virus scanning), a can of air, etc etc... Ideally you would want to find a way to cause the system to crash, because that would narrow down your list alot.

But after that, you really need replacement parts (cheap ones) that you know to work...

 

[forumsviewer]

He ran hard drive utility from WD. No issues. He also ran memory test for over 14 hours. No issues.

Must be something with the motherboard, graphics card or cpu I would imagine. None of which are easy to test because the problem doesn't occur consistently (it may be many hours before it happens or it may be a day or two)

 

[phanbuey]

He ran hard drive utility from WD. No issues. He also ran memory test for over 14 hours. No issues.

Must be something with the motherboard, graphics card or cpu I would imagine. None of which are easy to test because the problem doesn't occur consistently (it may be many hours before it happens or it may be a day or two)

sounds like the mobo... My system used to just randomly shut off... usually after Oblivion - turned out to be the mobo.

 

[TIGR]

Completely random freezing/shutdowns on a system with no recent hardware changes, and hard drive, memory, and PSU fine, usually means a software issue (file corruption or malicious software). I suggest following DavidV's recommendations.

 

[forumsviewer]

Completely random freezing/shutdowns on a system with no recent hardware changes, and hard drive, memory, and PSU fine, usually means a software issue (file corruption or malicious software). I suggest following DavidV's recommendations.

Hrrmmmm... but wouldnt formatting and reinstalling windows fix the issue since the virus/trojan/spyware isnt on the system any longer?

 

[TIGR]

Yes that would fix a software problem assuming the install disc was fine. I didn't realize the system had been reformmated/reinstalled (if that's the case)!

 

[blkhogan]

I would lean towards maybe a temp issue or possibly a mobo issue. You have ran all the tests, replaced the psu, formated and reloaded os and so on. What I usually do with a problematic system is strip her down and start from scratch. Install what is needed to boot and work from there. Its going to be something simple, usually is. Good luck.

 

[DavidV]

Hey forumsviewer if it was a virus/among other badies reformatting would help. I have done that many times and it will work. blkhogan has some good sugestion's also.

 

[{JNT}Raptor]

+1 on the Mobo.....and a new PSU in as well.......like most else here....I've seen it before....higher % call is Mobo.

Hope it helps.

 

[forumsviewer]

Anything to "test" the motherboard without swapping it?

 

[Thrackan]

Anything to "test" the motherboard without swapping it?

Testing the entire motherboard would be a problem, but you can test it piece by piece.

I see you already tested memory and HDD, so that should be alright.

You can run stress tests on the CPU and GFX card.
I'd go with OCCT for CPU stress testing and FurMark for GFX.

Try the OCCT test that also stresses memory usage (puts some strain on the northbridge) to rule the northbridge of the motherboard out as well.

And if it's the PSU, testing Furmark and OCCT at the same time will probably cause stability issues since that draws a lot of power.

I would also check for viruses and malware. Even if they can't be removed, if you find them you might have also found your cause. Reinstalling and directly updating Windows and a virus scanner/firewall would get rid of any malware/virus that's not a rootkit.