Computer shuts off unexpectedly

Status
Not open for further replies.

y-aji

Distinguished
Feb 20, 2009
26
0
18,530
Hey all,

Thanks for taking a second to review my problem. I've built a computer for about 8 months now and have had repeated problems with it. I have been building computers for some years with great success, but I think I've finally stumbled into a problem I can't fix.

I have had the motherboard in for replacement and MSI ended up giving me a new one saying that there were some fried parts, not necessarily as a direct result of overheating.

I have had the processor sent in to AMD and they returned me a fresh one saying the core was reading a little higher than they would like. They told me to monitor it and if it passed about 58oC to consider a different heat sink.. I've never seen this new processor pass 57oC, so I'm assuming it's okay with the stock heat sink.

I dont overclock ANYTHING that wasn't overclocked out of box.

Specs:

OCZ Platinum 4GB (2 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Dual Channel Kit Desktop Memory Model OCZ3P16004GK

AMD Phenom II X4 955 Black Edition Deneb 3.2GHz Socket AM3 125W Quad-Core Processor Model HDZ955FBGIBOX

OCZ ModXStream Pro OCZ700MXSP 700W ATX12V V2.2 / EPS12V SLI Certified CrossFire Ready 80 PLUS Certified Modular Active PFC

MSI 790GX-G65 AM3 AMD 790GX HDMI ATX AMD Motherboard

EVGA 896-P3-1260-AR GeForce GTX 260 896MB 448-bit GDDR3 PCI Express 2.0 x16 HDCP Ready SLI Support Video Card

3x 500gb hdd

1x DVD Burner

5x 120mm fans

Antec Performance One P180 Silver cold rolled steel ATX Mid Tower Computer Case

Problem:

The computer shuts down only when I am really dogging it. For instance, yesterday, I had Photoshop/Unreal Dev Kit/Web Browser/Messenger/Movie tut/Documentation all running on Windows 7 and was using close to 3gb of memory and had just opened the Unreal Dev test mode and after about 3 minutes of traveling around to diagnose my map, it shut down.

I was monitoring the temperatures as I was fully aware I had a lot going on and was definitely pushing it as hard as I could. Video showed about 75oC (I've called EVGA and they claim this is right about where it should be at peak temperature); my core was at about 56-57oC (again, I've called this in and AMD claims that's about perfect at peak); and my general system temperature was about 53oC and suddenly, poof, the computer went off.

The board didn't seem to receive any damage and came right back on and ran without any complaints, it was just that it seemed to say "I have had enough, I'm going down".

I don't see any event logs in Windows Event viewer outside unexpected power outage, so I'm assuming it was delegated by the board, not windows.

I have run prime95 for 24 hours to make sure that the processor isn't having trouble, and I ran 3dmark for 6 hours to make sure it wasn't video trouble and neither item alone pops the computer.

Reading around, I don't believe I'm using over 450w-500w of power, and have a 700w power supply, that wattage calculation came from a wattage calculator and I tried to get at least 200 over the suggested amount.

How can I figure out what is shutting down this dang computer!? Any suggestions? Should I just buy a second power supply and try it? is 700w grossly underpowered for this kind of setup?

I'm going in for another prime95 test this weekend in fear that the board got damaged again. Does anyone have any other applications I should use to test what is causing this?
 
Solution
It definitely SOUNDS like a heat or power issue if it's just shutting down with no warning. Especially under load. It sounds like you've ruled out heat, though -- 57C seems a little high but not alarming, and the fact that the system can turn right back on and run fine is another strike against heat; normally, it'd only run very briefly the second time and then shut off again.

I would guess the power supply -- 700W is plenty for your needs, but if the unit is going bad, that won't matter; it could cause exactly this kind of problem. If you have access to another PSU, I'd try that and see if it works ... buying a new one would be an expensive way to diagnose the problem, though,.
Have you run MemTest86+? It seems you've tested everything else I would guess is wrong. That leaves RAM as the easiest thing to test right now.

Set the RAM to the manufacturer's specs. Download the ISO file from memtest.org and burn it to CD. Then boot from the disc and let it run over night.
 
It definitely SOUNDS like a heat or power issue if it's just shutting down with no warning. Especially under load. It sounds like you've ruled out heat, though -- 57C seems a little high but not alarming, and the fact that the system can turn right back on and run fine is another strike against heat; normally, it'd only run very briefly the second time and then shut off again.

I would guess the power supply -- 700W is plenty for your needs, but if the unit is going bad, that won't matter; it could cause exactly this kind of problem. If you have access to another PSU, I'd try that and see if it works ... buying a new one would be an expensive way to diagnose the problem, though,.
 
Solution

y-aji

Distinguished
Feb 20, 2009
26
0
18,530


Come to think of it, if it were memory, I would probably get some kind of error from windows or some type of blue screen.. I think, for the heck of it, I'll run memtest but I went ahead and submitted a ticket to OCZ to see if they have any suggestions as I think you're right, the PSU must be going bad.

Lol, indeed you are correct.. That would be quite costly to just replace the PSU, which I don't have anything near 700 laying around. I've got a 5 y/o 500, but dont think that's gonna cut it.. This thing has a 3 year warranty, I'll just call them.

On a side note, I've always been unhappy with the temp of this monster, but I don't know how to cool it any further w/o water cooling! :( I spent 160$ on a case and 80$ on fans.. lol.. how much more does it want!? Dual 1680x1050 monitor immediately causes the video card to jump about 8oC which warms the entire rest of the case 3-4oC.. I think that's part of the problem, but I've always used dual monitors..

The hard drives are a bit old and run about 2-4oC higher than they probably should... w/ 3 of these older ones, they're probably not helping, but I've got them sectioned off on the bottom with direct exhaust into the psu out the back of the case.

My central air conditioner can't keep up with it in the summer at all.. I have a ridiculous bill coming off this thing.
 
Not a fan of OCZ PSU's...try this....download OCCT and run the PSU torture tests and see what happens......might as well do GPU / CPU while you're at it.

Look at graphs for temp and voltages as each one finishes and see if anything "looks odd". My guess is PSU is having trouble keeping stable voltages.
 

y-aji

Distinguished
Feb 20, 2009
26
0
18,530
20 minutes in. The core is hot, but stable...

It's sitting at 64-65oC back and forth every 30 seconds... I would be concerned, but it has sat at that temperature for 20 minutes w/o complaining.. (Maybe I do need something other than the stock heat sink on this).

Maybe I remember wrong (it has been a few months), they must have said it shouldn't pass 65. The video card is sitting just as it always does, at 75-80 back and forth..

So far, the power supply test isn't proving very taxing on the power supply but is definately showing obviously high heat levels..

Thanks again, JackNaylor, for the app suggestion.. I see what it's doing, it just goes from full load to half load every 15 seconds..
 

leon2006

Distinguished
Detail the specs of your machine including the brand and model of your PSU. Its hard to read your info with the details all over the place. Just make a list.

If your supply voltages are dropping then its short of capacity. None of your load is defective to cause such condition.
 

y-aji

Distinguished
Feb 20, 2009
26
0
18,530
That's listed in the first item I submitted. Just look above. I don't even know, now Speedfan says 9.4v all the time for 12+, but in bios it says 12.22v, same w/ 5v.. I think speedfan is giving me inconsistent results in Windows 7.

I can't get the problem to reoccur, so for the moment, I'm going to drop it and see what OCZ has to say about it. I think chances are high that the PSU is inconsistant, but I don't think anything I have is going to benchmark it accurately enough to give me answers.

I ran a quick memtest and got no errors, I'll run it longer tonight, but it's gotta be down to the motherboard or the power supply..

Thank you everyone for the help. I'll post a reply if I end up finding out the cause.
 
Status
Not open for further replies.