Computer shuts down seemingly when under load, but tests say everything's fine. Where do I go?

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
Hello!
I've been having some issues with my PC lately and I'm at my wits end. The main problem is that my PC seems to randomly shut down when under load and I can't quite explain to myself why that happens.

As an introduction, during the course of this last year I've changed three PSUs and two different sets of RAM- so either I'm incredibly unlucky, or something really wrong is going on.

My current specs are:
Gigabyte H87-HD3 Mobo
Intel i5-4670 3.4ghz
Nvidia GTX 970
8 GB Ram
Cooler Master G700 PSU
Windows 7 64 bit


I started noticing this problem a monthish ago when my PC would shut down by itself. That only happened once, so didn't think that was serious. After a few days I upgraded my GPU and got a 970, and the problem represented itself- the PC would shut down under load (mostly while playing games), stay off for less than a minute, then turn itself on again.

After a while of this I thought there was something wrong with the PSU, as in it couldn't handle the new GPU (I had a CM 550W, which should have been enough). Eventually, after shutting down a few times, the PSU decided it had enough and stopped working entirely- I tested this on another computer and it wouldn't start.

So at this point I pretty much narrowed it down to the PSU not being able to handle the new GPU, so I replaced it with a 700W thinking that would do the trick. Well, it didn't- same exact problem as before.
At that point I ran some other tests and it turned out that my CPU was overheating a little bit. I ran the standard OCCT test and the temp shot up to over 85° in just a minute, so thinking that might have been the problem I cleaned off the old thermal paste, applied some new one, and installed a Cooler Master 212 Evo cooling system. I ran the OCCT Linpack test and the CPU temp never went over 70°. So then again I thought I had fixed the problem, but it's still here.

It's worth noting that with every single test I made, everything seems to be in the norm. The power consumption is ok, the GPU temp is ok and so is the CPU temp. I kept scanning stuff with HWMonitor, GPU-Z while playing games, I ran a memtest that gave back no errors, and I ran the AIDA64 stress test which again gave absolutely no indication that something is wrong.

Still though, the random shutdowns happen. Sometimes they don't occur at all, sometimes they occur after having played for one hour, sometimes after 15 minutes, and so on. I can't seem to pinpoint once specific action that causes the shutdowns- they do, however, only happen when I'm playing games. I thought at first the more demanding games were causing it, but I recently discovered that even older and less demanding games are causing shutdowns now.

The one thing left that I haven't tested yet is the possibility of power surges. As I mentioned earlier I already had to replace several components during this past year- and in December 2013, exactly one year ago, I built this rig from scratch after that the mobo on my old one died (how? no idea).
I'm still quite skeptical about this option though since I can't quite explain why surges would only hit when I'm playing games. No other appliances in the house seem to suffer from this, and the computer never shuts down if it's not under load (I could use it for hours for work or browsing and nothing ever happens).

How should I move from here? Any help would be greatly appreciated! I'm afraid that those abrupt shutdowns will end up damaging some of the components in the long run.
 

Nerdboy

Reputable
Jun 22, 2014
192
0
4,710
Hmm you adding volts or trying to over clock it a bit? That heat sink should be fine as long as fans are on it... But if it's hitting 70 that sounds like heat issues... Check and see if it beeps when it shuts down if so then it's heating issues
 

JustCallMeAllen

Distinguished
Nov 30, 2014
195
0
18,760

Thanks for that info. I wouldn't worry about the voltage settings. It sounds to me like you have a bad MOBO. Unfortunately the cause of the problem can be many things. In fact, it might even be the Processor, but since you have it cooled, I don't see it being a problem now. There should be no reason for that processor to be reaching 80 under load. It would be understandable if you were rendering big videos or running 3 instances of Far Cry 3, but the motherboard sounds to be the root of the problem. How long have you had it, and is there a warranty still on it? Perhaps you need to RMA it.

 

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
Can't quite remember the exact day, but I remember clearly I got it last December so pretty much one full year ago.
I might be able to get my hands on a UPS with a surge protector tomorrow and I'll try some more, see if I can rule out the power surge option. I was also considering trying things by taking out one stick of ram at a time, and eventually reformatting just to make sure there's nothing in conflict software-wise.

For completeness sakes here are two screenshots- one is taken from OCCT after almost 1hr of the Linpack stress test and another one is from HWMonitor. I'm actually noticing that something might be wrong with the voltage on the mobo, if that makes any sense.

Can you see anything from those?

http://imgur.com/cMGgiJ5,s1xwqZg
 

JustCallMeAllen

Distinguished
Nov 30, 2014
195
0
18,760


Some of the voltages are a little odd, but nothing that would indicate something major. Go ahead and check for surges and see what you come up with. I am genuinely stumped here, but still think it might be a motherboard problem. Does the PC get slow right before this crash happens?
 

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
No, there's absolutely no indication that something is about to happen. One second I'm playing, the next everything just shuts down. It seriously all happens within the arc of one second- no freezes, nothing slowing down, no crashes. It's as if someone suddenly pulls out the power cord, fiddles around with it for 30 seconds, then plugs it back in. Must be spirits, or something.
 

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
Here's another update. The shutdown happened again after 26 minutes of gaming. I've also been noticing though that as soon as the PC reboots, the fans tend to spin like crazy and the entire machine gets quite noisy. After a few minutes, everything goes back down to normal.
Here's a Speedfan screenshot post-shutdown.

http://i.imgur.com/UNLaXsz.jpg

 

Nerdboy

Reputable
Jun 22, 2014
192
0
4,710
Your fans are only at 27% ... And it shows your gpu burning alive 0.o also do you have any fan controls installed like asus has (fans will speed up and start that way) but still sounds like psu or board issue
 

JustCallMeAllen

Distinguished
Nov 30, 2014
195
0
18,760

I agree with you, I wouldn't say "Burning Alive" But its pretty close. It is most definitely a Board issue since you have already gone through 3 PSU's


 

JustCallMeAllen

Distinguished
Nov 30, 2014
195
0
18,760


No I know, I just don't think that is the cause of his problem (Which I knew you weren't implying anyways). Anywho, it is most definitely a motherboard problem and I would try to get an REM asap. That can go dangerous on you quickly. Example, fry your CPU, GPU, RAM, PSU (As it has already done to your old 550w),remove or destroy HDD data and or HDD itslef, etc. Definitely get that checked out.
 

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
Those 56 degrees on the GPU came up after 26 minutes of gaming, it usually idles at 35 degrees. I'm no expert, but I think that's a normal temperature while gaming.
In any event, would there be any way to know for sure whether or not the mobo is the actual problem? I'm almost convinced about it, but before I take any steps I'd like to know for sure.
 

JustCallMeAllen

Distinguished
Nov 30, 2014
195
0
18,760


You are right. 56 degrees is about normal, maybe a bit high, but that is still too high for that little of time. As we have stated it is most likely a board issue. Otherwise, it is going to be multiple components failing all-together. Hopefully you can get it sorted out. If you want to check if it is the motherboard, constantly monitor the temps that you are getting. Open up Task manager and go to the "Performance" tab and constantly monitor that as well. If you notice any unusual spikes and that when these spikes happen, it stays that way, then it is most definitely the motherboard. And by it "staying that way" I mean if you see a spike in the usage and it stays at that percent of usage before it shuts off, then I would almost swear my left arm on it that it is the motherboard.
 

Edo Rosencrantz

Reputable
Dec 2, 2014
10
0
4,510
I think I see what the problem is, or at least, I hope I see.
My PSU fan isn't spinning, for some reason. Apparently it's one of those smart fans that are supposed to change their speed according to the load, but mine ain't spinning at all so I guess it's not so smart after all.
Any steps I could take from here?
 

westom

Distinguished
BANNED
Mar 30, 2009
931
0
19,160

If a not spinning fan has failed, then CPU temperatures far above 50 some degrees exist. If an Intel CPU, then it simply slows down. Did you see both symptoms?

You keep accusing suspects. Then try to fix those suspects without first collecting facts that identify that suspect as guilty. For example, you shotgunned power supplies on an accusation that, for example, ignored many other parts of a power system. Did you know about a power controller that decides when to power on or off a computer? That even decides when the CPU can exexecute? IOW nothing is replaced or even disconnected until one minute of labor obtains some criticaly numbers. Then those numbers say what is suspect or exonerated ... immediately. Without even disconnecting one wire.

If the power system is unknown, then every other subsystem can act defective. So what was accomplished? You do not even know if any changes solved it or just created another defect. Name any subsystem or part that is, without doubt, known good? IOW any conclusions have too much doubt.. One minute with a meter means answers without doubt.

Back to that fan as an example. If it has failed, then temperature numbers say it should be spinning and is not. But the numbers you posted said that fan is doing what it is suppose to do.

Your system is shutting off. What determines that? Power controller. Why accuse other parts that do not make that decision? Only way to see what happens is one minute using a meter. Using instructions requested from the fewer who know how this stuff really works. Otherwise shotgun: just keep replacing good parts until something works. No reason to post. Since your choice of suspects is as good or better as anyone elses choice.