Random Shut Offs :(

[marco_89]

Hey guys, new person here.
Always lurking...but first time posting....and im afraid it's not good news

Monday morning of this week i got an HIS HD 6850 in my machine, and for the entire day it was flawless. Played EVERYTHING at full spec beautifully [NB, my monitor max res is 1440x900, also, im upgrading from an 8500GT]
Wednesday evening however...while playing Crysis, my PC just shut off OUT OF THE BLUE. This never happened before.
It's like electricity completely left the machine. It wouldn't power up when i pushed the power button, I had to plug the power out, flick the switch off and back on, plug it back in and then power would come on...Then it didn't happen again.....so i didn't really sweat it...

Until the day after...Yesterday...i was playing Borderlands....and it randomly shut off again....so i had to do the same unplug switchy thingy to get it back on...and since then the PC doesn't necessarily need to be playing a game to randomly shut off...
i could just be messing around online, listening to music. Once or Twice it even shut off while it was booting up

Since then I've been doing everything i possibly can to try and see what the problem

I disconnected all case fans, case lighting, optical drive from the PSU etc [everything that isn't required for the PC to boot], just to see if the PSU wasn't providing enough power [doubt it though....it's a corsair tx650]
...no dice

My mobo downclocks my RAM to 1333, so earlier this year i just bumped them up to their 1600 ratings.
i downclocked them back
...no dice

Switched the RAM between the slots...tried 1 stick each...
....no dice

I switched the graphics card to the other PCI-E slot
....no dice

i disconnected the extra hard drive....no dice


I did some detective work....ran Crysis, and on another occasion COD-MW2, in windowed mode to watch my cpu + gpu temps while gaming, till the pc shut off again.

cpu never went past 47C, the gpu [when it was under user defined automatic fan control] never passed 66C.....while under factory automatic fan control went up to 77C, although i read that they go up to 80's with no issue....so i'm pretty sure it's not heat...

the only thing i haven't checked is the MOBO temps...but the PC is up in BIOS now (been on for about an hour now, no shutdowns so far *fingers crossed*), and im watching the CPU and MB temps and its nothing to make a note about


I know some of the things i tried might have been pointless, i'm not THAT knowledgeable with PC's....I'm just really worried and i really need to get to the bottom of this...

some help would be greatly appreciated guys

 

[marco_89]

OH, also

my pc specs

AMD Phenom II X4 955 BE
ASROCK 870 Extreme3
G.Skill 2x2GB DDR3(1600)
HIS HD 6850
WD Caviar Blue 500GB
Fujitsu 100GB laptop HD
HP DVD-RW
Corsair TX650
NZXT M59 Case

idk if these are necessary -

4 x 120mm rosewill blue led case fans
2 x 120mm nzxt case fans [one with blue led's]


additional information:
I built the system myself in december, the only things that have changed since then were the changes made this week, which were the HD 6850 and the 4 rosewill fans

upgraded from an 8500gt that failed [js crashed one day and never displayed correct after that....wrong aspect ratio/had weird discoloured stripes on the display]

here's a pic

 

[iam2thecrowe]

get HWmonitor to monitor all temps while in game. play for 10 mins or so then check it. it also logs highest and lowest temps and you can check for voltage abnormalities with it (although its not always accurate for voltages but you should be able to see any odd drops in voltage under load. If its not overheating then its gonna be a faulty psu.

 

[JackNaylorPE]

Those switchey around things aren't really good ideas.

1. Make sure RAM is in right slots (check manual). Run memtest.

2. Download OCCT .... run GPU and CPU tests monitoring temps AND voltages.

3. Check Event Viewer for messages

 

[Helltech]

Download and run Memtest 86+

Download and run OCCT, see if any errors occur.

EDIT - Hivemind!

 

[bystander]

77C on the GPU is on the hot side, but probably not what would do this, but keep in mind, it's not the only part of the card that gets hot. Your VRM's could be overheating still. Try the user defined profile instead, and keep it down to near 70C.

 

[iam2thecrowe]

overheating graphics cards dont normally kill all power to the computer. normally cpu or faulty PSU. Check all your fans are spinning, including PSU fan.

 

[marco_89]

thanks so much for the suggestions guys...i'm gonna try them all today, i'll keep you posted

 

[marco_89]

All fans are spinning by the way, including the PSU fan.
Currently running memtest

 

[marco_89]

Also, the computer was on in BIOS for nearly 4 hours yesterday, no shutting off whatsoever.
memtest complete, no errors found

gna do the OCCT cpu n gpu tests

 

[marco_89]

wow...tests take an hour. guess i'll just leave and come back later.
I also notice a high pitched sound coming from the PC. I don't remember exactly when it started...but i think it's safe to say it started after i got the graphics card.

i checked the pci-e connector to see if it wasn't pushed all the way in...but it was. or at least i'm pretty sure...if i pushed any harder i fear i might break something :lol:

 

[marco_89]

CPU test finished, no errors

idle temps round 37 - 43C, maxes out at 56C

VCore voltages always between 1.4 and 1.41

3.3V voltages always between 3.36 and 3.38

5V voltages always between 5.14 and 5.16

12V voltages permanently 12.95

How's that sound to everybody?
I assume its all normal....because the PC stayed on for the entire stress test lol

onto the gpu test....

 

[marco_89]

gpu test completed...no errors....(sigh of relief)

started the gpumemtest, but it keeps saying CudaMemTest stopped working
is it because cuda is nvidia?

I'm gonna skip this and test the psu now..

EDIT: upon the start of the PSU test the gpu temps gradually increased.
now, my user defined profile for the gpu fan is that once the temp hits 65C, fans go up to 100%
it hit 65C.....and then increased...gradually....to 78C...BUT MY CASE IS OPEN!!! my side panel is off!! whatisthisidonteven

....its been revolving between 78C and 75C now for the past 10 minutes or so...

EDIT AGAIN: only 2 6850 reviews show temps under load to be around 50C - 55C, which are tom's hardware's review and bit-tech's review

all other reviews i read (hexus.net, overclock3d.net, guru3d, techpowerup, all showed between 74C and 78C..

 

[marco_89]

I'm reading there are other users who's temps under load are even higher than mine. i'm seeing 85C and such.
I also been reading [although i can't verify the validity] that the 6850 can handle up to 100C? sounds like BS to me...

 

[bystander]

Not all reviews show the card at room temperatures. Some use a Delta temps, where they subtract the room temperature from the cards temperature. The idea is that you add your room temperature to the readings they gave. Those 50C-55C readings are likely using that method.

 

[marco_89]

Not all reviews show the card at room temperatures. Some use a Delta temps, where they subtract the room temperature from the cards temperature. The idea is that you add your room temperature to the readings they gave. Those 50C-55C readings are likely using that method.

oh..thanks...so basically they all say around 70somthing degrees under load then....so temperature isn't my problem at all..




so this final test completed....no errors....
Now I've noticed the computer's been on since 9 o clock this morning, been through multiple stress tests....and not a single shut down.....

i'm so confused

 

[marco_89]

Best answer selected by marco_89.

 

[Mousemonkey]

This topic has been closed by Mousemonkey