PC turns off with no error message

-JD-

Reputable
Jun 17, 2014
32
0
4,530
First my PC stats

Windows 7 64 bit
i7 2600k
mobo - MSI z77
Ram - 2 x 4gb kingston
PSU corsair cx600
GTX460 (evga superclocked)

My PC has been shutting off in the middle of gameplay. The PC takes 30-60 mins to turn on again even with external fan cooling.

All fans CPU/GFX/PSU fans seem to be operating as normal to the eye.

First instance was a few months ago, I cant remember what was happening at the time but it turned on again after a while so I just forgot about it.

5 - 6 days ago playing dragon age on high settings it happened again, I updated the GFX drivers to see if this would help.

Updated drivers prompted me to optimise some of my games so I let it. Playing Infinite Crisis with new higher settings the PC again turned off without warning only 2 - 3 mins in to the game. This had not happened before with this game. I thought it may be due to the new settings so when my PC eventually turned on again I loaded up Civ 5 which I had played with the same settings for 3 - 5 hours the day before and on other occasions without problem. About 10 minutes in to the game it happened again.

The current build of my PC had been stable without problems for about 2 years. There have been no recent changes to it.

At this point I thought it may be the graphics card as it seems to idle at a hot temp (around 60 celsius) and when tested playing civ 5 again with the fan maxed out it worked fine. Then i played it again the next day and it crashed about 10 mins in. When it turned on again i tried the game again with the fan cranked up to max but this time it still crashed after about 10 mins. When it would finally turn on again I just had chrome with a bunch of tabs open as i always do one of them was streaming and it crashed AGAIN. This was the first time it had happened while not running a game.

So to recap the basics. When this happens the computer just turns off as if unplugged. It then wont turn on again for 20 - 60 mins. When it turns on again the only sign something was wrong is the "you didnt shut down properly do you want to start in safe mode" bit. Then all is normal. The windows event viewer shows a kernel error 41 with the description "The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly." I cant find any other error messages but maybe I dont know where to look.

At this stage i dont feel like its an overheating issue with the graphics card or the CPU although it definately could be. Ive been told that the max temp for the card is something like 104 celsius. Although I could have missed it going up I have been running something that keeps my GFX temp on screen and I havent seen it go above around 82. In theory it would have had to spike considerably without me noticing for it to cause shutdown from overheating.

It crashed the night before last when doing nothing but having chrome tabs open (as mentioned) and i left it till morning. When I went to turn it on nothing happened. I looked inside the case and the power button on the mobo was lit up. I tried turning the power on and off a few times and nothing happened. When I turned the power off for long enough that the button light turned off (10 - 15 s) then turned it on again the PC booted up. It then ran stable for about 4 hours of just browsing/watching streams.

I then put in my old gfx card to test as i didnt know what else to do. Im not sure if it would really prove the other card is the issue as this card is so old and low power that if the issue was something else like the PSU was failing when drawing alot of power or something then that wouldnt show up now either as the load on the PSU will be much lower. Someone also recommended I record the system temp and voltages after using it for a while with this old low power card which were as follows.

cpu temp 34 sys temp 29

cpu core voltage 1.12V
cpu i/o 1.048V
gpu 0.00V
DRAM 1.488V
3.3V - 3.312V/3.296V (flickers between the two)
5v - 5.087V
12v 12.056V

Thats what the BIOS showed after playing League of Legends for about 40 - 60 mins. But im going from a card that plays on max settings with 60+ FPS and has two power connections and running dual monitors to a card that is running one small monitor and the game runs only on "very low" graphics settings with about 20FPS and the only power draw is directly from the mobo.

Is there anything I can run to eliminate any possibilities or something i can run that will record (temps/voltages/something else relevant) in the background so when it crashes and I boot it up again I will have some more info to go on?

Im really stuck and dont know what to do!

Any help would be greatly appreciated.
 

-JD-

Reputable
Jun 17, 2014
32
0
4,530
My case is actually really spacious and almost dust free. All the outlets have grills on them that stops nearly all dust getting in and ive checked the graphics, power and cpu fans and they are all clean from everything i can see short of opening up the PSU/GFX card or removing the heatsink from the CPU. I would have thought if there was enough dust inside to cause problems i would be able to see signs without opening up.

I know its a lot of info I just wanted to include everything to help best diagnose the problem.

Thanks for your input. Any more thoughts? Is there any kind of monitoring program that will record temps and maybe also voltages that when it crashes it will have recorded it in logs for me to look at when i boot it back up?
 

-JD-

Reputable
Jun 17, 2014
32
0
4,530
Thanks, I couldnt find a log function on Open HardWare Monitor and the only logs that i could find that SpeedFan was creating were garbage. In this instance i need to log temps from multiple sources and multiple voltages and the SpeedFan log was putting all the data in to one cell in the excel log which made it totally unreadable.

Any other ideas for logging programs or otherwise?

From the trawling ive been doing most people have said similar problems were a PSU issue. If it was a PSU issue would it show up with my lower graphics card installed or possibly not because its obviously now using alot less power?

I read somewhere about a program called OCCT that stress tests individual parts of the system. Is this something you know about and can recommend? Is this an avenue worth pursuing in general, finding something that will stress individual parts of the system theoretically causing the issue to show when only one part of the system is under stress? or does it not work like that?

Lastly i have been wondering if this shutdown issue could be damaging other hardware when it occurs?
 
OCCT is a high-stress test program. If not used carefully it will destroy your computer. I have it, but I don't often use it and don't know much about it. Thre is a website for it somewhere.


Not sure why you're having difficulty with the speedfan logs. Can you parse the data after import?

Can you perhaps upload some data somewhere so I can have a look at it? There are some tutorials on the web.
 

-JD-

Reputable
Jun 17, 2014
32
0
4,530
OCCT is off the table then as i definately dont know enough about what im doing to risk wrecking the machine totally.

Maybe i was just using SpeedFan incorrectly. I set it to log and selected all the things i want in the log which included all the temp sensors (mobo/psu/gpu/cpu) and the various voltage sensors. There seems to be no way to open up the logs back in speedfan itsself and they are defaulting to read in excel which makes sense. The problem is when i open them in excel all the data for each row is in one cell making it extremely difficult to read. Ive attached a screenshot.

mbmwsz.jpg


I may be totally missing something here, ill trawl around to see if i can find more info on using speedfan.

Edit - ive set it to log just GPU temp to compare the logs. Now i can see what its doing i can look back at the above picture and see all the data is there its just really difficult to read the way its formatted and it would take a long time to separate it manually. Surely there is a better way than that?
 

-JD-

Reputable
Jun 17, 2014
32
0
4,530
Ok so after doing a little searching ive found that changing the file extension of the log to .txt puts it in a readable format where its organised properly in columns. Ill either sort this out and post the logs tonight or tomorrow morning. Huzzah!
 

-JD-

Reputable
Jun 17, 2014
32
0
4,530
Ok.

So ive linked some files. The jpgs of HWmonitor and speedfan open. This jpg was about 5 mins before it died this jpg was 1 - 2 mins before it died.

Link to the speedfan log up to when it died. The speedfan log is quite short as i had just restarted it when launching the game i suspected would crash the PC.

The only thing i noticed really was that the GFX card got hot up to 90C but it is supposed to be ok up to 104. I also noticed one of the MOBO heat sensors was reporting much higher temps than the others averaging almost double temp which seemed odd to me although i have no idea if thats normal.

I also did some testing when it died. Previously it took 20+ mins to turn on. This time as soon as it died i took out the power cable. After about 10 - 15s the power button light on the mobo switched off. I plugged it back in and pressed to turn it on. It started to turn on for about half a second then stopped. I dont know if that is useful info. It seems if i dont unplug it takes a long time before it will turn on again if i do unplug it will at least try to turn on again but fail.

Thanks for any help you are able to provide!