PDA

View Full Version : Computer Overheating



dpDesignz
12-06-2011, 09:42 PM
Hey I have a mate whos computer freezes after about an hour of gaming, just freezes. Would this be the graphics card or overheating?

mikebartnz
12-06-2011, 09:49 PM
Hey I have a mate whos computer freezes after about an hour of gaming, just freezes. Would this be the graphics card or overheating?
If he goes into the bios he can see what the temperatures are.
Could be the power supply.

dpDesignz
12-06-2011, 09:59 PM
The power supply? Explain more please?

sahilcc7
12-06-2011, 10:57 PM
Prob not power supply (could be), what model is the pc (Desktop or laptop)? Post temp from bios or speecy/speedfan.

Hydroksyde
12-06-2011, 11:12 PM
I take it this is a desktop PC, not a laptop?

Based on that it happens after an hour, I would guess this is an overheating issue... A hardware fault would likely crash after specific operations, rather than a relatively uniform period of time.

You can view the temperatures in your BIOS setup program (usually F1, F2, or DEL just after you turn your computer on. You may see a message on the POST screen of "Press <some key> to enter setup). Or you can install temperature monitoring software that can run within windows.

If it seems to be a temperature issue, I'd suggest opening up the computer and checking all of the fans are running. Also use a can of compressed air (Jaycar cat. no. NA1018) to blow any dust out that may be clogging fans or heatsinks.

I had a buddy with a toshiba laptop which would crash within minutes of starting up. I opened the system up and blew the dust off the CPU heatsink (which was completely clogging it up), which solved the problem.

The power supply is a possibility I guess, but IMO that would more likely cause mild instability and unpredictable behavior - rather than stopping after an hour. The power supply is the metal box at the top of the tower, where the power cord plugs in to. If the total current draw of all of the components installed inside your computer (as well as USB devices that don't have their own power source) exceeds the rating of the power supply, you will have problems. Also cheap, off brand power supplies (including ones sold with the case) sometimes cause stability issues. Past experience has taught me never to trust these, and I buy a decent PSU whenever I build a new PC.

mikebartnz
12-06-2011, 11:32 PM
The power supply? Explain more please?
A dodgy power supply will usually cause a shut down/ reboot but I wouldn't rule it out if the temps are OK.
That is the power supply within the PC.

Chilling_Silence
13-06-2011, 08:42 AM
BIOS temperatures are a bad idea. Going by them won't give you a good indication, because after 30 seconds it's taken you to start muttering and cursing at the PC for freezing, restarting into the BIOS and then you getting into the Hardware Info section, your PC will have cooled significantly, in some cases by 30-odd degrees.

Easiest way to check this is to download a trial of AIDA64 (Previously EVEREST) or a free download of Speccy, and have that running while your friend games. Play games offline in some non-serious manner, and Alt + Tab every 3-4 minutes and see what temperature your CPU & GPU are sitting at.

Most CPU / GPU's will happily sit around 65 degrees, but once you start getting much higher than that, it's time to invest in some more cooling. Of course some can keep going up to around 90 degrees, but just how ideal or safe that is, is another story altogether ;)

Power supply is definitely another option, as is drivers for the GPU, though both are probably less likely. If it was PSU then I'd suggest it'd happen a lot sooner as soon as there's significant load as opposed to slowly increasing in temperature over time.

Chilling_Silence
13-06-2011, 08:42 AM
BIOS temperatures are a bad idea. Going by them won't give you a good indication, because after 30 seconds it's taken you to start muttering and cursing at the PC for freezing, restarting into the BIOS and then you getting into the Hardware Info section, your PC will have cooled significantly, in some cases by 30-odd degrees.

Easiest way to check this is to download a trial of AIDA64 (Previously EVEREST) or a free download of Speccy, and have that running while your friend games. Play games offline in some non-serious manner, and Alt + Tab every 3-4 minutes and see what temperature your CPU & GPU are sitting at.

Most CPU / GPU's will happily sit around 65 degrees, but once you start getting much higher than that, it's time to invest in some more cooling. Of course some can keep going up to around 90 degrees, but just how ideal or safe that is, is another story altogether ;)

Power supply is definitely another option, as is drivers for the GPU, though both are probably less likely. If it was PSU then I'd suggest it'd happen a lot sooner as soon as there's significant load as opposed to slowly increasing in temperature over time.

mikebartnz
13-06-2011, 01:09 PM
Got the DT's again Chill.:D

Paul.Cov
13-06-2011, 01:38 PM
Another option is to simply leave the side off the case while gaming (assuming no kids or pets around to poke their noses into the computer).
If it doesn't freeze with the side off (better cooling) then the evidence points further toward heat issues.

mikebartnz
13-06-2011, 02:04 PM
Another option is to simply leave the side off the case while gaming (assuming no kids or pets around to poke their noses into the computer).
If it doesn't freeze with the side off (better cooling) then the evidence points further toward heat issues.
I don't really recommend that as you can alter the desired air flow from the fans. You are better off making sure all dust is removed and if still over heating improve the air flow with another fan.

dugimodo
13-06-2011, 02:32 PM
I don't really recommend that as you can alter the desired air flow from the fans. You are better off making sure all dust is removed and if still over heating improve the air flow with another fan.

Well actually it works most of the time, only really awesome airflow cools better than opening the case to the world. It can theoretically cause hot spots but will reduce the overall temps. And if you have a generic PC case with a single extraction fan like many do it can lower the temps significantly.

My flatmate's pc overheats regularly despite many attempts to fix it so he leaves the side off and has a desktop fan blowing into it while gaming. Crude but effective. In any case it can't hurt to try it and see.


EDIT: - actually if you wan't to stress test the graphics download furmark and run it. It will load the GPU to 100% and give you a graph of the temperature. If you see it climbing past 90 or the pc freezes you may be onto something (my GTX460's hit 99 before I changed motherboards!)
Incidentally the Graphics driver may already have the capability of displaying the temp without downloading anything -open catalyst control centre or nvidia display settings and have a look - if it's there leave it open and launch the game & alt tab as suggested.

gary67
13-06-2011, 03:18 PM
I converted an old P4 to a server for our house I used a crappy old case I had that had bad airflow, the server was getting really hot, so I pulled both side panels of and drilled loads of 3mm holes in them. It made a huge difference

Chilling_Silence
13-06-2011, 04:57 PM
Got the DT's again Chill.:D

Yeah PF1 thinks my words are so wise it posts it twice just to make sure my message gets through :p

dpDesignz
16-06-2011, 02:14 PM
Hi guys. sorry I have been out working at the fieldays hence the late reply.

Thanks for all the feedback. It's a Tower Acer Aspire M3100. I have only just got it back up and running so will have to check stats shortly.

Will let you know how I get on. Thanks

Oh and thanks for the idea about the compressed air. Dust came flying out!

1101
16-06-2011, 02:59 PM
I don't really recommend that as you can alter the desired air flow from the fans.

Only if the case is really well designed with perfect airflow.
Most arnt

Taking the side panel off for a test run is a really good idea. Lay it over so the hot air gets out straight up .

Use the finger method to see hot the vid card gets(avoid the fan)
Check the vid card fan is working. Check for bulging caps on the motherboard & vid card. Check power supply fan is working

If the heatsinks are gunked up with dust they will need to be cleaned out.
:badpc:

dpDesignz
16-06-2011, 03:05 PM
Well I installed a gpu monitor, and the graphics fan is currently at 0rpm. I think that may be the case. How would I go about installing another fan? They don't have another fan port, so can I get an expander or something? And where can I install it?

Agent_24
16-06-2011, 03:42 PM
Check the video card fan with your eyes.

Mine reports 0rpm as it doesn't have an RPM sensor, but it's definitely running.

Have you tried Speedfan to check other temperatures?

dpDesignz
16-06-2011, 03:52 PM
Check the video card fan with your eyes.

Mine reports 0rpm as it doesn't have an RPM sensor, but it's definitely running.

Have you tried Speedfan to check other temperatures?

:). So I checked my fan, and um, the reason there is no rpm, there is NO fan! Which would explain all the issues. hmm. Now to work out how to get a fan in there....

Agent_24
16-06-2011, 04:03 PM
It doesn't really explain the issues.

I've had a passively cooled 8600GT and before that a passive cooled 6600GT - neither of them caused freezing after an hour of gaming.

However passive cards do rely on good airflow in the rest of the case so if your fans are blocked with dust etc that might cause a problem.


What actual temperatures do you get on CPU, Motherboard and Video card while gaming?

Chilling_Silence
16-06-2011, 04:09 PM
...passive cards do rely on good airflow in the rest of the case...

What actual temperatures do you get on CPU, Motherboard and Video card while gaming?

I'll second that. Those are the details we need. Post some *exact* temperatures, not fan speeds. Fan speed != relevant. Temperatures are.

dpDesignz
16-06-2011, 04:36 PM
approx 30C each cpu (2x). That is with a game running in the background, it hasn't actually been played yet. How would I check motherboard and video card temp?

btw the video card is a GeForce 7300 GT

Chilling_Silence
16-06-2011, 04:41 PM
It's more likely to increase when the game is actually being played and the GPU is being stressed.

http://pressf1.co.nz/showpost.php?p=1024257&postcount=8

dpDesignz
16-06-2011, 04:51 PM
Wow thanks. Speccy looks like it's going to be a useful little tool. Ohk so the CPU is at 30C (but was hovering around 38-40 when installing), the Motherboard is 28C, and the Graphics is at 55C. This is all running idle. I will try a game for about 30 min and post specs then

Agent_24
16-06-2011, 05:04 PM
Does that program support logging?

I would log the temperatures at an interval of 1 second, then play the game until crash.

Check to see what the temperatures were just before it crashed, to see if they might have been the problem.

dpDesignz
16-06-2011, 05:06 PM
not sure. how would I go about checking/doing this?

dpDesignz
16-06-2011, 06:32 PM
Ohk it's been trying to load a game for about an hour now, cpa @ 30C, Motherboard at 27C, & graphics at 60C

Agent_24
16-06-2011, 08:33 PM
They're pretty good temperatures then.
I doubt overheating is causing the freezing problem, most likely something else.

Check the board\graphics card and PSU (if you're game) for bad capacitors and then run Memtest86+ if they all look OK.

Chilling_Silence
16-06-2011, 11:16 PM
I'm with Agent_24, 60 is pretty acceptable.
You could always try prime95 and see if that stresses the CPU? It's a long shot there, but if memtest86+ fails then I'd go with Power Supply personally?

Agent_24
16-06-2011, 11:34 PM
Yeah my old 8600GT ran about 75-80 degrees while gaming so 60 degrees sounds good to me.

Power supply could be checked two ways... properly with load tester and 'scope or by substitution with a known good unit.

A visual inspection may also be useful, if it's full of leaky capacitors then it's probably dumping a ton of ripple into the motherboard\GPU VRMs, not to mention everything else.

dpDesignz
17-06-2011, 10:59 PM
Well memtest86+ has passed, so just need to try the prime95. :). Thanks for all your help so far guys

dpDesignz
22-06-2011, 12:37 AM
So all my tests went fine. I couldn't get it to freeze anymore, so gave it back to my mate. Within an hour of him having it, he had frozen it! Playing CIV. See screenshot for comp dets. Any ideas?

Chilling_Silence
22-06-2011, 07:59 AM
Is that *while* playing the game?

I'm still going to hedge my bets on the GPU...

I take it it's not blue-screening, but everything just goes black?

dpDesignz
22-06-2011, 10:43 AM
while playing the game. No blue screen or anything. It basically just freezes in spot and you need to crash to get running again.

SolMiester
22-06-2011, 10:50 AM
Freezing is usually memory related in my experience, whether or not it passes memory test.
You could try giving the memory a tweak in voltage, is it running at stock speeds?

dpDesignz
22-06-2011, 10:55 AM
No idea. How would I do this? I'm pretty sure it is at stock

dpDesignz
22-06-2011, 10:55 AM
So just to confirm. No one thinks it's overheating at all?

Chilling_Silence
22-06-2011, 11:14 AM
I still do think it's the graphics card (overheating?). Does your mate have Onboard video?

Pull the graphics card out if you can and run it for a day like that.

Better still, see if you can borrow a replacement card from another mate or something for a day and see how they go.

SolMiester
22-06-2011, 11:17 AM
No idea. How would I do this? I'm pretty sure it is at stock

if you select manual setting in bios for memory, you should be able bump the voltage, but as i dont know the motherboard, i cant really say for sure how!

dpDesignz
22-06-2011, 11:19 AM
thanks @Chilling_Silence. They do have onboard, but the onboard doesn't work. Would this just be because the graphics card is in there?

Good idea though thanks. Will give it a go.

@SolMiester I'll give it a go

Agent_24
22-06-2011, 11:42 AM
thanks @Chilling_Silence. They do have onboard, but the onboard doesn't work. Would this just be because the graphics card is in there?

Yes most boards disable onboard graphics when a separate card is installed.

Well at the temperatures you describe for the GPU I can't say I would point the finger at it, because my 8600GT as I said ran 75-80 degrees while gaming and never crashed.

I managed to get it up to 99 degrees and probably more as the temperature meter only went to 2 digits on ATiTool (similar to furmark) and it never crashed then either.

But then your 60 degrees is while idling.


Install SpeedFan and log the temperatures while gaming. Only then will you know what's really going on.

dugimodo
22-06-2011, 11:46 AM
Did you ever try furmark? I suggested it a while back. It will stress the Graphics card and graph the temperature at the same time. If it survives that it's not overheating. No game pushes a card as hard as furmark does.

My cards now top out at 80~85 degrees under furmark, which is acceptable (previously hitting 99 degrees until I changed mb - thermal limiting prevents them going higher).

Also have you tried updating or reinstalling video drivers? sometimes certain driver releases cause problems with some games.

dpDesignz
22-06-2011, 11:56 AM
@Agent_24 that temp was while running.

@dugimodo no didn't see that? Where can I download it?

dugimodo
22-06-2011, 01:57 PM
http://www.ozone3d.net/benchmarks/fur/

dpDesignz
22-06-2011, 02:01 PM
thanks. will try that as well