PDA

View Full Version : PC possibly overheating?



Richardd150
11-12-2010, 09:59 PM
Hi all.

Over the last few weeks my PC has started misbehaving.

Playing games (Fallout 3 and Sacred 2) the games will crash after playing a while - Sacred just kicks me out and then restarts from where it was, Fallout starts getting graphic glitches before dying. Both games start running very slowly before this. I do get an error message in my taskbar, which say's something about the vid card driver having failed and restarted succesfully.

I have been monitoring temp with coretemp32 and have established this is happening when the cores get around 61 degrees (don't know about the graphics card temp) so had put it down to warmer weather, although never had this problem before. PCWizard confirmed the temps. Gave the inside of the case a good clean and carried on.

Problem is that the issue is continuing and now I getting the occasional:
-complete shutdown (PC turns off)
-Instances of screen going blank with a message that screen is auto-adjusting, and needing a hard reboot to fix

I've started thinking that maybe the vid card is on it's way out (hasn't been so warm here recently) or perhaps the PSU could be failing but hey, what do I know? By the way the PSU is a 250watt jobo that cam with the system.

Any suggestions as to how I could narrow down the cause of the problem? I'm fairly confident it's not malware or virus related as I've scanned with Malwarebytes, MSE and SuperAntispyware.

Ta very muchly in advance:)

WarNox
11-12-2010, 11:39 PM
Try running Prime95 (http://files.extremeoverclocking.com/file.php?f=103), that should stress your RAM/CPU so in turn the motherboard. If the PC doesn't crash you'll know the problem is the Graphic Card. As far as I know Prime95 doesn't stress the GPU.

Screenshot of my PC with Everest Ultimate, no game running.

qazwsxokmijn
11-12-2010, 11:43 PM
You're running a dual core 5000+ AMD and an X1950 Pro off a nameless 250W PSU.....I would say failing or not you ought to replace it.

pctek
12-12-2010, 08:06 AM
I do get an error message in my taskbar, which say's something about the vid card driver having failed and restarted succesfully.

By the way the PSU is a 250watt jobo that cam with the system.
)

OMG!!! I'm with qazwsxokmijn. You cannot use such a lame PSU even with your older equipment.


The GPU will be working hard, as you say, you get an error about it, that could be a problem with the card (but I think not) but for sure get yourself a decent PSU.

never use case bundled PSUs. Stick with quality brands, Corsair, Enermax, Antec.

Get something around 500w.

Richardd150
12-12-2010, 09:14 AM
Try running Prime95 (http://files.extremeoverclocking.com/file.php?f=103), that should stress your RAM/CPU so in turn the motherboard. If the PC doesn't crash you'll know the problem is the Graphic Card. As far as I know Prime95 doesn't stress the GPU.

Screenshot of my PC with Everest Ultimate, no game running.

Thanks, I'll give that a go.

Trev
12-12-2010, 09:17 AM
Get a new power supply unit. A 250watt one is overtaxing the system. Get at least a 500watt one.
:)

Richardd150
12-12-2010, 09:18 AM
OMG!!! I'm with qazwsxokmijn. You cannot use such a lame PSU even with your older equipment.


The GPU will be working hard, as you say, you get an error about it, that could be a problem with the card (but I think not) but for sure get yourself a decent PSU.

never use case bundled PSUs. Stick with quality brands, Corsair, Enermax, Antec.

Get something around 500w.

Thanks qazwsxokmijn and pctek. I was considering upgrading the PSU but was looking for some good advice - I got it! Thanks for the advice on the PSU brands.

Richardd150
12-12-2010, 02:52 PM
Get a new power supply unit. A 250watt one is overtaxing the system. Get at least a 500watt one.
:)

Boy was I misquoting the wattage, although not by much! It's actually 305 watts... but I've just put in an order for Antec EarthWatts 650W PSU ATX12V from Playtech anyway, as I thought it was a good idea.

Richardd150
14-12-2010, 08:46 PM
Try running Prime95 (http://files.extremeoverclocking.com/file.php?f=103), that should stress your RAM/CPU so in turn the motherboard. If the PC doesn't crash you'll know the problem is the Graphic Card. As far as I know Prime95 doesn't stress the GPU.

Screenshot of my PC with Everest Ultimate, no game running.

Hi WarNox.

I went ahead and replaced the PSU as a starting point on the good advice of fellow forum members and it hasn't made any difference, possibly even getting worse - I figure I needed it anyway on my first build in a year or so's time or if I upgrade my vid card.

I used Prime95 for a stress test and thankfully there was no shutting down going on. The cores peeked at 70c and hovered around 68c until I killed Prime95 - idle temps are around 38-41c.

From what I've been able to read the idle temps look ok but it doesn't seem as clear with the upper temps, is 68-70c too high? When I game the core temps get to around 61c.

If that's not too high then I suspect that my vid card (which I got 2nd hand a couple of years ago) may be on the way out and has become overly heat sensitive.

Cheers and thanks for your advice:)

Richardd150
15-12-2010, 02:37 PM
Hi all.
Anyone else able to contribute anything on appropriate tempature levels?

dugimodo
15-12-2010, 02:44 PM
it's sounded like a graphics card problem to me from the first post, replacing that PSU was a good Idea though.

60-70 degrees for a cpu core is not dangerous and shouldn't cause a crash, at around 80 I start getting nervous. Most motherboard & CPU combos will shut down if the temp get's too high, which is not quite what you describe.

When you cleaned out your case did you also clean the heatsink of the graphics card ? did you check the graphics card fan is spinning freely and not making odd noises etc ?

Also sometimes windows just gets bloated and full of crap over time and gets unstable, yours does sound like a hardware issue though.

EDIT: - A crude kiwi shed type troubleshooting method for PC overheating problems. Take the side off, get a desktop fan and sit it blowing air into the case, see if the situation improves.

nedkelly
15-12-2010, 02:59 PM
A crude kiwi shed type troubleshooting method for PC overheating problems. Take the side off, get a desktop fan and sit it blowing air into the case, see if the situation improves.

Hey thats my xbox 360 cooling system

SolMiester
15-12-2010, 03:33 PM
OMG!!! I'm with qazwsxokmijn. You cannot use such a lame PSU even with your older equipment.


The GPU will be working hard, as you say, you get an error about it, that could be a problem with the card (but I think not) but for sure get yourself a decent PSU.

never use case bundled PSUs. Stick with quality brands, Corsair, Enermax, Antec.

Get something around 500w.

I had a x1950 pro (30a required on 12v rail)and had issues running that on a TT460w, so I am surprised you had it running at all. You could well have stuffed the card running like that...

Richardd150
15-12-2010, 03:46 PM
it's sounded like a graphics card problem to me from the first post, replacing that PSU was a good Idea though.

60-70 degrees for a cpu core is not dangerous and shouldn't cause a crash, at around 80 I start getting nervous. Most motherboard & CPU combos will shut down if the temp get's too high, which is not quite what you describe.

When you cleaned out your case did you also clean the heatsink of the graphics card ? did you check the graphics card fan is spinning freely and not making odd noises etc ?

Also sometimes windows just gets bloated and full of crap over time and gets unstable, yours does sound like a hardware issue though.

EDIT: - A crude kiwi shed type troubleshooting method for PC overheating problems. Take the side off, get a desktop fan and sit it blowing air into the case, see if the situation improves.

Thanks! I did a general clean out of the case but I'll give the card some extra TLC.

Richardd150
15-12-2010, 03:48 PM
I had a x1950 pro (30a required on 12v rail)and had issues running that on a TT460w, so I am surprised you had it running at all. You could well have stuffed the card running like that...
And possibly have:o Strangely, it ran fine for a couple of years.

SolMiester
15-12-2010, 03:49 PM
Thanks! I did a general clean out of the case but I'll give the card some extra TLC.

You well have to do more than that Im afraid!.....unless you dont mind taking out the rest of the system.

Overdrawing on that PSU has probably weaked it to that state where you are getting your restarts. It wont stay like that for long, next time it may spike and take the mainboard and whatever else with it......

Upgrade the PSU immediately!

Richardd150
15-12-2010, 03:56 PM
You well have to do more than that Im afraid!.....unless you dont mind taking out the rest of the system.

Overdrawing on that PSU has probably weaked it to that state where you are getting your restarts. It wont stay like that for long, next time it may spike and take the mainboard and whatever else with it......

Upgrade the PSU immediately!

I agree, and did so in posts 8 & 9 (Antec EarthWatts 650W PSU ATX12V). I'm hoping that'll do me for a while:)

Paul.Cov
15-12-2010, 06:27 PM
Dugimodo's suggestion is a good one. Running without the case side in place will reveal a lot more fan noise, but you can also check that the GPU fan is indeed turning quickly.

My old system was used without the case side for over 12 months when I had concerns about heat. Removing the side easily dropped temps by 15 degrees... if memory serves.

If you can game without breakdown with the side off, then you've probably narrowed it down to GPU heat at this stage of things.
Might be worth looking for signs of bad capacitors on the mobo and graphics card as well.

Richardd150
16-12-2010, 02:49 PM
Dugimodo's suggestion is a good one. Running without the case side in place will reveal a lot more fan noise, but you can also check that the GPU fan is indeed turning quickly.

My old system was used without the case side for over 12 months when I had concerns about heat. Removing the side easily dropped temps by 15 degrees... if memory serves.

If you can game without breakdown with the side off, then you've probably narrowed it down to GPU heat at this stage of things.
Might be worth looking for signs of bad capacitors on the mobo and graphics card as well.

Hi ya. The fan is spinniing sweetly, no problems there. I touched (after grounding) several parts of the card and it didn't seem particularly hot.

Removing the side cover initially dropped the idle temp to 35c (troughed at 33c) but it gradually moved back up to the 38-40c level. Under stress testing still hovered around 68c with a spike to 70c - no real change there.

Funnily enough the "display driver failed and has recovered successfully" error hapened while I was reading PF1, so it's not just during games and under load.

Playing fallout 3 went alright (no crash or slow down but there was the odd colour 'bloom' on screen) with temp hitting 60-61c, Sacred 2 played for a little longer than usual then the PC re-booted without warning (no error messages).

By the way I'm testing the temp with PCWizard and Coretemp and they agree on the results - the temps are from the processor cores, the card doesn't show up so I assume it doesn't have a sensor.

So, a little unclear as to whether a card, MB or processor issue so time to trouble shoot.

Overheating? with the tests I've done and the advice I've been given it doesn't seem likely.
PSU prob? Doesn't seem likely with new PSU.
So next I'll put my old card (radeon x1300) back in and see what happens, as well as checking the caps.

dugimodo
16-12-2010, 03:50 PM
Speedfan was always a good program to detect graphics card temperatures, haven't used it for a while though so not sure if it's still around. Sometimes the driver software itself will tell you the temperature as well.

Swapping the Graphics card does seem like the best option right no as you have one to test with, should pretty quickly prove it one way or the other.

SolMiester
16-12-2010, 03:54 PM
Dugimodo's suggestion is a good one. Running without the case side in place will reveal a lot more fan noise, but you can also check that the GPU fan is indeed turning quickly.

My old system was used without the case side for over 12 months when I had concerns about heat. Removing the side easily dropped temps by 15 degrees... if memory serves.

If you can game without breakdown with the side off, then you've probably narrowed it down to GPU heat at this stage of things.
Might be worth looking for signs of bad capacitors on the mobo and graphics card as well.

You know, I really cant recommend taking the sides off a PC as a solution....PC cases are design for airflow from front to back and some with positive\negative pressure so air passes over the components rather than just the components being exposed to fresh air.

dugimodo
16-12-2010, 04:05 PM
I never really intended it as a solution, just a diagnostic tool. Yes I agree about airflow but only in decent cases, many older machines and cheap generic cases have no airflow to speak of with maybe a single 80mm fan at rear if that.

By taking the side off and directing a desktop fan into the case you can generate much more cool air into the components than most cases manage. I wouldn't recommend running like that long term but it could possibly be useful to prove wether or not temperature was the problem

SolMiester
16-12-2010, 04:23 PM
Its not overheating...he has killed the PSU by overloading it, and probably stuffed the GPU into the bargin!

Richardd150
16-12-2010, 05:12 PM
Its not overheating...he has killed the PSU by overloading it, and probably stuffed the GPU into the bargin!
It's not dead, just resting. The card, on the other hand may well be on it's way out. That's what I'm trying to figure out.

SolMiester
16-12-2010, 05:21 PM
Well, it is hard to diagnose hardware issues from a web site, however my experience is screaming the PSU has been fried due to excessive overloading from the GPU, and this usually has a detrimental effect of the card also as it cant get the power to run at the required amperage...
If you attempt to use the PSU again, I'd say your a fool!...as for the card, with another PSU, it maybe fine for 2d, but not 3D

Richardd150
16-12-2010, 05:47 PM
Well, it is hard to diagnose hardware issues from a web site, however my experience is screaming the PSU has been fried due to excessive overloading from the GPU, and this usually has a detrimental effect of the card also as it cant get the power to run at the required amperage...
If you attempt to use the PSU again, I'd say your a fool!...as for the card, with another PSU, it maybe fine for 2d, but not 3D
Thanks SolMiester, it's refreshing to know somebody is prepared to call me names if I do a foolish thing:)
The old PSU is dead, deceased, departed, gone, an ex PSU, as it were. I will not attempt to nail it to the perch.

Richardd150
17-12-2010, 08:17 PM
Well... The award for culprit in my PC woes goes to... Capacitors!!! (on the graphics card). Plus possibly my abuse of an underpowered PSU.

Three of them are bulging and split and I don't know how I managed not to see them before - I noticed them when I was removing the card to replace it with the X1300 (which is running the games fine with nary a sign of crash or a driver failure)

Now to persuade the wife that I need that HD5850 or HD6870 as a replacement ($400 odd compared to the $1000 to 1,200 I wanted to spend on a new case, ram, motherboard and CPU and graphics card!)

Thanks all for your advice.