When I was playing a game of UT2k4 the other night the sound cut out. The game didn't exit however which I thought was odd. Anyway I tried to restart the game to see if that would fix it but the game would not even start up again. I've had BSODs in the game before due to what I believe is a driver issue (I had a suspicion it was to do with Cool'n'Quiet which I have enabled). Anyway, I thought I'd check my graphics card temp too see if anything was wrong there only to find the core was about 90C. 
Since last night I've been monitoring the temp to see if maybe it was a one off and the problem would go away. Wishful thinking I know and it turns out the problem hasn't gone away. I've been monitoring the temp while the computer is on the desktop and I've measured readings ranging from about 60C to 75C idle! I just now had dinner while leaving this computer on and when I came back and moved my mouse to disable my screensaver I noticed that the core is running at 96C and the ambient is 58C. This soon settled down however to about 69C/43C.
Now my processor temp has been fairly normal (just under 30C while idle with Cool'n'Quiet enabled), so I'm led to believe that the graphics card is at fault somehow. I can remember a while back when I was measuring temps that my 6800GT would hit somewhere in the 80s under maximum load, but I've never seen the GPU core this hot before especially idle!
I've used about half a can of compressed air trying to clean up as much dust out of my case as possible (I've ran out now), but I've seen no decrease in temps at all.
My comp is running in a Midi sized Chieftec Dragon which has 4 case fans. 3 of them of the regular 80mm Ystech variety and the other a low cfm one that came with the case. I have this one running at full power pointing at the graphics card. I have one of the others placed blowing air over the HDD at the front of the case and the other two at the back drawing air out. These three are connected to the fan-only connectors on the power supply. This means that the power supply controls the speed of the fans. Now yes I know that's probably not the best thing to do as I believe the temp reading comes from the power supply itself, but I've managed like this for a while now and everything has ran fine for just under a year now (I had a previous set up before, but changed it as I overused the fan-only connectors and buggered the power supply up :/) . Today I switched the front and one rear fan to full power and connected the others to fan-only connectors. Can't remember exactly, but I didn't really have any significant drop in temps and the fans didn't half make a racket at full speed!
I'm not quite sure how long the GPU has been running this hot for, but I did start noticing the same problem in UT2k4 a few weeks ago, but I put this down to a driver issue (might still be and the temp problem is completely unrelated).
The 6800GT is running with the stock cooler as is my processor (more specs in my sig). I also have a Antec Truepower 430W PSU and I'm currently running the 81.95 set of drivers with my 6800GT. I've also booted in to Ubuntu Linux to see if it was a driver problem, but I'm getting similar idle temps there as well (not sure what version the drivers are, but they're the latest ones available on the Synaptic package manager)
Now I'm obviously quite worried about this and last night I was looking at after market GPU coolers and CPU HSFs (been meaning to try to quieten this rig down for ages anyway as it's quite loud imo). However, the stock cooler for the 6800GT shouldn't be this lousy anyway, so I'm probably ignoring the problem at hand.
I can't really afford to take my computer out of action as I have two pieces of uni work to do this weekend and another two next week, so I'll be using my computer a lot over the next week if only for programming and using word. I can go on to my Uni's campus and do the work if I have to if it means saving my computer, but I obviously don't want to do this not unless I have to as it's a bit of hassle traveling to and staying on campus all day.
I don't think my card is under warranty anymore (had it about two years now), so I can't RMA it. Like I said this couldn't have happened to me at a worse time as I'm under enough pressure as it is trying to get a load of work I left done in time without having to worry about my computer as well.
So is there anything I can try short of replacing the card to fix this problem (I'm the only person in my house with a desktop, so there is no chance of trying the card in another machine)? Any constructive help/advice would be most appreciated.
Since last night I've been monitoring the temp to see if maybe it was a one off and the problem would go away. Wishful thinking I know and it turns out the problem hasn't gone away. I've been monitoring the temp while the computer is on the desktop and I've measured readings ranging from about 60C to 75C idle! I just now had dinner while leaving this computer on and when I came back and moved my mouse to disable my screensaver I noticed that the core is running at 96C and the ambient is 58C.
This soon settled down however to about 69C/43C.Now my processor temp has been fairly normal (just under 30C while idle with Cool'n'Quiet enabled), so I'm led to believe that the graphics card is at fault somehow. I can remember a while back when I was measuring temps that my 6800GT would hit somewhere in the 80s under maximum load, but I've never seen the GPU core this hot before especially idle!
I've used about half a can of compressed air trying to clean up as much dust out of my case as possible (I've ran out now), but I've seen no decrease in temps at all.
My comp is running in a Midi sized Chieftec Dragon which has 4 case fans. 3 of them of the regular 80mm Ystech variety and the other a low cfm one that came with the case. I have this one running at full power pointing at the graphics card. I have one of the others placed blowing air over the HDD at the front of the case and the other two at the back drawing air out. These three are connected to the fan-only connectors on the power supply. This means that the power supply controls the speed of the fans. Now yes I know that's probably not the best thing to do as I believe the temp reading comes from the power supply itself, but I've managed like this for a while now and everything has ran fine for just under a year now (I had a previous set up before, but changed it as I overused the fan-only connectors and buggered the power supply up :/) . Today I switched the front and one rear fan to full power and connected the others to fan-only connectors. Can't remember exactly, but I didn't really have any significant drop in temps and the fans didn't half make a racket at full speed!
I'm not quite sure how long the GPU has been running this hot for, but I did start noticing the same problem in UT2k4 a few weeks ago, but I put this down to a driver issue (might still be and the temp problem is completely unrelated).
The 6800GT is running with the stock cooler as is my processor (more specs in my sig). I also have a Antec Truepower 430W PSU and I'm currently running the 81.95 set of drivers with my 6800GT. I've also booted in to Ubuntu Linux to see if it was a driver problem, but I'm getting similar idle temps there as well (not sure what version the drivers are, but they're the latest ones available on the Synaptic package manager)
Now I'm obviously quite worried about this and last night I was looking at after market GPU coolers and CPU HSFs (been meaning to try to quieten this rig down for ages anyway as it's quite loud imo). However, the stock cooler for the 6800GT shouldn't be this lousy anyway, so I'm probably ignoring the problem at hand.
I can't really afford to take my computer out of action as I have two pieces of uni work to do this weekend and another two next week, so I'll be using my computer a lot over the next week if only for programming and using word. I can go on to my Uni's campus and do the work if I have to if it means saving my computer, but I obviously don't want to do this not unless I have to as it's a bit of hassle traveling to and staying on campus all day.
I don't think my card is under warranty anymore (had it about two years now), so I can't RMA it. Like I said this couldn't have happened to me at a worse time as I'm under enough pressure as it is trying to get a load of work I left done in time without having to worry about my computer as well.
So is there anything I can try short of replacing the card to fix this problem (I'm the only person in my house with a desktop, so there is no chance of trying the card in another machine)? Any constructive help/advice would be most appreciated.
Last edited:
Don't have any compressed air left anyway so it would have been a job to get rid of the dust under the plate. I think one of my housemates mentioned the other week that he has large set of screwdrivers that he uses for electronics stuff, but he's out at the moment and I can't get hold of him... Hopefully he has them with him and has the one I need, so I can get this sorted asap. 
) and I couldn't believe how much dust was clogged under there! The heatsink was absolutely caked with dust, so it's no wonder I was getting high temps. Just played a few games of UT2k4 and the core reached a maximum of 65C while the ambient was stuck around 40C. Currently idling now at 52C/33C.