Sign in with
Sign up | Sign in
Your question

X1950 Pro fan problems and overheating

Last response: in Graphics & Displays
Share
February 13, 2007 2:11:26 PM

Hey

I bought an ASUS Radeon X1950 Pro recently and have had some problems with it. My screen would get patterns when having played a certain game for a few minutes and crash after that, I thought it was the game's fault because it didn't appear in other games. Now it looks like the other games just weren't graphics intensive enough, I've tried a few other demanding games and they have made my computer crash the same way and sometimes in other ways too, but always after having played for a while.

So I went to get ATITool and RivaTuner to see what's going on, and it's not a big surprise that the card seems to overheat quite quickly. I had to drop the GPU frequency down to about 2/5 of the default to keep the heat under control. Now, the reason for the overheating seems to be the fan, ATITool shows that its speed stays at 20% constantly, and RivaTuner says it to be at 19.57% (~1455 RPM). I find it strange that the fan itself works but the changing of the speed doesn't.

Before taking it to the store, I'd like to hear how the fan speed should correspond to the temperature, and what are the usual max and min speeds are. And of course any suggestions for possible solutions are welcome too.
February 13, 2007 2:46:39 PM

By default, the fan speed is very conservative because if it should run at real speed to counter temps, then it would be real loud.

So, I'm guessing your fan speed ratio isn't too great, and most people use custom fan speed settings.

The fact that it overheats quickly means that a.) cooler isn't properly seated or b.) your case temp is high due to inadequate cooling scheme (no offence meant).

I'd say, RMA the card and then check it again. If the issue persists, then look at the air flow in ur case.

Also, confirm exact temps if possible. It might also be memory overheating as opposed to gpu.
February 13, 2007 3:14:08 PM

I tried the 'scan for artifacts' test in ATITool and it ran for about a minute before artifacts started appearing, the temperature at that point was around 75 degrees celsius. It rose to 79 and then crashed my computer. I think I got a big higher earlier, and the artifacts kept coming for a longer period. This time there were only 2 wrong pixels at max, in the previous run there were 56 at best.

My case cooling relies on fans. The case is an Antec P150 and the fans in there are the CPU fan, PSU fan, GPU fan and the rear fan, they all seem to be working. Other than that, I can't say much about the cooling.

How do those temperatures sound?
Related resources
February 13, 2007 3:15:29 PM

Quote:
By default, the fan speed is very conservative because if it should run at real speed to counter temps, then it would be real loud.

So, I'm guessing your fan speed ratio isn't too great, and most people use custom fan speed settings.

The fact that it overheats quickly means that a.) cooler isn't properly seated or b.) your case temp is high due to inadequate cooling scheme (no offence meant).

I'd say, RMA the card and then check it again. If the issue persists, then look at the air flow in ur case.

Also, confirm exact temps if possible. It might also be memory overheating as opposed to gpu.


what he said also what is your Case ventilation like *how many fans and where
February 13, 2007 3:25:54 PM

Heh, slightly ahead of you :) 

Also, when not running any intensive applications, the GPU temperature seems to stay at around 45 degrees celcius.

The fans are the default ones that came with the parts. So the rear fan and the PSU (Neo HE) fan came with the case, the processor fan is what I got with this E6300 and the graphics card fan is also original.

I have now set the rear fan to high, previously it was on medium, doesn't seem to affect much.
February 13, 2007 3:43:20 PM

75C is my nominal using Arctic Cooling X2 - so it's definately your cooler. RMA it. I can run up to 90C with no issues, so if your at 75, sumthings wrong.
!