Geforce GTX 280 Shutting Down, Red Light at the Back

[the_picard]

Hi there.

So I've had this problem for basically the entire existence of my computer minus maybe half a year. I bought this PC from Alienware (yes I know, I'm going to build my own when I need a new one and I've learnt my lesson) in I think 2008 and at first everything was going well. The processor is the Q9450 with liquid cooling which I have overclocked to 3ghz. This problem was happening before I overclocked so I don't see how it could be caused by that. I have 4Gb DDR3, an nForce 790i SLI (not ultra), 1000w power supply, a 300Gb Raptor hd, Windows 7 64bit (came with Vista 32bit) and I believe that's everything worth mentioning.

My initial and current theory is heat. When it first happened I didn't think much of it, just a freak incident, stuff happens. Then it happened a while later and again and again. So I felt the back area of my PC in it's place in my computer desk and it was really hot, and I tried blowing a fan to circulate the air out from there. To my surprise it seemed to work, but months later it started happening again. I then pulled my computer out of its desk slot and had the fan blowing directly on the back, which also seemed to work. Again more months pass and it starts again so I took off my case side and I just have a house fan that blows directly into my case now. Also at some point I installed a program to turn my GPU fan to maximum at all times. Occasionally after I restart my computer and forget to turn my GPU fan to max and play a game it will basically always eventually shut down the GPU. I have a temperature monitor and it seems like when it reaches 70 degrees it shuts down (which BTW seems outrageously low). I don't think I've ever seen my temperature go above 70 but I find it hard to believe that the card can't handle 70 degrees of heat, but the evidence seems to say otherwise. Also I realize 70 might not be the actual temp but whatever 70 really is on the monitor it never seems to go above it.

On the other hand all the information I have got through the interwebs seems to say that the red light turning on is because of power issues. Now my evidence seems to point to heat but I investigated the power possibility anyway. So I have 1000w of power and only 1 GPU plus the standard other components a typical PC might have, oh except an internet gaming card... surely that can't take up much power? Everything is how Alienware set it up and I haven't changed anything on the inside but I did notice that instead of the 6 pin connector and 8 pin connector for power to the GPU, it has two 6 pins with one of them being converted to 8 pins through a short adapter cable. IIRC a 8 pin connector is similar to a 6 pin connector but without the grounding (I think I remember that from the research I did like a year ago about this problem so apologies if I'm wrong or am generally getting some of my own specifics wrong).

So I'm asking for help because I don't think there is anything more I can do to cool my computer and temporarily delay this any further. On a thread about a similar issue someone said reducing their voltage fixed their problem. I'm sad to say that I've lost a lot of the knowledge I gained while I was trying to solve this problem before and I've lost the knowledge I had about overclocking... so how will lowering my voltage affect things on my computer and will it alter/hurt my overclocked processor in any way? I plan to try I just wanted to know what might happen if I do before I actually try. What else can I try or do I just need a new graphics card? I don't want to buy one and find out I have the same problem. Why did it seem to get better when I used more cooling on it?


TL;DR I think it shuts down because of heat, based on the evidence, but it seems the internet says it's a power issue. What can I try to fix it?

 

[AMutedScream]

Seem like you have an interesting problem there. The red light does indeed indicate insufficient voltage to the card. It cutting out at a reliable 70 degree's is a bit odd though. In my experience voltage problems often show instability during operation, rather than a reliable and predictable cutout point. That little adapter you mentioned going from 6 pin to 8 pin would be worth checking out. the 6 pin spec is rated for 75watts, the 8 pin is 150 watts. Now the 6 pin can handle more than 75, but the 8 pin was designed with additional grounding and lower resistance to support the increased wattage. Slapping an adapter to the end of it is not going to change the spec. Do you not have any 8 pin, or 6+2 pin, PCIe connectors coming off the power supply? If not, you may want to drop by a local shop and pick up a Y adapter that uses 2 6 pin connectors to feed one 8 pin, that would give you more juice to work with (and its a returnable 10$ part if it doesn't help). I'm assuming your 1000 watt supply comes with more than just the two PCIe connectors (most do). While I'm thinking about it, you might just want to try different PCIe connectors for power (if again my assumption that you have extra on your psu are correct). Some power supplies have more than one "rail" delivering power, and using a different rail may help as well. I hope these ideas help.

Also what people mean about "reducing voltage" is actually "reducing your cards need for voltage." IE removing or reducing any overclocking done to your graphics card, which will lower the power demand for the card somewhat. There's really no need to touch the settings you have on your processor, what overclocking you already have done to it shouldn't be enough to starve the rest of your system of power, with 1000 watts available.

 

[the_picard]

Best answer selected by the_picard.