Hey, guys!
I have a problem to solve and I can't quite figure it out by myself. Maybe you could help me with this. Let me preface a bit!
In March, 2011 I bought my current setup. In my signature, you can see the rough layout of my build. Maybe about a year in I started getting GPU crashes, albeit very, very seldom. Maybe once in a month or two. What would happen was my screen would go black and the picture would be restored in roughly 1 - 3 seconds. I didn't think too much of this, since it didn't really affect what I was doing. Over the course of last year, it had become a bit more frequent. This would happen let's say once in 3 - 4 weeks.
It was a beautiful evening maybe 6 or so weeks ago. I was laying on my bed with my PC running at the other side of the room. Suddenly, my ear would catch a tiny crackling sound that was clearly coming from the direction of my PC. I had been cleaning my PC from dust earlier the same day and I had forgotten to close the side panel, due to which every little noise the PC made was very audible. I was starting to be rather alarmed at this point and I pounced through my room to make a diagnosis. The sound was very faint and I though it was coming from the direction of my PSU. It was late and I decided to call it a night and shut down my PC.
The next morning, I observed the sound more. I tried pinpointing the location of the sound and finally determined that it had to be one of my hard drives. I thought that as long as no absolutely necessary component was malfunctioning, I shouldn't be too worried.
About 10 days ago, we had a LAN party with a group of friends. The crackling sound was now almost ever-present, whereas before it had only been occasional. For the first time in a long, long time, my computer froze. I wasn't doing anything specific at that time. Just surfing the web. My mouse wouldn't move but I could still all the sounds that were on. This happened I guess twice during the two-day party. The GPU errors would also happen a few times and a message along the following lines would be displayed - just as it had always: "NVIDIA xxx.x Kernel Mode Driver has stopped working, and has recovered".
I started growing uneasy and these problems would become ever more frequent. Yesterday I had two freeze-ups in two hours and a couple of driver malfunctions. Today, I've had three.
With this in mind, I started running test. Here are some results:
Windows Memory Diagnostic: Pass
Seatools short and long generic tests for SSD and two HDDs: Pass
Furmark: My video card GIGABYTE GV-N560OC-1GI GeForce GTX 560 Ti OC would run different sorts of tests but always the temps would go up to 95 degrees celcius. The device probably has a self-turn mechanism and the PC would reboot itself always, and only at this point. I'm not sure what to make out of these results. I would have thought that at some low benchmark settings, I would have gotten a stabilised temperature and a decent framerate. Or does Furmark always try to squeeze every last bit of power from the card.
Prime95: The program wouldn't run the Blend test at all due to hardware malfunction, at first. I tried twice. Then I tried the Small FFTs test, which would run okay for 30 minutes. Also Blend would run now and it worked fine for 90 - 120 minutes or so.
Today, I observed the sound again. I still can't tell if it was coming from the PSU or the GPU but I'm pretty sure one of them is failing. I decided to unplug the GPU in order to clean it up. The sound disappeared and I'm now writing this message with my CPU on-board graphics and I haven't heard a thing during this whole time. However, I'm not 100 % convinced. It might be that now that the PSU is under a decreased amount of stress, it is working fine and my rig has only been overloading the thing. My power supply is a Corsair Builder Series CX500 V2, 500 Watt. I've sometimes wondered that my PSU might not give enough power to the components although I haven't overclocked anything as of yet. Could I be running out of power here? I don't have any extra PCI-E cards. Just an SSD, two HDDs, GPU, CPU, case fans, 3 sticks of RAM, an aftermarket cooler and a fan controller.
I'm was going to upgrade my GPU at some point but I would like to invest that money to the component that is actually failing.
What do you guys think of this? Have you had similar problems? Can I run more tests to test out the GPU or the PSU? Can we draw a conclusion?
Thank you in advance your input!
-mon4ro
More about :
gpu psu failing mystery solved