PC monitors shut off and system hangs while playing 3D games, but sound continues – Diagnosis


Two days ago, I started running into a problem with my Windows PC:

  • The PC's two connected monitors simultaneously lose signal and go black (as though the PC had been powered off).
  • The keyboard's Numlock, Capslock, and Scroll Lights will become "stuck" in their current positions, as though the PC is hung. (For example, the Numlock light on the keyboard remains lit regardless of me pressing the Numlock key repeatedly.)
  • No keyboard input does anything. (Ctrl+Alt+Del, Ctrl+Shift+Esc, Ctrl+C, etc.)
  • However — Whatever sound/music the PC was playing continues to play, and the PC's fans continue running, so the PC hasn't powered itself off or rebooted itself.
  • Opening up the case, the graphics card is pretty hot to the touch.

I had this happen 3 times in one evening. In all cases, I was playing a game with 3D graphics when the problem occurred (Torchlight, Minecraft, Magic: The Gathering 2012, Avadon: The Black Fortress demo). I have yet to have the problem happen when I'm not playing a game.

This system has been running stable for about 2.5 years prior to this. I didn't make any changes to the system prior to the problem starting to occur.

System specs:

  • OS: Windows 7 64-bit
  • Processor: Intel Core 2 Duo E7200 Wolfdale 2.53GHz
  • Video Card: XFX GeForce 9800 GT 512 MB
  • Motherboard: Foxconn P45A-S LGA 775 Intel ATX
  • RAM: Corsair 4 GB (2x 2GB) DDR2-800 (PC2 6400)

Full specs: New PC 2008

Troubleshooting tried so far (the problem occurred again after taking each of these steps, one at a time):

  • Updated the video drivers with the latest drivers from NVidia's site.
  • Opened up the case and cleaned out the video card and processor fans (both were pretty dirty).
  • Installed and ran temperature monitor software. The processor idles at about 50 degrees C, and goes up to about 63 degrees C while playing a game (seems on the warm side, but not excessively so?). The software wasn't able to report the temperature of the GPU — not sure this particular GPU supports software temperature readout?

My initial diagnosis is that maybe the GPU is on its last legs (given that it seems to be running pretty hot, and the problem only occurs while playing 3D games). Does this seem likely? Or is it likely that this problem is caused by the processor, RAM, or motherboard? Or could this be a software issue of some kind?

Best Answer

Could you try running the GPU-Z software and monitor the temperature of the graphics card instead of the CPU. To me this sounds quite like the graphics card is overheating and potentially failing.

If the graphics card is above 80 degrees for a significant portion of time then I would consider opening up your machine, and checking for dust around the graphics card.

In my limited experience of graphics card temperatures:

  • 70 degrees = a bit on the warm side (would prefer it lower but not a big problem)

  • 80 degrees = potential problem

  • 90 degrees = real problem

  • 100 degrees = potentially fatal problem

On the "Sensors" tab you can check the two options "Log to file" and "Continue refreshing..." So that you have (at least some) data available after the crash. When you click "Log to file" it will ask for a location to save a file, so no need to go hunting for where it stores it.

enter image description here

Related Question