Sound Loops and Blank Blue Screen

[Ustauk]

I've been having an intermittent issue with my computer.  Every once in a while, while I'm playing games the computer will lock with a sound loop and a blank blue screen.  I've had this issue with Battlefield 2 and City of Villains, but not with Battlefront 2 to my knowledge.  I have the latest Radeon drivers. This usually happens after a fair bit of gaming, but can reoccur within a short period after rebooting.  It seems worse on Battlefield 2, where at one point I would crash after only a short time in the game.  Sometimes I have no crashes at all, yesterday being an example.  I've run Memtest86, the Prime95 torture test and the Sisoft sandra burn in, as well as done some media encoding, and haven't had any issues, so I don't think its the hard drive or RAM.  If it is the RAM, I'm wondering if its overheating because of the video card next door and is in high use, since Battlefront 2 works fine but it isn't overly RAM intensive.  Or it could be something wrong with the vid card itself.  Has anyone seen anything like this before?

[Mr. Analog]

I was using the wrong audio drivers, after I got the updated ones it worked just fine.

[Melbosa]

This is sometimes an IRQ conflict as well.  Try repositioning your Audio card in a different PCI slot.

[Mr. Analog]

Another thing to ensure is that if you are using a PCI sound card is that if you have a built-in soundcard on your mobo that it is turned off (both the jumper setting and the soft setting). Then log into Windows and make sure that you only see one audio hardware device installed.

[Ustauk]

I'll try and swap the sound card slot tonight.  That might be it, since I think I might have moved the sound card when I put in the vid card, to better position a new usb bracket I put in at the same time.



As far as I can remember, I disabled the onboard sound card in the bios, and there's no hardware jumper for it.  I'll double check tonight.



Thanks for the help.

[Cova]

It sounds like a thermal problem to me.  While the system is running pull the side of the case off and make sure all of your fans are actually spinning - especially small fans (chipset and GPU) die a lot without people noticing.  While your looking in there make sure that all your heatsinks are reasonably dust-free - a layer of dust is like putting a nice warm blanket over the heatsink.  Since it's occuring in games, and seems to occur more after heavy use, my first guess would be gpu overheating, with PSU overheating close behind.  Video ram problems tend to show up as video corruption before causing lockups, so they're probably ok.

[Ustauk]

I tried swapping the sound card slot in case it was an IRQ problem and reinstalled my drivers, but that didn't help. I downloaded ATITool last night and ran the artifact detection for a bit, but found none. I'll try and leave it running longer tonight. The highest temperature I observed using the tool was 75 degrees, the average was about 65, and the idle was 55. Does 75 seem high for a X850 Pro chip? And if I use ATITool to step down the clock speed to see if that helps, would that void my warranty?

[Lazybones]

GPU temps are hard to gauge without the specs for the chip.



What are your CPU and Chipset temps?



How old is your PSU? Power problems can often cause VERY random issues.

[Cova]

GPU's normally run hotter than CPU's and other chips, 75 probably won't cause any problems but its hard to say for sure.  Does that tool just display the temp on screen, or does it keep a high-low log that you can check back to after running a game.  It's amazing how fast chips cool back off when you remove the load from them - in the time it takes for you to alt-tab out of a game and go find that tool (all the while the gpu is practically idle again with no 3D on-screen) your GPU could be dropping 10 degree's - I've seen 10 degree drops in under 5 seconds when I removed the load from CPU's.  If you can easily do it, while your system has a load on the gpu (leave 3dmark or a game demo looping) touch the video card heatsink, and the back-side of the vid-card PCB right where the GPU is.  If you can't hold your finger there for at least a couple seconds you've probably got a heat issue.  If you burn yourself in under a second, whatever you touched is over 100 and you DO have a heat problem.  Don't worry - I use my fingers all the time to diagnose heat problems while overclocking and have yet to burn myself  You did confirm that all the fans are still spinning right?



I wouldn't worry about aritfact detection - you said you hadn't seen any in games yet (or at least I got that impression from somewhere), and if you did have some it would indicate a video memory problem (corrupted textures and geometry).  It's amazing how bad you can corrupt video memory and not crash - I've had games look so bad you can't tell what they are anymore, but they still run fine, whlie playing with over-clocking video memory and finding where my card max's out while being stable.

[Ustauk]

The tool returns the current temperature, as well as showing an average/min/max temperature reading.  It can also record a log file, so I'll attempt to recreate the crash tonight to see if I can record the temperature at crash time.  The weird part was I can get through 3Dmark 2005 no problem, and I think that really stresses the vid card.



I have a Sparkle 400 Watt power supply.  I'll see if I can get the model and specs tonight.



I'll disable EAX tonight on Battlefield and see if that helps too.  Maybe the load on the sound card is enough to put the heat over the edge.  The weird part is some Battlefield maps are fine, too.  I'll let you guys know how it goes.