Hard Light Productions Forums
Off-Topic Discussion => General Discussion => Topic started by: MP-Ryan on October 06, 2011, 12:20:21 pm
-
OK, fellow geeks, I need your technical advice. My desktop system has recently started some off behaviour. Four times in the last week, the system has frozen without warning, the monitor has begun flashing "analog-digital" like it's not receiving a signal from the video card, and the sound (if active) has begun looping. Three out of the four have happened while Team Fortress 2 was active. The only way to fix is to hit the reset on the tower. I've tried accessing it through the network while the monitor is flashing away, and it's not showing up on the network (leading me to think Windows has crashed), but when the system is rebooted the only entries in the event log refer to the unexpected system restart. There is nothing timestamped in any of the logs to indicate any sort of problem leading up to the reset.
I've updated a bunch of drivers since and cannot get the problem to replicate (despite running a series of high-load games last night), but the spontaneous arrival of the problem (without any software/driver updates beforehand to trigger it) makes me think it might be a hardware fault.
Does this sound like a failing video card or PSU to you folks? The original PSU was replaced with a fancier Corsair 18 months ago which has run flawlessly since (and still has 3.5 years left in warranty).
Appreciate any input anyone has. I'm contemplating buying a PSU tester if this keeps up.
-
Had the same problem and nobody figured out what it was. Someone told me to replace the CMOS-battery (which didn't do the trick). I ended up checking all connections and dusting of the hardware (and the registry).
But freezing + soundlooping is almost definitely a graphics-card overheating sign (well at least it was one on my end).
Now with everything cleaned and properly cooled - I encounter no more of those strange problems.
-
Forgot to mention that - the GPU temps are well within normal, and I've had none of the other usual indications (artifacts, sluggish GPU performance, elevated case temps, etc).
Also, I did use a can of compressed air in their last night just in case. I have an air compressor out in the garage that I routinely use to blow all the dust out of the case, but it's not all that dusty right now.
I may swap some of the molex connections and see if there's a fault there.
-
Could also be corrupted DirectX or openGL libraries, if it's crashing in game. Honestly can't remember which TF2 uses as I don't play.
edit: You might want to run a simultaneous Furmark and Prime95 as that would stress the PSU as much as gaming with full CPU usage would, especially if the GPU temps stay within reasonable ranges while running it. And that may save you the cost of a PSU tester and/or expose errors in the GPU.
-
I had a dying vid card cause all kinds of weird stuff a few months ago. I'd try swapping out the card if you have a spare around.
-
Update:
-MemTest came back with no errors and no crashes.
-Prime95 processor stress test came back with no induced crashes and no crazy temps.
-FurMark GPU stress test came back with no induced crashes and no crazy temps.
-OCCT CPU/GPU/PSU stress test came back with no induced crashes, crazy temperatures, or voltages way out of spec.
-After completing the tests, TF2 ran just fine for over an hour with no stupidity.
-Shut down system, opened case, no leaking capacitors, everything is seated just fine, no damage to PSU connections, blew out minimal dust with a can of compressed air.
-Booted system, checked all temps and voltages... everything was normal.
-Ran TF2, joined a server, connection timed out, sat in server browser trying to refresh (it wouldn't), and the crash happened.
-Event log is still blank except for noting an unexpected system restart (this timestamped after the crash), and unfortunately I did not have monitoring software active in the background this time.
New hypothesis (input appreciated!): Crashes are cause by a conflict with or dying onboard network card. Evidence: Every single crash has occurred during [intensive] internet activity (TF2, downloads), several programs have been timing out or dropping connections to remote servers (Steam, Mumble, TF2 - this started happening a month or two ago and hasn't gotten better with an ISP plan upgrade), and no games have caused the crash other than TF2 (which is currently my only multiplayer game).
An update was deployed for my network card through Windows update last night as optional, which I installed following the last crash. I plan to try things further today.
Anyone able to comment?
-
when having issue with games its not uncommon to blame graphics card/cpu/chipset overheats. network issues can break a game just as easily. ive had at least one instance of where a network driver obtained though windows update was incompatible with the hardware it was supposed to be for. if you could manually update the driver and see what happens, might fix it. check cabling (if wireless check antennas and signal strength, ive never had a wireless problem that couldn't be solved with a bigger antenna), make sure its not a router issue. could also try using an alternate network interface (like bluetooth or wifi, or an extra network card) and see if that works. if its onboard check the bios and see if there are any settings for it that you can tweak with.
-
You might want to check what people say about your motherboard online as well. I stupidly bought a motherboard which was discounted because it was faulty. One of the problems it gave me was that it's PCI and AGP slots were constantly conflicting with each other. So when I tried to run even Minecraft multiplayer, I would get a BSOD after about 5 minutes in. I could run more graphically intensive stuff easily, but if I tried multiplayer there was always a high chance of a BSOD.