Looking for PC Advice and Troubleshooting

These forums have a pretty knowledgeable and helpful group of people, so I'm going to ask for a little advice. My current rig was built for me less than two years ago by a friend who owed me some favors.

It's been a real workhorse, but I've started having occasional problems with it, usually (but not always) related to playing CoH. In about one out of four sessions the machine will cut straight to a black screen and reboot. I've checked pretty much everything I can think of, and can't determine the culprit. Any ideas?

Processor: Intel Core2 Extreme QX6850 @3.0GHz
Motherboard: Gigabyte X38T-DQ6
Graphics Slot: PCI-Express x16
RAM: 4GB DDR3 (Patriot PDC32G 1600LLK)
Main Disc: 320GB Western Digital 7200rpm
Graphics Card: NVIDIA GeForce GTX 295 (ASUS ENGTX295/2DI/1792MD3)
Power Supply: Turbo-Cool 1 Kilowatt (1100 W Peak)

For a monitor I'm using a Samsung 40" TV running at 1920x1600, if that makes a difference, although I can't imagine how it would.

Temperature-wise it runs a little hot, but the monitors indicate that the GPU peaks at about 70-72C, which should be okay. (That's fighting a room full of Carnies with all my shaders and particles and what-not cranked to max, too.)

Any ideas as to where I should start looking to track this down? Thanks in advance.

cutting to a black screen and rebooting would normally suggest an overheating error, which is certianly possible given the hardware in the computer.

One of the big problems with software temperature monitors is that they are not accurate. There's a reason most sites that do "serious" thermal testing run a lead right onto a processors heat sink. Even though your temp sensors in the computer may say everything's fine... it probably isn't. 70-72c is also really LOW for a GTX 295 that isn't being water-cooled. Okay, in all fairness, I've ever actually worked on 2 GTX 295's in the entire time they've been out... but on those two I had temp. reportings in the range of 170-185 F... and that's a bit above 72c.

Tracking down a thermal fail-over can be a bit hard to do though.

The easist way is to force a downclock on all components. You should be able to set lower clock speeds for the 295 through the Nvidia-Settings in the Windows control panel, and you should be able to downclock your processor in the bios.

If your system remains stable with lowered speeds, then there's your answer, you are reaching a thermal cut-out.

If the system still crashes with forced downclocks, we have other problems to think about.

Thanks, I'll try downclocking. After making that post I noticed an oddity about the way the builder set it up: the PSU's exhaust fan is right beside the GPUs intake fan. Not perhaps the way I would have structured it, myself.

I'll let you know if downclocking solves it. Thanks again!

Hmm, I'm less confident that it's a heat issue now. I logged into CoH from a cold boot, after the machine was powered down overnight. It blew up (i.e., instant power-interruption type reboot) before it even finished logging me into the game. After the reboot I was able to get in, fought a few mobs (on-line for maybe 15 minutes), and then it blew up on zoning via the train.

Could there be a power issue? I would think a 1000W PSU would be sufficient.

Well 1000 watts doesnt mean anything if one of your capacitors blew a cap. You may want to physically check the motherboard, and look for things that look like little batteries. Make sure none of them have a dome top (they should be flat). If one or more of them DOES have a pushed-up cap, or worse, it's leaking acid, you'll want a new mobo ASAP. Try not to use your computer at ALL, bad power going through your system can damage all your other components.

Thanks for the input, Steele. I don't see anything looking funky on the mobo, but I'm going to tear the box down completely and look for anything else weird. The most recent reset it was sitting idle without any running applications at all.

This is getting frustrating. I've been out of work for six months. I can barely afford to keep my account current. Building a new box will be beyond my means, and going back to playing on a laptop will su- will be less than optimal.

I updated the system from XP Pro to Win7 Home Premium, but it worked fine for a few months before the trouble started. I upgraded the video drivers at that time (but not to the card-killer driver they pulled back). The Windows crash analysis tool pointed out that there had been ~some~ driver failures, but upgrading to the latest-n-greatest doesn't seem to have changed anything.

I'm going to tear the box apart and look over all the components for signs of failure. I've already checked the fans and they all appear to be hunky-dory.

Thanks for the information!

I hope I can catch you before you take that step...not sure why no one has asked this yet.

Do you have it set to automatically restart on error? You should be seeing a blue screen of some type. Here's how you check:

Control Panel->System->Advanced System Settings->Advanced Tab->Settings button under 'Startup and Recovery'->Remove the checkmark beside "Automatically restart"

If you can see a blue screen, it might shed some light as to what's actually happening. If it just powers itself off or restarts without a blue screen and you have it set to not automatically restart, then we're looking for a hardware failure of some type in most cases.

Hope this helps.

We'll see....

You can also check the Event Viewer (Control Panel --> Administrative Tools --> Event Viewer). Error messages from past crashes ought to be recorded there as well which might point a bit to what is going on.

I had previously looked at the logs. The last 58(!) critical events are all the same, varying only by timestamp:

Log Name:      System
Source:        Microsoft-Windows-Kernel-Power
Date:          3/16/2010 1:20:28 AM
Event ID:      41
Task Category: (63)
Level:         Critical
Keywords:      (2)
User:          SYSTEM
Computer:      Nyarlathotep
Description:
The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

1000w power supply is more than sufficient for that rig. That doesn't mean the ps isn't going bad on you though.

Couple of things to check before you disassemble the whole machine.

1) Have you cleaned all your heatsinks/dusted the interior lately? Fuzz-bunnies and computers don't mix well.

2) You listed your GPU temps, how about your CPU though? If your cpu cooler has worked it's way loose or the fans are having troubles, that could cause an intermittant heat-fail.

3) Make sure all your capacitors look OK (on both MB and Video card if you can see them).

4) Double check all of your connections between the power supply and the components. Be sure to check the mb main connector and all the fan pins to make sure none have come partially off.

5) Try running with the case open so you can easily check and make sure all fans are spinning and aren't clogged with gunk.

6) Check your CPU cooler and Power Supply as well to make sure they clean and venting properly.

7) You can look at getting one of the various software monitors to track your utilization/heat levels/and voltages in real-time.

8) If you're the electrical type, check the actual output of your PS to make sure it's operating within specs. If you're not, DON'T mess with it unless/until you learn how to do so correctly (without ruining your PS or yourself via electrocution).

Did some searching on your error and while my search was hardly exhaustive, a common issue I did see was with the audio drivers causing this. In some cases, updating them fixed the issue, in others, they had multiple apparent audio drivers that were conflicting with each other. Disabling the extras seemed to fix the crashes.

In the first case, the common denominator seemed to be with Realtek audio and a search of your mobo indicates it has a Realtek audio chip. This may be your problem.

In the second, it seems at least some ATI cards have some sort of audio driver with them (not quite sure why) and disabling it fixes the conflict. As you have an Nvidia card, this may not be relevant. Not sure if Nvidias have such as well. In any case, you might look for additional audio type devices and try either updating or disabling them as well.

Keep in mind the above only *may* be your issue. It could easily be a HW problem of some sort but those seem to be common solutions that I found in short searching.

Interestingly, I run Win 7 and would technically fall into both categories (Realtek + ATI) but fortunately I'm not having your problem. (And I hope I never do.) Something to check out and see what you find.

Well, last night I tore the rig down to loose components and rebuilt it from scratch. (It needed to be done, anyway. As I mentioned, a friend built it for me so I needed to get a better feel for how it was assembled.)

In checking around and moving components (like putting the graphics card at the bottom of the case for better air flow, and taking out the never-used BD-Data drive), I found that the SATA port running the C drive may have gone bad. No visible issue on the mobo, but I had noticed an intermittent failure on init for SATAII01. I moved the drive to a new port and it ran for hours (including a couple of hours playing a grav controller) without a reboot.

Was that the problem? I don't really know. I ended up changing so many things it would be hard to be sure. But I wouldn't have even started looking in that direction if not for the assistance you people provided. Thank you all very much!

Now let's just hope it keeps working...

Good to see you seemed to have fixed it. Here's hoping you truly have.