Random Crashes

[ApolloX]

Hey Guys,

 

My computer for the last week or two has started crashing for almost no reason. I've run all the basic diagnostic stuff but come up with zip. *DESPERATELY* hoping someone can give me advice on how to resolve it.

 

Symptoms: First, the system locks up but the mouse is still responsive, then 1-2 seconds later mouse locks completely, a few seconds after that the screen goes black and often times (although not always) the computer starts beeping. Most of the time when it does beep its a single beep, although one time it beeped multiple times alternating between high and low pitches.

 

Frequency: Seems like it occurs if I use the computer for more than 20 mins at a time. If I leave it alone for hours it does crash. No specific program seems to be responsible. Sometimes there are days between crashes, other times minutes.

 

Temperatures: All low (30C). I did have the temperature monitor open during 2 of the crahses and nothing was hot. All fans are running and I've checked the seating of the CPU cooler.

 

Testing: I ran MemoryTest86 as well as 2xPrime95 and neither demonstrated any problems.

 

I can't even determine if this is a hardware or software issue. The way its crashing makes me think software but the fact that the computer is beeping and screen goes black, makes me think hardware. Anyone have any ideas as to what I should do next?

 

PS I've turned off all overclocking and using loose timings, but with no change. Also, nothing in the Windows Event Viewer records any errors.

[ApolloX]

UPDATE: I found it it was the nv4_disp.dll BSOD causing the problem. I assumed by the beeps and black screen there was no BSOD coming but I started waiting 20 seconds and it appeared.

 

Anyone know why this error would suddenly start appearing in normal windows behavior (heck, just using aim) very frequently out of nowwhere? Is it systemic of a hardware failure? I've tried reinstalling the drives but no luck.

[jo blo]

Hmmm, I was trolling for a solution to a similar problem I've been having for months. I've reinstalled to OS probably 10 times. Back at the beginning, I made several changes to my system almost simultaneously: New CPU (from 3500+ single core), new pair of WD drives in RAID-0, two new PCI cards (one now not installed), and updated BIOS. About a month ago I upgraded the RAM from 2x512 Crucial, and that hasn't improved anything.

 

The differences are that for me I only heard beeps at the beginning, months ago, and don't anymore, and mine never goes to a black screen. It just freezes permanently on the Windows screen. The striking similarity is that it's not at all clear if it's hardware or software. Like you, it is not load dependent, I've passed memtest and 2X Prime95 overnight. It doesn't seem to be caused by any specific program. In a common scenario, I open a bunch of tabs in IE or Firefox (it's oddly, definitely worse when these are Anandtech), moments later the tabs stop updating, but the progress gizmos keep spinning, and then after several seconds the mouse pointer freezes. The progress thingies can keep spinning forever though. The mouse pointer is pretty low level, but I have never heard before of a windows crash that doesn't result in a blue screen or reboot. There are no entries in the system log.

 

I upgraded my power supply several weeks ago from a 430W Seasonic, and it had been stable until last night. I thought I had solved the problem, but it froze (application, then mouse) when I cancelled an operation in DVD decrypter. The elapsed time counter in the application window went for about half an hour before freezing. This is damn frustrating. This is my first build and I wish I had bought a Dell instead. It has turned into a huge time and money sink. It's demoralizing. I was going to disable the RAID and boot from a single drive, but I see that that's not your problem. Next I think I'm going to downgrade the BIOS, and if that fails, revert to the old single core CPU. It has been a problem both with and without all the AMD/MS Dual core updates.

 

Anyone? A way to nail down a hardware vs. software issue would be momentus.

[ApolloX]

Wish I knew. Also wish I knew why this suddenly started happening with some frequency (1-2 a day variable) out of nowwhere. Anyone?

[ApolloX]

I should also mention that sometimes after the crash, windows fails to boot with the message "NTLDR is missing". I have to do a cold restart in order to get the system to boot after that. Anyone have any ideas?

[tasr]

First thing that comes to my mind is memory timings. Then browsing your signature a AeroCool Turbine 550W in a Expert, dual-core, 7900XX and 3 HDD. You need a lot of good clean power.

 

PSU Brands to Avoid!

 

For a DFI Expert are all three (3) of the power connectors plugged into the motherboard.

24-pin main power connector

8-pin 12v secondary connector or EPS

4-pin 5v/12v floppy connection or FDD-type

[jo blo]

Re: memory timings, if that was it, we should get errors in Memtest, right? I've run memtest overnight and window memory tester for hours with two different sets of ram, and I've had no errors. I ran 2X Prime95 overnight with no cpu errors. I put a Fluke multimeter on my power supply and measured all rails at poweron, startup, heavy load, idle, and shutdown, and all was rock solid. Then I bought a new power supply anyway, because I didn't have the "official" 480W+. Still have the problem.

 

Just happened to me tonight, the second time since the new PSU. This time it nuked Windows. I had nothing open except two Windows explorer windows of folders of digital images. In one of them I was looking at the pics in filmstrip mode. I was rotating a pics one by one as needed as I went through the images. Then it gets slow. Other applications respond fine while I wait for the hourglass. I close them. Finally finishes. Rotate another, hourglass, seconds later no mouse response. Frozen. There was nothing else running. No audio, no downloads that I know of (maybe win update or something similar, wasn't monitoring it). Upon reboot machine gets stuck at Windows scrolling green bar bootup screen with no HD activity. Reboot, NVRaid shows ERROR. Reboot again, Nvraid showing healthy but have to set Nvraid as boot disk in BIOS, reboot again. Stuck at bootup screen again with no HD activity. Try to boot in safe mode, it freezes at the first screen where the screen is full of file paths of dll's and things. Looks like I'll be reinstalling Windows for the 15th time.

 

Before the crash, there was something screwy with discrepencies between folders that were showing in the folder tree on the left and what was showing in the main window. Usually this refreshes when you navigate somewhere else and come back, but this time it wasn't. I was also getting file not found errors with images I was trying to move from the main explorer window, that I was seeing in filmstrip mode right in front of me. I messed around with it for a while, closed and reopened explorer or something and got around it. But wierd. That and the one time raid status ERROR has made the Nvidia IDE controller/drivers a prime suspect in my mind. I know they're WHQL, and I thought that was the end of the story, but I know of Nvidia's poor IDE driver stability in the past. A guy in this thread bluntly says to just get rid of the Nvidia IDE drivers to make a similar problem go away: http://www.diy-street.com/forum/showthread.php?t=72266

 

ApolloX, how did you determine that nv4_disp.dll is the culprit? So did you try using different video drivers to see if your problem went away? Are you using the nvidia IDE drivers? Since you're not using raid, it would be easy for you just uninstall them to see if you're problem went away. I still need data off a different partion of the RAID array, so the next windows install still needs to be on the existing array, so I won't be running without the nvidia ide drivers soon. Hopefully within a week though. After the 16th install of Windows.

 

A guy here http://forums.winamp.com/showthread.php?threadid=72895 had a similar problem whenever winamp was installed. Not running, just installed. Turned out it was crappy drivers for his VIA chipset motherboard. That's the only thread anywhere where someone conclusively solved this problem. Try a google search for "Windows lockup freeze". A bunch of people crying for help mostly, but evidently it is a problem that's been around. I think I read somewhere on the forums that a guy solved a similar problem by unplugging case fans and his DVD drive. Not exactly a solution, really.

[ApolloX]

Actually I forgot to change my signature, I'm upgraded my PSU to OCZ 600 awhile ago.

 

I've determined its definitely related to the video although whether its hardware or software I cannot tell. A few minutes ago the screen went black but instead of getting a BSOD it returned to the OS. About 5-10 seconds later the screen went black again. This continued about 20 times before (going black for 1 second then recovering for 5) before I decided it was time to reset.

 

Since the system never fully shut down (mouse/keyboard were responsive) I was able to conduct some ad-hock tests. First off, when I dropped to a dos prompt and full-screened the flickering stopped all together. When I went back windows gui they returned. I tried doing something simple like a screen saver (3d flower box, 3d text) and the results I saw were odd. Aside from the continued shutting down every couple of seconds, the graphics while they were being used were hicupping back and forth. So for the flower box I saw jittery images. Its definitely related to the video card at this point. I got luck in that it did not fully shut down.

 

Did my card go bad or do I need to reinstall windows? I tried switching drivers but no luck.

[wizurd of awwz]

What is your PCIe frequency in BIOS?

Which slot is the card in?

Ran any DirectX tests?

[ApolloX]

What is your PCIe frequency in BIOS?

Which slot is the card in?

Ran any DirectX tests?

 

100,1,Yes

 

The tests run just fine under normal conditions but then at some point (anywhere from 1-40 hours) the video dies, the screen goes black, and the machine shuts down with an nvp BSOD. That's been the consistent shape every time except the last time where the screen went black, but then revived, only to go black every few seconds until I finally restarted. At that point I'm sure directx tests would have failed since I couldn't even run a screensaver normally.

[ApolloX]

First thing that comes to my mind is memory timings. Then browsing your signature a AeroCool Turbine 550W in a Expert, dual-core, 7900XX and 3 HDD. You need a lot of good clean power.

 

PSU cleanliness is highly over-rated, especially on this forum. That being said I replaced the psu last year so nope, not that.

[Gunlance]

have you tried new nvidia drivers? Chipset and video.