Page fault in nonpaged area

Status
Not open for further replies.

ondrej.castek

Posts: 13   +0
Hi,
sorry to bringing up thread similar to others, but I haven't found help in those others.
My PC (WinXPsp2, MB MSI 648MAX, GeForceMX440,P4,1x 512MB ram;all about 5yrs old, OS 2yrs not reinstalled) goes to BSOD after a short while of working.
It started after I restarted DSL modem with the PC running - PC got frozen - I restarted PC - got BSOD after while.
At first I thought of IP collision, so changed it. Didn't help, so I removed UTP cable, then net cards, went into safe mode and went back to a recovery point, uninstalled antivirus, installed new drivers for vga card, ide drives. None of these steps has helped, I tried a "normal mode" after each of them.
I don't suspect hw malfunction as in safe mode, either with and without network, the pc was fine.
I am adding minidumps to this post and plead for help...
(I really need to work at home)
Thank you for advices in advance.

(I use FF 2.0.0.7, IE6, Kerio personal firewall, Avast antivirus, Spybot SD)
 

Attachments

  • Minidumps.zip
    71.9 KB · Views: 5
Because you can get into Safe Mode normally, and your minidumps note a problem with the Nvidia video drivers, I suspect an Nvidia driver is causing the BSOD instability. When you are in the Safe Mode Windows uses generic video drivers not your Nvidia drivers...

Try reverting back to the older video driver. Is Windows fully updated to SP2, and subsequent critical updates?
 
Thanks Tmagic. I will try to do so. Unfortunately, I have only the original drivers on cd and the newest one from www, so it will be a jump back in time.
Weird is, that I haven't changed the vga drivers for ages. But the truth is, that I experienced bit strange behaviour may be connected with vga. The resolution of desktop has changed itself few times. Once also with the frequency of monitor.

I have also remembered, that the problems didn't started with "Page fault in nonpaged area" but "irql less or equal..." and the bsod reffered to different files such like win32.sys etc. I hope I remember it well.

As soon as I get to reinstalling the driver I will post result. THX ONCE MORE!
 
Ok ondrej.castek,
the Win32k.sys, NTOSKRNL.exe and such are Windows system files, and they can point to just about ANY unstable condition... Unfortunately, Nvidia and Windows have long been at odds
 
Hi Tmagic,
bad news. I installed the original driver (btw., the VGA producer is Manli). Helped for about 5 mins. At least the PC was able to send error messages to Microsoft. It didn't help much anyway, mostly the MS site said, that it is hw failure. Once (from 4 cases) it mentioned GPU and told me, that there is no solution available.
Then I tried to install the newest drivers and didn't help again.

I'd like to ask few more questions:
a) can we be sure that the cause is GPU
b) is it GPU itself or drivers?
- I tried the oldest and newest drivers. I don't have any other :-/. It is old vga (GeForce MX 440).
c) would it help to reinstall WinXP, install Vista or to change vga?

I am attaching new minidumps (6). Still, I can work alright in safe mode.
Sorry for bothering you again :-(.
 
So, I borrowed vga card from a friend. It was GeForce MX 2 400. I installed original drivers enclosed on cd. After that, I was able to work fine for some time, maybe 20 minutes. But then Firefox has colapsed (I don't know, whether it is connected with the root cause or just unsuccessful upgrade of FF to 2.0.0.7) and I got BSOD in about 2 minutes.
I tried to download and install new version of Avast antivirus, but it was impossible in normal mode (BSODs). So I did it in safe mode and tried to run it, but got BSOD even in the safe mode. Next time I changed the resolution to the lowest (this was strange, contrary to work with previous vga the safe mode had always the lowest resolution and probably 256 colors) and colors to 16 mil and even run screen saver so the display was black and hoped the antivirus test will finish. But it was unsuccessful again - BSOD.
Do you have any ideas? Tmagic or anybody?
 
Ok,
with all you have done, the latest minidumps suggest a corrupt file on the hard drive or an impending hard drive failure, a corrupt Nvidia video driver, or memory problems.

I would begin again by testing or swapping your memory. Safe Mode loads basic drivers and does not fail to show anything unless there is a catastrophic hardware failure or disk corruption... So:

1> Test or swap memory modules
2> Test the hard drive for errors/impending failures/ SMART status
3> Clear and reset motherboard CMOS settings
 
also one other thing. Look at the caps on the board and see if they are swelled or cracked on the top

http://www.badcaps.net/pages.php?vid=5

Most problems that I see in my work have to do with bad caps from boards 5-7 years old.

120329818_0e20370795_m.jpg
 
Nice post Firedawg1971,
I had an old IBM desktop that would post and parttially load Windows then stop and throw a BSOD. The electrolytics near the CPU socket were swelled and leaking. 12 in all were replaced... I don't usually do this because of the time involved. This IBM board was proprietary, so it was repaired. It's been well over a year, and it's still up and running.

It might be good for ondrej.castek to check for the swollen and leaking caps. If he finds none, replace the CPU Heatsink/fan. On older systems, this is the first thing I do. If he finds bad caps, trash the motherboard
 
Well, in Memtest-86 v3.1, 9th pass 16 errors in test no. 5. Removed Ram into bank no. 3 (from no. 1), running the tests again.
 
So I swapped the dimm banks and in third bank problems occurred in fifth pass and in second bank even in first. Always in test no. 5 if it matters.
If the RAM module is faulty, why did the PC work fine in safe mode? (except of last time with the borrowed vga card) Are the tests in Memtest significant?

I checked the motherboard as well and it looks fine.

I will test harddrives after weekend, have to leave the city now.
 
When you are in safe mode the memory usage is low. Therefor the memory was most likely not used in the bad locations on the stick. Do you have other memory you can test with the system? borrow friends? test one stick at a time also.

As for the vga card from your friend. go ahead and return that to him. and steal his ram :)
 
Thanks. I´ve got only one RAM module. I will try to "steal" another one somewhere. Finally, I think I would be glad if it was only RAM.
 
I managed to get another RAM module and time to test it only tomorrow. The RAM I borrowed and tested with Memtest-86 v3.1 showed errors in test no. 5 even in 1st pass. It was the same test, as before with my RAM module and I believe this borrowed module is fine as my friend has never had a problem with it.
Is it possible, that Memtest shows wrong results?
 
More facts

More facts: Both Dimms worked fine on another machine, tested with the same version of Memtest from the very same floppy.

That makes me think that wrong is motherboard or CPU. I wanted to update Bios, but there is no newer than the one I already have. MB seems to me to be more likely in fault and that would mean to buy another one, which can be quite a task for socket 478. Am I correct with these assumptions?

Or can it be also problem of power supply? How can I check, that the voltage and amperage are steady?
 
PROBLEM SOLVED
Well, the problem is gone. Solution was easy, my RAM was overclocked.
As I wrote in the problem description, its' evolution was:
"It started after I restarted DSL modem with the PC running - PC got frozen - I restarted PC - got BSOD after while."
But I forgot to mention, that after the restart I got warning that there was possible case intrusion and settings may have been changed. Because I didn't notice any changes in the posts screen, I didn't pay attention to it.
A week ago I reset the CMOS, as last attempt. The CPU was than underclocked so I had to change the FSB and saw, that there is wrong ratio of FSB and RAM frequency. I corrected it and since that it is all fine.
Thanks to all, who tried to help.
Hope it can help somebody else.
 
Status
Not open for further replies.
Back