Random Crashes-BSOD and Illegal Ops

[stealth_monkey]

Hi guys. I've been attempting to solve this problem for about 6 months now on and off, but I pretty much give up, I'm posting here (sorry if this is the wrong section) in the hope someone out there can help me solve it.

Problem: I'm having random crashes in games and out of games, some crashes are bsod's, some crashes are illegal operations and some are simply silent quits. The most common crash is that firefox will perform an illegal operation, and after this, will constantly perform illegal operations after the first within 5-10 seconds of restarting the browser. Initial I-op sometimes takes 5 minutes after boot, sometimes several hours or days. A restart solves this problem. When playing games, I'll usually get silent exits, but bsod's are fairly common. BSOD's generally point to either p16x.sys (file related to my sound card) or nv4.dsp (file related to my video card) as having IRQ not less or equal errors. These errors occur in both windows xp and ubuntu (instead of bsod's ubuntu programs simply silent quit). These errors are random, and so far I have no reliable way of reproducing them.

What I know it's not:
- RAM, ran an 8 hour block of memtest, 1 error (during the entire test, did not repeat) which I was unable to reproduce testing each stick individually. Swapped in a friend's ram, problem persisted
- Sound Card- Originally used AC97 onboard audio, upgraded to a sb live! 5.1, error persisted (although bsod's started mentioning the card as a culprit)
- Video Card- Upgraded from geforce 5600mx to 7600gt, problem persisted
- IRQ conflict- Moved soundcard pci slots, error persisted.
- Hard drive (probably)- scanned hard drive with xp disk checker, no errors, problem persisted.
- Audio Acceleration- Dropped audio to emulation only in dxdiag, problem persisted.
- Windows Install- Reinstalled windows, problem persisted.
- Virus/Spyware- Problem occured with clean formatted partition.


System Specs
PSU- Brand/model unknown. 400w, 1x 12v rail
Mobo- Asus P4S800-MX SE, socket 478
CPU- Intel Pentium 4 "Prescott" @ 3gz
Ram- 2x 512mb Kingston DDR PC3200 sticks @ 200mhz
VCard- Galaxy Geforce 7600gt 128mb, agp
HDD- Seagate 230gb SATA
OS- Dual Boot Ubunty Fiesty (primary) with Windows XP Pro SP2.

If you guys want any more info, just ask. I've got enough error reports to fill a truck with.

Thanks in advance for any help you can provide.

[pharoah]

have you scanned for viruses,and spyware?also that power supply may need replaced try this http://www.techsupportforum.com/arti...rsviewpro.html post back the temp,and voltage readings.

[stealth_monkey]

Hi Pharaoh, as I posted, I'm certain it's not a virus. I've formatted my hard drive twice with the problem occurring on a fresh install and it's affecting both a linux and windows partition.

Thanks for the tip about the sensorview program, output is here

[pharoah]

at 5% cpu usage from that screen shot,and the low 12v.id say its power supply.check the voltages again under load.

[stealth_monkey]

Created cpu load with prime95 torture test. All tests stop within minutes due to an error. Good news though is that if it isn't the psu, I've found a way to recreate the problem at will.

Output

I noticed that whilst undergoing the tests, the 12v rail (I was paying attention to that one since you mentioned it) would hover around 11.5v at 50% cpu cap, dropping down to 11.4 only occasionally , and for a few seconds at a time. Didn't find any set point the error would occur though, sometimes the voltage would be around 11.6 and stop, sometimes 11.4

[pharoah]

with the voltage dropping like that id say it is the power supply.

[stealth_monkey]

Thanks for the help Pharaoh. This, however, leads me to another problem. A phone call to my friends revealed none of them have a spare psu lying around that I can test with, and I'm unable to justify spending the cash for a new psu without being 100% sure that's the problem. However, I figure that logically, if my psu isn't up to scratch, I should be able to disconnect my dvd-roms (1 rom and a burner) and floppy drive and should be able to notice an increase in stability. Am I on to something or have I missed some fundamental part of psu operations (not saying that sarcastically, I'm one of many computer gamers who tend to ignore the little silver box unless it starts bursting into flames)

[pharoah]

lets take the test a bit farther.download rthdribl which is a graphics test.run that,and prime95 at the same time see how far that 12v drops then.with the cpu being maxed with prime,and the video card being stressed with rthdribl that will tell alot.i can already say with the voltage drop under cpu stressing alone.the power supply is dropping to below tolerance.

[koala]

Removing the ROM and floppy drives won't help if the +12V line is struggling. Open your case and post back with the amps listed for +12V. And a make/model if possible.

Enter your full system specs into this power calculator, and add 30% to the total to allow for PSU inefficiency. Is the total anywhere near 400W?

Can you give us some examples of the BSOD stop codes? As you know, nv4.dsp is related to your graphics card/driver, which would suggest an incompatible or corrupt driver. But having the same crashes in both XP and Ubuntu could mean a hardware problem with the card or motherboard.

[pharoah]

i agree koala but it looks like power supply.with the cpu at 50% his 12v is dropping to 11.4v.

[stealth_monkey]

Originally Posted by pharoah

lets take the test a bit farther.download rthdribl which is a graphics test.run that,and prime95 at the same time see how far that 12v drops then.with the cpu being maxed with prime,and the video card being stressed with rthdribl that will tell alot.i can already say with the voltage drop under cpu stressing alone.the power supply is dropping to below tolerance.

Output
Only difference that I noticed was that the 12v rail was constantly at 11.4, but would occasionally jump back up to 11.5 or 11.6 before settling down again

Can you give us some examples of the BSOD stop codes?

Sure thing, I'm not 100% sure what you mean. The standard error message is IRQL_NOT_LESS_OR_EQUAL, here's the output of event viewer:

Error code 100000d1, parameter1 0000005c, parameter2 00000009, parameter3 00000001, parameter4 f70040c4.

Error code 1000008e, parameter1 c0000005, parameter2 0b5c0001, parameter3 f2503b3c, parameter4 00000000.

Error code 100000d1, parameter1 00000044, parameter2 00000009, parameter3 00000001, parameter4 f700407c.

Error code 1000008e, parameter1 c0000005, parameter2 bf9c339e, parameter3 f23d3bdc, parameter4 00000000.

Error code 100000be, parameter1 bfb172f7, parameter2 0d755021, parameter3 f79aeb24, parameter4 0000000c.



As for the psu calculator, it turned up at 284w, when I factored in 10% due to psu age (6 month old psu, extremely heavy use) 313ish, ending up at 406w. Which means I'm almost positive it's the psu, since mine is only pulling at best 400w. I'm going over to a mate's place on the weekend and we're going to plug his 430w into my tower for a few hours, I'll let you guys know how it goes

[pharoah]

its power supply running prime,and rthdribl at the same time.is just about like playing a game,and the voltages droping to 11.4v.that is 99.9% likely the cause of the crashes.