Unstable machine, NOT OCed, passes memtest86 but completely crashes Prime95

[psyched550]

Hello I've been a spectator on sites like this for a while but don't post very often as I'm not a hardcore overclocker, although I've been attempting it moderately with my machine, however I have some stability problems with my machine even when everything is set to default.

My machine is a bit dated, and it has been overclocked previously, but never really successfully, the most I ever got out of it was 2.8 to 3.0Ghz while stock its 2.4, but it has never been very stable so Its been running stock most of the time. A few times a week I get a bsod or random unexpected reboots when doing stuff with photoshop or maya. I suspected it was the RAM for the longest time so I just bought some new ram but the problem still exists. Again, at the moment nothing is overclocked, all voltages are set to auto right now, (not that it made any improvement when I manually tweaked these in the past)and i reset pretty much everything to default in the bios. I ran Memtest86 for 3 passes with no errors. When I run Prime95 I'm lucky if i get a few min out of Blend and then I get a BSOD with a memory dump, I can maybe get a min out of large FFT torture test, but man when I run the small FFT test I don't even get a BSOD, its just a straight reboot...according to core temp i'm not going over 70c when running prime95, and most of the time its running 45c idle and 50-60c under load, this is all stock cooler btw. I originally had liquid cooling but the system wasn't stable then either and it proved to be a waste since I couldn't OC anyway. Here are my specs:

Q6600 2.4ghz Core 2 Quad - kentsfield, 65nm, B3, running 1.28v according to CPU-z

ABIT IN9 32X-MAX LGA 775 Nvidia nForce 680i SLI (680i-MCP55PXE)

G.Skill 4gb (2x2gb) DDR2 800 PC2-6400 (F2-6400CL5-2GBPQ) 5-5-5-15-24 1.8V

Nvidia GeForce 8800GTS PCI-E

Western Digital Raptor WD1500ADFD 150GB 10000 RPM (primary)
2x500gb Western Digital 7200rpm RAID0 (storage)
1XTB 7200 Western Digital (backup)

PSU is Thermaltake...I can't see the max power from within my case and I can't remember its specs since its 6 years old but I'm sure it was adequate enough at the time for the equipment.

Antec 900 case, plenty of fans.

Windows 7 professional 64-bit

Did I leave anything out? Any other information I can provide for someone to help me narrow the problem? I'm suspecting either my CPU or MB, but more likely the MB, since not many people complain about the Q6600...plus despite the fact that the abit IN9 32max got mostly decent reviews before it was discontinued, there were a few people that spoke of it being unstable. If I don't get any replies I think i just might go ahead and replace it with a used Gigabyte EP45-UD3P I saw on craigslist for $50 and hope that does the trick. What do you guys think? I'll probably build and all new system next year when I have the cash, but for now I have to make do with this...thanks!! any help would be appreciated!

 

[Automata]

Reset the CMOS to make sure you don't have a setting that is causing it. Try one stick of RAM at a time, and rotate them through the memory slots. This will identify a back stick or a bad memory slot.

If that doesn't work, disconnect all unnecessary components that you don't need to boot the system into Windows and try again.

If that still doesn't work, I'd suspect the power supply or motherboard.

 

[psyched550]

thanks, are you saying to do all this to achieve success in prime95 or still do memtest as well? my previous ram was OCZ pc2 6400, 4x 1gig sticks...and that was just as unstable as the 2x 2 gig sticks of G.Skill i've got right now, but you know i didn't think about the slots and i'll try the individual tests

 

[Automata]

Knowing that you swapped RAM sticks, I would guess that the memory itself is not the issue. I'm betting on the motherboard (RAM slots, capacitors, etc) or the power supply.

 

[psyched550]

and i'm guessing the only sure way to find out which of the two is just to replace one of them and see what happens? or is there a way to detect a faulty power supply?

 

[Automata]

If the power supply is on the edge of failing, disconnecting components or putting in lower draw components (video card, etc) may make it stable. For example, if you had onboard video, you could remove your video card. If everything suddenly starts working, it is very likely the power supply.

 

[Knufire]

Another way to try to test the PSU would be to run Furmark on the GPU, that should cause a pretty big jump in power consumption, just like Prime95 would.

 

[psyched550]

seems to be doing ok with the furmark...well i mean it sucks at it and scores low cuz its an old card but its not crashing at least. How long should i run it and should i do the extreme burn-in or just the regular? any special settings i should tweak to maximize it? i'm still betting that the problem is the mobo, but that might also be because i really want a better board anyway. Although I don't want to spend much. that deal on the Gigabyte EP45-UD3P fell through, anyone selling one or know a place online that still sells them for cheap(ish)? Ebay has used ones already approaching $100...not sure its worth that for its age and use, but man no wonder that craigslist deal for $50 went so fast!