Persistent crashing

[ryanblak]

I'm having persistent issues with my PC crashing in various ways.

In the beginning it would crash out randomly with no particular pattern, though this was more frequent during things like gaming. This would usually be a blue screen (if my memory is right) and wouldn't always be the same blue screen message.

The system seemed to keep corrupting various Windows and Linux installs to the point they were unusable and would require a full format and re-install. I put up with this for a while before changing mobo after reading about people having similar problems with my old mobo.

With my new mobo, problems still persist, though it usually just crashes to a black screen requiring a hard reboot. On occasion I have seen a flash of a blue screen message before the system goes on to restart itself. The system can no longer be hibernated as when it turns back on, it crashes and restarts (though this is very recent and I suspect more to do with the current OS install being shot). Did a MemTest on install of my new mobo and found 2 RAM sticks were riddled with errors, so dropped to 2 and returned no errors.

The problem currently occurs fairly infrequently, unless I'm gaming. WoW really seems to give the computer some issues, it generally crashes within 30 seconds on entering the game with a maximum of 5 minutes play. Tested WoW on various settings to ascertain whether it was the GPU:

1280 x 1024 Highest: Crash
1280 x 1024 Lowest: Crash
1024 x 768 Lowest: Crash
800 x 600 Lowest: Crash

Here are the specs:

CPU: Intel Core 2 Quad Q6600 G0
Old MOBO: Asus P5N-E SLI 650i
New MOBO: Asus P5Q PRO
PSU: Enermax Pro82+ 425W PSU
RAM: 2x 1GB Crucial DDR2 800MHz/PC2-6400 Ballistix Memory Non-ECC Unbuffered CL4
GPU: Point Of View 8800GT EXO Overclocked Edition 512MB
OS: Windows XP 32-bit

Recently did a MemTest and that returned no errors.

Seems to run quite hot on GPU and CPU. SpeedFan numbers:

GPU: 62C Idle 70+ Load
CPU: 47C Idle
Core0: 51C
Core1: 52C
Core2: 48C
Core3: 50C
Core: 62C

These numbers reminded me that when I put my CPU back in with my new mobo, I used some cheap generic brand paste, because I had misplaced my Arctic Silver (not the smartest idea, I know).

I really want to get it sorted, but would like some help on where to start with that.

Thanks in advance.

 

[doormatderek]

POWER SUPPLY, see if you can find one to try swapping it out with. Bet you $100 that is the problem...
Those temps aren't anything to be worried about

 

[ryanblak]

Thanks doormatderek, I think you are probably going to be right about that. Unfortunately, I don't have any spare PSUs lying around, but I will see if any mates do tomorrow. I'm guessing anything over 500W should do the trick?

 

[pjmelect]

Classic symptoms of a faulty power supply, the faulty RAM that you removed is probably perfectly ok.

 

[jsc]

I'm guessing anything over 500W should do the trick?

A good 500 watt PSU.

And if you are overclocking, reset the BIOS to factory defaults until you get your problem sorted.

 

[capt_taco]

Checking the PSU is a good idea, but before you spend any money ... did you ever adjust your RAM voltage manually in the BIOS? Yours looks like takes 2.2V, which is way over the standard 1.8V for DDR2, so if you didn't already tell your motherboard to adjust the voltage, it's likely to give you problems. Also, if you've got 4 sticks installed, and it sounds like you do, you may need to give it another 0.05-0.1V above that.

However, even though memory voltage is probably one of the biggest causes of random freezing, in my experience it usually just freezes or gives a blue screen. Crashing to a black screen is rare, but I suppose it's possible. Which is why checking the PSU is a good idea as well.

 

[ryanblak]

Thanks for all the responses guys. Nobody I knew had a spare PSU lying around, so I took the plunge and got myself a Corsair 650W to replace my Enermax. I've been running the PC with the new unit in for ~4 hours (~1.5 of which I ran WoW on 1280 x 1024 with high graphics as a direct comparison test) and I haven't had a single blip, so all seems well considering the frequency of crashes previously.

I haven't tried the other 2 sticks of RAM yet, but I suspect you're right about the false positive there pjmelect.

@capt_taco: I've always run the voltages on the [auto] setting in the BIOS because I don't trust my abilities of fiddling with that. Everything seems fine now, but if the problems persist, the RAM will be the first thing I look at.

Again, thanks guys, I'm really glad to have this sorted.

 

[capt_taco]

You know what, it just occurred to me there might not be anything physically WRONG with your old power supply per se -- it may just have been too small.

A 425W PSU might have been fine to handle a regular 8800GT, but I completely skipped past the part where it was overclocked. That can get your system up over 300W easy, and some power supplies start struggling around 75-80% load. Plus you can run into problems like drawing too many amps and overloading one of the 12V rails.

I got curious enough to look into it, and sure enough, these guys who tested it i a similar system recommended a 500W power supply with 28A on the 12V rails:
http://www.guru3d.com/article/geforce-8800-gts-512mb-bfg--pov-review/4

And here's your PSU at 425W and 25A per rail (actually this focuses on the 525W version), so there are a number of things that could have been causing trouble.
http://www.anandtech.com/casecoolingpsus/showdoc.aspx?i=3270&p=2

Basically, I think this means you can ignore everything I said about RAM voltage unless it starts giving you trouble again. If it works, don't mess with it!