Apophis

So. I've been having hardware issues for the past couple weeks that have completely eliminated my ability to play any sort of game. Well. Any game that actually requires CPU usage.

Many months ago I bought an Intel Core 2 805 CPU. It was the processor that was getting rave overclocking reviews as its normal clock speed was 2.66Ghz but could be clocked up to 4.1Ghz. When I bought the CPU and my new ASUS motherboard I also picked up a CoolerMaster water cooling kit with integrated pump, external radiator, and 120mm fan. Good stuff.

I set this thing up and played around with pushing the chip to the limit. I didn't hit 4.0Ghz but I got pretty damn close. My overclocking endeavor was primarily for curiosity purposes. Once I found my max stable clock speed. I dropped the clock rate down to 3.0Ghz and eventually back to 2.8Ghz to gain increased stability in GRAW. All was well for quite a long time.

A couple weeks ago I noticed that my temps were climbing. Nothing critical, but my peak temp at 100% utilization on both cores was around 64°C. Three days ago my machine started to spontaneously shut down and failed to reboot unless I did a COMPLETE cold restart. This meant unplugging the power supply, letting the caps on the board discharge, and plugging the whole thing back in again. A major PITA.

Two nights ago I decide that my temp issues had to be related to dust and such clogging up the radiator. I pulled the radiator and 120mm fan off the back plate on the case and cleaned it out and put it back in. No change in temps. In fact, after running the machine for a few more hours the thing was getting even hotter. At this point I was at 58°C with less than 2% utilization on both cores. IE: IDLE temps! Fine.. My pump must be going in the watercooler (it's a closed system, so it shouldn't have an evaporation issue).

Last night I decided to remedy the situation by pulling the motherboard out of the case, removing the mounting brackets for the watercooling system, cleaned off the thermal grease off the top of the CPU core heat spreader, and installed the original heatsink/fan that came with the CPU. The HSF was still in the box and never used so it was in new condition and should remedy the cooling issues. Makes sense, no? While I was at it, I decided to clean up excess cables inside my case and using zip-ties to neatly route all the cabling. I normally do this when I build my machines, but in the case of this last machine I never cleaned it up as I was doing intensive overclocking and figured I might need to replace some hardware (or the core itself) if I blew it up.

End result? Well. The machine started at a 52°C idle temp with the CPU fan running at about 2k RPM. I left the thing on and went upstairs for an hour. Came back down and now I was at 54°C idle with the fan at 3.6k RPM. This didn't look good, so for giggles I fired up SETI to kick the cores into high gear and see where my temps ended up. Less than 2 minutes later my machine was running at 71°C and the thermal protection in the core kicked in to a point where I could do math calculations faster with an abacus than my machine was capable of.

I shut the whole thing down and pulled the HSF off the CPU. The thermal pad seemed to have applied itself properly, but I scraped it off the heatsink, re-lapped the base of the heatsink using 2000 grit fiber optic SC/SC connector polishing paper, applied some AS3 and put it back together. After going through the same process again. I still hit a point where the core applied its thermal protection but it took an extra minute.

I've been building machines for about 12 years and have encountered all sorts of issues with hardware compatibility, cooling issues and overclocking woes. At this point in time I'm quite stumped. At the point where my CPU first showed signs of overheating issues, I had not been overclocking beyond the 2.8Ghz mark.

I would like to keep this hardware around to use for something (maybe a real-time transcoding media server for my Xbox 360) although I did order a new Core 2 Duo E6800 CPU and ASUS P5WDG2-WS PRO motherboard. I'm planning on building the new machine tomorrow evening.

So what's wrong? What did I miss? What didn't I test? What didn't I replace (other than the core itself)?

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

ednos

I've heard of a problem like this, with my friend's Intel's Core 2 processor. (Hint: don't buy Intel's products in the first or second generation.) Basically, he had to send the chip in for replacement. Now, he had only been using the stock heatsink and fan, and his temperatures were higher than yours, but it appears to be a related issue.

__________________
The trouble with the world is that the stupid are cocksure and the intelligent are full of doubt. ~ Bertrand Russell
I have a tendency to key out three or four things and then let them battle for supremacy while I key, so there's a lot of backspacing as potential statements are slaughtered and eaten by the victors. ~ Magna Centipede

Apophis

It's just been a big pain in the arse. I'm not sure if my CPU is eligible for replacement, but I'll certainly look into that. I'd like to keep the hardware for something else. I've got another Core 2 Celeron that I was using to play with MythTV that I could use for the XBox 360 media server should this hardware end up as junk.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

Dick Blonov

Unless your mobo failed somehow and your vcore has climbed up ?

Might be worth checking before putting in another CPU.

DB

__________________
|TG-6th|Blonov

ednos

Good point; I had an older MSI motherboard that caused higher temperatures than normal, and a little instability to boot... but it never got that bad. Then again, AMD processors don't go quite as ballistic as their Intel counterparts.

__________________
The trouble with the world is that the stupid are cocksure and the intelligent are full of doubt. ~ Bertrand Russell
I have a tendency to key out three or four things and then let them battle for supremacy while I key, so there's a lot of backspacing as potential statements are slaughtered and eaten by the victors. ~ Magna Centipede

BleedingKnee

This is just a stab in the dark, but maybe the CPU thermometer is busted.

You have a good old air cooled thing on your CPU now right? Well, touch the CPU heatsink (make sure you are grounded first) with your finger. If the CPU is at 75C and the heatsink is mounted properly, the heatsink should be too hot to touch.

__________________
"Even the very wise cannot see all ends" - Gandalf

Apophis

Oh, the thing is really as hot as it claims. The heatsink itself is extremely hot to the point of burning my finger.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

BleedingKnee

Here's another guess: The CPU voltage is higher than it claims (I have no idea how you could check this without probing the mobo).

__________________
"Even the very wise cannot see all ends" - Gandalf

jaymind

Like others have said, I would be looking at a CPU voltage issue. I'm guessing you had to increase the volts quite a bit on the original overclock? Did you reset them when moving to air?

The original issue very well may have had to do with some instability in the motherboard or ram that just managed to crop up. Usually, I've seen issues with shutdowns and reboots associated with bad ram or power issues.

However, I have seen one of my old AMD CPU's go to crap after 1 year on a high overclock - which I guessed had to do with the voltages I needed to run to get that high overclock.

Drizzid

Yup I would have to agree with the others. Either a voltage issue or the motherboard is toast which is causing the overheating. Have you reset the bios back to default?

__________________

JMJ

I'm going to go with the CPU voltage also. These processors do not run hot with stock cooling/voltage/frequencies. Check your bios and make sure that it's being reported correctly.

Apophis

I did have to slightly increase my voltage when I did the push for 4GHz, but it has been down to normal for a long time. IE: When I finished my OC run and dropped it down to a more practical clock speed I set the voltage back to normal.

The shutdowns are definitely heat related. Of that I'm certain.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

Apophis

Good ideas. I think once I get my new machine built tonight I'll re-flash the bios on the old board, re-set everything back to stock settings, and see where that gets me. There's also no reason why I can't manually check the voltage with an external meter while I'm at it.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

Apophis

FYI:

http://www.tacticalgamer.com/hardwar...an-true-3.html

That was the original thread that prompted me into building this machine. Post #6 on details the specs. Posts #28 and #29 have some of the original OC results. Post #33 has the temps based on running at a stable 3.6GHz.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.

Apophis

FYI:

http://www.tacticalgamer.com/hardwar...an-true-3.html

That's the original thread that prompted me to build this machine. Posts #6, 28, 29 and 33 are the most relevant.

__________________
Diplomacy is the art of saying "good doggie" while looking for a bigger stick.