Serious heat problem :( Help!

[voss]

Hi all,

I've been having no problems with my 2500+ Barton for a few weeks now, everything working stable and relatively cool. Until now ;(

I noticed the diode and socket temps were getting higher than they were in the last few weeks, abouts 5C higher on average for the diode. I have not changed anything on the PC since I got it, same cooling, same BIOS and same settings... basically using 210x10 or there about with 1.65vcore the whole time.

It's summer here in Europe but my ambient room temp is definitely the same, my house is air conditioned and my mobo temps are always the same 28-30C.

So I decided to take the heatsink off, Volcano 7+, and check the thermal paste. It looked fine so I just cleaned the die and heatsink and put some more on... same Radio Shack stuff I've had for a while... a few years maybe. I am sure the hsf is on correctly (since I had problems with that before ) but now the temps are even WORSE!

Now on default settings 11x166 1.65vcore I have an idle diode temp of 53-54C which was my old load temp on the default speed Load is up to 70C!!

This is definitely terribly wrong, but I can't work out why? The heatsink gets very hot, so I know it's conducting the heat from the die. All my fans are working.
Does anyone have any clues as to what's happening here? Can the thermal grease decompose with time and become completely useless? What else should I check?

Thanks

 

[mechBgon]

The current Radio Shack thermal grease is very poor quality. Can you find some Arctic Alumina, Arctic Silver 3 or another high-quality type?

Also double-check that you have the clip into your heatsink in the proper direction The longer leg goes towards the raised, solid-plastic end of the CPU socket.

 

[o1die]

Is it possible you applied too much thermal grease, or installed the heatsink backwards? If you've never used thermal grease before, the arctic silver website has instructions on how to apply it. I use a single edge razor blade to apply the thinnest possible layer over the die only, then do the same on the center of the heatsink. Radio shack thermal grease isn't that bad. It's only 2-3 degrees warmer than artic silver.

 

[voss]

No I can't get any of the 'real' stuff here. I'd need to order online.

I spread a thin amount with a peice of thin plastic, credit card like, and made an even layer, just enouhg to cover the die... less than a mm thickness.

My hfs is on right since I can't put it on backwards and have the Tt letters the right way up. That's how i put it together and it worked fine until just recently?

It's not ossible for the CPU to start dieing on me slowly like this is it? It wouldn't get hotter, just stop working stable, right?

I am compltely out of ideas... This is not a good situation at all.

 

[o1die]

I'm also having a heat problem with my albatron board, and will rma it soon. If your heatsink fan has a temperature sensor for the fan rpms, I would monitor the fan speed to be sure it's running at the correct rpms. I use a zalman "fan mate 1" for my amd heatsink, as the stock rpm is too high. I dropped mine from 4500 to 3200, even though the cpu temp rose a few degrees. I prefer quiet over cooling.

 

[oko]

I have the exact same problem as voss. I was wondering if the diode in Barton is not
calibrated. My temps go up to 70C under load, but system is stable (prime95 gives no errors).
The socket temperatures are about 55 when diode is 70, isn't this a big difference?

 

[VitoVonAntwon]

Hey, I get the same problem, I have a p4 2.8c o/c to 3 ghz with 3500 hyper x from best buy.

Machine was a bitch to get working stable, It likes 3.0ghz rather than 2.8. I did bump the voltage of the mem to 2.7 and after that everything was behaving rather nicely. But the Heat is High.
Idle is 51, and At load is 67 about. I'm going to buy more case fans and a rheostat, Hopefully that will get it lower.

VitoVonAntwon