Pea sized blob method for Q6600?

Hi guys

I have been stress testing my Q6600 with IBT. These are the temps I got today and IBT produced error on the 8th pass.



Yesterday I ran 20 passes of the test with same settings and it passed successfully. Temps never went higher than 69C on the hottest cores. It seems that when the temps remain below 70C IBT doesn't mind. AS soon as I hit 70C+ it crashes.

The other thing is as you noticed there is a big temp difference of 8C! between core0,1 and core2,3. I am currently using MX-4 and the application method is horizontal line. I used the same method with AS5 and temps were more or less similar. I am thinking that core0,1 naturally run hotter and these possibily consist of one chip while core2,3 consist of the other chip.

I would like to know if any of you using blob method for core 2 quad and if you can post pics of how much it should be as this method seems to be the best. I would really like to minimize the temp difference.

I just put a blob on and smooth it across the surface and then fit the heatsink

Pea size, literally. A blob the size of a pea, as near to the middle of the chip as you can get it, then push the heatsink ontop to do the spreading. Don't move the heatsink too much once you've applied it and if you have to remove the heatsink you'll need to reapply the paste.

Then I usually do a couple of sensor tests in RealTemp (20 minute prime small FTTs basically) to get it to heat up and cool down a bit. Then you should be good to go.

I've tried the line, grain of rice, spreading with credit card/finger wrapped in clingfilm methods and the one above is the best by far.

The base of my IFX-14 is slightly concave from the bottom. So when I place it directly on cpu surface, the ends of the heatsink make contact but there is slight gap between the cpu surface and the heatsink base in the middle. So I am thinking if I use the blob method in the middle, that should easily cover the gap and hopefully results should be better.

Surely thats a defect?

I'd be careful with the blob method if that is the case then, it might not spread as well as if the base of the heatsink was flat. No experience with it wonky heatsinks, thats just a theory, it could be a load of rubbish but worth bearing in mind.

Yeah lol. I have been reading that thermalright always do that intentionally with their heatsinks and that they are never flat. I don't want to do lapping as I don't have any experience with that. The gap is vey tiny.

I guess give it a go and see what happens then!

Hzza said:

I guess give it a go and see what happens then!

Click to expand...

. Will post results.

Good luck!

A pea sized blob is a very large amount of paste... is that really the best way of doing it? It just doesn't seem right. Previous received wisdom was to use a credit card and cover it using as little an amount as possible.

I wouldn't use quite as much as the size of a pea, more the size of a rice grain or thereabouts and then just let the cooler spread the paste.

yea i found the blob method to work best with my q6600

i tryed the ccard and minimal paste wasn't really impressed

it could have just been a case it slightly uneven base on cooler but i did manget to get my temps down by about 2c - 5c depending on core

My Q6600 was the same with Cores 0 and 1 being significantly higher that 2 and 3.

No method is going to give you considerably better temps, there might be scientific reasons that a pea size sized blob is best but in practice you're unlikely to notice much difference either way, your best way to improve temps would be to keep ambient temps down (window wide open, heating off etc) or fit a better heatsink.

Plus if your CPU is running 66-68C and it becomes unstable at 70C that doesn't leave much headroom come the summer.

philtorrens said:

I wouldn't use quite as much as the size of a pea, more the size of a rice grain or thereabouts and then just let the cooler spread the paste.

Click to expand...

That seems far more plausible.

When I was watercooling I'd remove the IHS and lap the waterblock down to a near mirror finish (near is better than mirror apparently). And when they started welding them on I'd just lap to copper. I'd use as little paste as possible.

I'm much less anal about it now... perhaps that's why extra paste is needed to cover the sins of an uneven surface.

philtorrens said:

I wouldn't use quite as much as the size of a pea, more the size of a rice grain or thereabouts and then just let the cooler spread the paste.

Click to expand...

I tried using less (grain of rice sized blob) and got an uneven coverage which was shown when cores 0&1 were running up to 10 degrees hotter then 2&3 and when I removed the heatsink you could visably see there wasn't great coverage. Reapplyed using a pea sized blob and got much better coverage and cooler temps. Theres still a difference but now its only 2-3 degrees which I understand is about normal.

It all sounds very well in theory, but in practice it's never quite as simple as dropping the heatsink onto the CPU. There's usually some movement before you get it clamped down.

There's nothing in my mind that makes me think the pea sized blob would be forced out in all directions, what's to stop it just going in one direction and the opposite direction getting no paste?

I'll do my usual incredibly thin layer of paste, and then put a grain of rice in the middle.

Ok took off the viper fans and you can see the dust on the IFX-14. I cleared it as much as i could.




In this picture I used paint to draw the horizontal line which I used with MX-4.




Now looking at the pics again, The darker areas are wet thermal paste while the white areas are where the thermal paste has almost completely dried up. Now with horizontal line method, I didn't expect the pattern to turn out quite like that.
The IFX-14 does have tricky mounting system plus it has the tendancy to 'slide' when I placed it on top of the MX-4. So I agree completely with the poster above that whenever placing heatsinks on thermal paste, there is bound to be some movement. However as long as you don't raise the heatsink then it shouldn't be much problem as air bubbles won't be formed.






Now I have used the grain rice method in the horizontal direction. Basically it is just smaller version of horizontal line method. Early on I applied pea-sized method and took off the heatsink to see what the pattern looked like. It was small circle/oval covering almost 60% of the cpu surface which I think is good enough.
So I am hoping that rice grain would give me similar pattern. I placed heatsink on top and then kept my hands on it for 2/3 minutes so that heatsink would 'join' with the MX-4. Even though it has a weight of over 800g, I noted the tendency to slide again hence keeping my hands on it.
Once or twice the heatsink rotated lightly as I tried placing screws into the mounting system. Would this affect the thermal paste?

However I would like to ask that when taking off heatsink, why thermal paste condition is similar to dry land that has been devoid of water so it is breaking if you know what I mean i.e. becoming crusty. Does this condition only occur when taking off heatsink or is it due to the heat generated by the cpu?

I have been doing some preliminary testing and it seems that gap has narrowed. However temps have risen higher. It has been warm in my room and ambient temperature was higher than average normal room temperature.

I think tomorrow I will run IBT again and then post results to see if there has been difference. For now guys what are your further thoughts on this matter and any further suggestions.

Many thanks

The blob in middle shouldn't be bigger than a black pepper ball size

Also those are very low volts for 3.4ghz could be why it isn't stable.

my q6600 has hit 90c under ibt during summer stress testing and was fine.

At the moment I am getting 75-85c under IBT 1.4 volts 3.5ghz on a tuniq tower 120 one fan in middle at 1500 rpm.

Every cpu and every heatsink is different though some concave etc you just have to find the best method for yours.

I tried spread method and lines and x's all sorts but found blob in middle size of a black pepper ball best for mine.
But that's using MX3.

I just can't bring myself to believe that the grain of rice/grain of black pepper/pea is the best way. It just doesn't convince me.

It doesn't seem possible that it could work unless the heatsink is positioned dead above and comes right down without movement.

Why would that be better than an exact coating of thermal paste that is as thin as possible without metal shining through? The exact coating would get pushed out of the way just as well as the pea. The only thing I can think of is trapped air.