Delid TIM 6700k

23JMZ07

23JMZ07

23JMZ07

Associate
Joined
5 Aug 2009
Posts
1,126
Location
England
Hi all,

I have just delidded my 6700k with the vice only method.

What's the general consensus for paste between IHS and Die??

Most recommend CLU, some CLP and some Kryonaut but I have seen Thermal Grizzly's new Conductonaut which looks to be insane but without any independent testing I'm not quite sold.

Will just be under custom loop WC no extremes for me.

Views??
 
I only done my 4670k. I used CLU between the die and IHS, then Kryonaut between IHS and block. It doesn't go above 48 degrees while being stress tested and that was with 1.35v.

Thermal Grizzly Conductonaut does look interesting but cant imagine a massive difference between that and CLU.
 
I have CLU between the die and IHS (nail varnish over the transistors first though) and Gelid Extreme between the IHS and block. My 4670k actually ran hotter naked (EK Naked Ivy mounting kit) than with the IHS glued back on.
 
Yeah mine ran like crap naked too, i found CLP works a treat on and under the IHS though, i also get temps in the 40s at [email protected]. I havent tried that Thermal Grizzly though, wouldnt mind seeing how that does, maybe you could be our guinea pig. :)
 
So is the waterblock, well one of mine is actually silver but I don't like to brag...

It's strange because you have to go core> TIM> copper> TIM> copper and every step adds a level of insulation.

Are you fellas sure you're getting enough pressure between the core and block?
 
jigger said:
So is the waterblock, well one of mine is actually silver but I don't like to brag...

It's strange because you have to go core> TIM> copper> TIM> copper and every step adds a level of insulation.

Are you fellas sure you're getting enough pressure between the core and block?
Click to expand...

One of the possible reason maybe due to the increase in surface area when you put the IHS back on, and the thermal "Impedence" is minimized by using CLU between the die and IHS.

This can proved if someone would use CLU between die and direct mounted block.
 
I still can't see how. The only change thats made is with the IHS the chip is partly encapsulated and the IHS would increase the pressure between the block and die because of the extra couple of MM.

I think it's all down to pressure myself.
 
Last edited:
P4Clock said:
why? its a block of copper for the heat to be passed to. also its not possible to get the same pressure with the cooler as the socket hase on the IHS
Click to expand...

jigger said:
I still can't see how. The only change thats made is with the IHS the chip is partly encapsulated and the IHS would increase the pressure between the block and die because of the extra couple of MM.

I think it's all down to pressure myself.
Click to expand...

mm i do too lol
 
jigger said:
So is the waterblock, well one of mine is actually silver but I don't like to brag...

It's strange because you have to go core> TIM> copper> TIM> copper and every step adds a level of insulation.

Are you fellas sure you're getting enough pressure between the core and block?
Click to expand...

It is down the pressure yes, naked it just doesn't apply enough, with the IHS it does. I guess the water blocks hits the surrounding socket or something, the chips are so thin nowadays.
 
I believe pressure with the naked ivy kit is the same as a normal mount. I also had the socket clamp off mine when I tried it naked. My reasoning was the same as Definology and that the surface area is bigger to dissapate the heat with the IHS on.
 
Thanks for the insights guys.

I tried Kryonaut and was impressed with the temperature difference but I suspect that is mainly down to the reduced gap from removing the glue/silicone stuff intel use. I have a tube of CLU on order so will see which works best for me.

Just to add, I found the vice only method perfectly safe and easy with my chip. I know the substrate is super thin but as long as you go slow and keep things square it's fine.
 
Dempsey22 said:
Thermal Grizzly Conductonaut does look interesting but cant imagine a massive difference between that and CLU.
Click to expand...

The TGC looks in it's appearance to be more like a liquid metal than a paste. And compare the thermal conductivity ratings:

CLU - 38.4w/mk
TGC - 73 W/mk
MX4 - 8.5 W/mk
CLP - 82 W/mk

It's relative thermal conductivity is closest to Pro, not Ultra. Therefore it is indeed probably most like CLP, more liquid than paste.

Hwever the TC rating doesn't seem to matter much. Because as we know, a standard paste such as MX4 still performs very well. Yet it's thermal conductivity rating is in fact multiples lower than CLU or CLP.

Conclusion:

Either CLU or a converntional paste likely still remains the more practical option. As not being a liquid means it is easier to handle and spread. Does not get everywhere and on other components.

So then this new thermal grizzly stuff would also have to be easier to use than CLP to present a tangible benefit. It's also priced a shade cheaper than CLP.

I would like to agree that the IHS gap is more important. But where's the double-run test showing a delided IHS + paste, then again as IHS + CLU (on the same delidded cpu)?
 
pastymuncher said:
I have CLU between the die and IHS (nail varnish over the transistors first though) and Gelid Extreme between the IHS and block. My 4670k actually ran hotter naked (EK Naked Ivy mounting kit) than with the IHS glued back on.
Click to expand...

Sounds like a stupid question, but do you mean nail polish or nail varnish, as we have plenty of nail varnish hear but its all coloured and not clear nail polish??

Does it matter if you actually use coloured nail varnish to cover the transistors?
 
You must log in or register to reply here.
Back
Top Bottom