FEATURED So I just de-lidded my 3770K...Results and pics inside!

[xhacker]

I delidded my 3770k this morning with a stanley razer and polished it with a soft tipped dremmel bit. I decided that using a hammer was probably not my best choice of tools.

Under load temps
core before after
0 82 50
1 80 56
2 79 58
3 75 58

I also removed the lid and retaining bracket and mounted (gently) the Koolance waterblock directly on the die.

I'll show the proof when I figure out how to upload jpegs (seems to want a url)

 

[LoganG]

click on go advanced then attach the file

 

[xhacker]

Before and after screen shots

 

[xhacker]

Pics of delidding

 

[xhacker]

Sorry about the poor picture quality.

 

[ivanlabrie]

What paste is that?

You can probably go much higher mhz wise, to benefit more from the temp drops.
And use liquid metal for the die, if you're not using it already.

 

[xhacker]

It koolance thermal paste. I ordered liquid pro ultra from frozencpu yesterday. I'll change the paste when it arrives.

 

[ivanlabrie]

Awesome!

Be sure to post results...people tend to look bad at me when I say it gives 25c temp drops from regular tim/glue.

 

[I.M.O.G.]

Nice results, thanks for sharing.

 

[xhacker]

No problem, I loved the experience. Makes you wonder what Intels motivation is behind having such a larger gap between the die and the lid.

 

[I.M.O.G.]

Cool.

I've heard other people talk about the gap being larger too, but I haven't seen anything proving the gap is larger other than seeing the statement repeated a lot by different people. No criticism intended towards anyone there, I'm just recognizing what I've seen so far and I'm not sure why. I don't know that its actually a larger gap at all. In the past, the gap was spanned by the solder attaching the IHS to die, now its spanned by paste. Maybe I'm missing an article or something that provides evidence that the gap is larger - I still think its just the difference in material between the IHS and die. Would help if I knew where this larger gap theory originated, and how the conclusion was arrived at.

Curious to see what the status is with Haswell too as its getting close.

 

[briansun1]

has intel given an officeal statment about the change and why?

 

[LoneWolf121188]

Cool.

I've heard other people talk about the gap being larger too, but I haven't seen anything proving the gap is larger other than seeing the statement repeated a lot by different people. No criticism intended towards anyone there, I'm just recognizing what I've seen so far and I'm not sure why. I don't know that its actually a larger gap at all. In the past, the gap was spanned by the solder attaching the IHS to die, now its spanned by paste. Maybe I'm missing an article or something that provides evidence that the gap is larger - I still think its just the difference in material between the IHS and die. Would help if I knew where this larger gap theory originated, and how the conclusion was arrived at.

Curious to see what the status is with Haswell too as its getting close.

This is where the gap theory comes from: http://forums.anandtech.com/showpost.php?p=34053183&postcount=566

 

[xhacker]

Here's my problem, right now i'm not using the IHS at all. My Koolance waterblock is mounted directly on the die. This is not a favorable situation since it easy to break the die without the IHS. I have some liquid pro on order will the liquid pro minus the glue plus the IHS still give me excellent core temps?

This is what my temps were like with MX-4 minus the glue with the IHS.

 

[Qlix]

Here's my problem, right now i'm not using the IHS at all. My Koolance waterblock is mounted directly on the die. This is not a favorable situation since it easy to break the die without the IHS. I have some liquid pro on order will the liquid pro minus the glue plus the IHS still give me excellent core temps?

This is what my temps were like with MX-4 minus the glue with the IHS.

From my own research and reading the consensus seems to be that there is no reason at all to do a direct die mount. The gain is ~2c and not worth the risk of crushing the die.

Put the ihs back on. You probably won't notice a difference.

 

[ivanlabrie]

Here's my problem, right now i'm not using the IHS at all. My Koolance waterblock is mounted directly on the die. This is not a favorable situation since it easy to break the die without the IHS. I have some liquid pro on order will the liquid pro minus the glue plus the IHS still give me excellent core temps?

This is what my temps were like with MX-4 minus the glue with the IHS.

Don't go direct die...ain't worth the hassle, really.
You may even get better temps with the ihs on, it's margin of error area right there.

Don't use Liquid Pro on the ihs neither, you'll damage your waterblock.

 

[Qlix]

Thought you would only damage the block if it was aluminum. Don't they recommend it on the copper blocks on Coolabrator site? Staining yes fusing/dissolving no was my understanding.

 

[ivanlabrie]

It won't damage it, but leave a NASTY stain. If you wanna resell you have to basically lap it to get rid of it. Or use the metal cleaning kit they sell. (or metal polish)

 

[LoneWolf121188]

Here's my problem, right now i'm not using the IHS at all. My Koolance waterblock is mounted directly on the die. This is not a favorable situation since it easy to break the die without the IHS. I have some liquid pro on order will the liquid pro minus the glue plus the IHS still give me excellent core temps?

This is what my temps were like with MX-4 minus the glue with the IHS.

If you weren't already planning on taking it apart to replace the TIM, I'd say just leave it...once it's mounted, it should be fine. Since you're taking it off and reapplying the HSF, I'd put the IHS back on.

 

[xhacker]

Thanks for the replies

I think i will put the IHS back on with metal pro. But I'll use a different thermal paste for the waterblock. It makes me nervous even touching my system without the IHS.