Coollaboratory TIM on 3770k and 7970

[Chazz]

Hello folks

I'm new to this forum, as a member i've just signed up but i've been browsing it for quite some time now and Techpowerup.com is my great source of news and info and i just love it.

Anyway, i haven't had much time to read all the posts related to my questions that i'm about to ask but i did browse a couple of threads here and there and also googled it and i'm pretty sure i got most of the things planned and what not.

As the title says, i want to de-lidd my 3770k, it's a decent batch as it seems from my own tests, can reach 4.8k with just 1.32 on air but it does get hotter in prime/ibt after 30 mins or so, 92-93C with a NH-14D. So i decided to "operate" it. Got me a nice thick razerblade, an old credit card to scrape de rubber, isopropyl alcohol and cotton pads to clean the cheap paste from the inside of the ihs and die and CL Pro for the die and MX-4 for IHS. Now i know people are going to say to use CL Pro/Ultra between the IHS and the cooler HS but i'd like to hope that once i'm done de-lidding and closing it back with some heat resistent gasket silicone i can "fool" the store i bought it in case it ever breaks somehow and have to take it in. And i know CL Pro/Ultra swipes the text from the IHS when you clean it. Anyway, i'm also switching from NH-14D to Swiftech H220 once the de-lidding is done.

Now my questions to you, nice folks are:

- how much time should i let my cpu rest after applying the CL Pro on the die and closing it back before putting it to use?
- is it safe or better to use CL Pro or Ultra on my graphics card also? It's a rev.B Sapphire 7970 Ghz Ed Vapor-X, custom pcb, black. While the Vapor-X cooling is really great, it never passes 70C in full load, it does get noisy after a while when the fans run at 60%+. So i was thinking if i replace the factory paste from Sapphire with CL Pro or Ultra, is it safe or better? Can it do any harm?
- how do you recommend to close 3770k back? I was thinking of adding a thick line of black silicone adhesive on the inside of the IHS, all along the margins, then using a cotton ear bud dipped in water&soap to evenly spread it and arrange it, then fixing it on the cpu pcb and pressing it firmly for a couple of minutes and then just leave it for the night to dry out. Any thoughts/suggestions?

Now i know this is a long post but i'm a really patient and meticulous guy when it comes to delicate matters like this one and even though i'm aware i might actually kill my cpu if i'm not careful i still want to do it. Doesn't hurt to ask and make sure before i start though. Thanks for taking the time to read it!

 

[AsRock]

Maybe ask here and there is another thread some were how you do it with a razor too
http://www.techpowerup.com/forums/showthread.php?t=182313&highlight=3770+delid

Razor mothered
http://www.techpowerup.com/forums/showthread.php?t=169489&highlight=Delidding

http://www.techpowerup.com/forums/showthread.php?t=182312&highlight=Delidding


If you do a search for delidding you should find more threads.

 

[Chazz]

Hehe, thanks for the links but i'm actually confident that i can de-lidd it safely if i take my time and have patience. I was looking for the answers for my questions but i'll browse those threads too - maybe something new pops out.

 

[Sasqui]

- how do you recommend to close 3770k back? I was thinking of adding a thick line of black silicone adhesive on the inside of the IHS, all along the margins, then using a cotton ear bud dipped in water&soap to evenly spread it and arrange it, then fixing it on the cpu pcb and pressing it firmly for a couple of minutes and then just leave it for the night to dry out. Any thoughts/suggestions?

Once you remove the black stock silicone from the IHS and PCB, the IHS will be resting on the top of the die, and there will be a gap where the original silicone was.

You can do a few things:
1. Put the TIM on, place the IHS on, carefully place it in the socket, lock the retainer mechanism on making sure the IHS doesn't move too much from where it's supposed to be. No silicone needed.

2. Add a dab of silicone on each corner of the IHS and follow step #1.

3. Spread silicon around the entire perimeter of the IHS where the old silicone used to be, and repeat step #1.

I tried #1 and #2, no difference in heat, great results either way. Be consrvative with the TIM and silicone (don't use too much of either.)

Make sure to run at stock first to verify temps are good and everything is placed right!

Here's my original thread with results: http://www.techpowerup.com/forums/showthread.php?t=182013

 

[RCoon]

Tempted to delid my 3570k this weekend now...

 

[Sasqui]

Tempted to delid my 3570k this weekend now...

Do it! My results here: http://www.techpowerup.com/forums/showthread.php?t=182013

 

[RCoon]

Do it! My results here: http://www.techpowerup.com/forums/showthread.php?t=182013

What razor is suitable? I have spare time Wednesday.

 

[Sasqui]

What razor is suitable? I have spare time Wednesday.

I used this:

But would recommend one of these:

 

[RCoon]

I used this:
http://www.floormatshop.com/Hardwar...placement-Utility-Knife-Blades-10-5-Packs.jpg

But would recommend one of these:
http://modelshipbuilder.com/e107_images/custom/Tools/single-edged razor blade.jpg

Assume i can get one from my friendly neighborhood anti-suicide hardware store?

 

[Sasqui]

Assume i can get one from my friendly neighborhood anti-suicide hardware store?

Lol, yes, just don't walk in with bandages on your wrists asking about the effectiveness of the blade on a jugular.

 

[Frogger]

So i was thinking if i replace the factory paste from Sapphire with CL Pro or Ultra, is it safe or better? Can it do any harm?

DON'T do it Both tim's are conductive 1 little drop in the wrong place & say goodbye to the card use MX4 or GELID EXtreme the temp diff is <2% not worth the card loss

 

[Arrakis9]

Assume i can get one from my friendly neighborhood anti-suicide hardware store?

Don't use a razor blade, there's a new method which is less risky

http://www.overclock.net/t/1376206/how-to-delid-your-ivy-bridge-cpu-with-out-a-razor-blade

theres even a youtube video of it which shows how easy it is How to Delid your ivybridge cpu (Razorless) - YouT...

 

[AsRock]

Don't use a razor blade, there's a new method which is less risky

http://www.overclock.net/t/1376206/how-to-delid-your-ivy-bridge-cpu-with-out-a-razor-blade

theres even a youtube video of it which shows how easy it is How to Delid your ivybridge cpu (Razorless) - YouT...

Yes because everyone has a heavy duty vice lol.

 

[Arrakis9]

Yes because everyone has a heavy duty vice lol.

no but a neighbor or a friend or family member might

Don't think inside the box

 

[Sasqui]

Yes because everyone has a heavy duty vice lol.

Yes, and using a hammer and 2x4 on your CPU is MUCH safer, lol

 

[Chazz]

DON'T do it Both tim's are conductive 1 little drop in the wrong place & say goodbye to the card use MX4 or GELID EXtreme the temp diff is <2% not worth the card loss

The Sapphire Thermal compound is that good?

 

[drdeathx]

Don't use a razor blade, there's a new method which is less risky

http://www.overclock.net/t/1376206/how-to-delid-your-ivy-bridge-cpu-with-out-a-razor-blade

theres even a youtube video of it which shows how easy it is How to Delid your ivybridge cpu (Razorless) - YouT...

Razorblade works fine. Take a few minutes FYI

 

[Chazz]

Yup razorblade is the way i am planning to go, i might have access to a very old vice but it's full of dirt and rust and might not have the proper edges to hold the cpu ihs in place.. i haven't seen it yet but i'm not confident about that method.

I could have been a doctor - my hdans arre vrey setady wen i dnot dnirk *hic!*.

 

[erocker]

Anyone try heating up the IHS/silicone glue before using a razor? Just wondering if it would make things easier or not.

 

[drdeathx]

Anyone try heating up the IHS/silicone glue before using a razor? Just wondering if it would make things easier or not.

No Rock. Just keep going around and slice the seal. Once the seal is cut, twist it and boing!

 

[erocker]

Yeah, I didn't think that heat would have much impact on it. Just wanted to make sure.

 

[drdeathx]

The seal goes in about 3/8 of an inch. just use a sharp blade and slice slowly not to let the blade go deep. A stanley blade works best and there is a lot of tolerance past the seal.

 

[Frogger]

The Sapphire Thermal compound is that good?

not the greatest using MX4 works fine about 5% better than The Sapphire

 

[Sasqui]

Anyone try heating up the IHS/silicone glue before using a razor? Just wondering if it would make things easier or not.

No Rock. Just keep going around and slice the seal. Once the seal is cut, twist it and boing!

Never even thought about doing that. Put in the oven @150f for 5 mins probably would soften it, but gloves would be mandatory.

I found that once you dug into 3 of the corners, the whole lid started lifting off pretty quick, so theres a first nerve wracking couple of minutes before it starts to give up.

 

[Orthello77]

I'm using LP straight to EK wb direct to die on cfx 7970 lightnings and ek supremecy cpu wb direct to ivy 3770k delidded die also. Had no issues what so ever with the 7970 but the cpu is on its second mount.

First mount with LP to the 3770k, degraded after about 48 hrs. Was gradually gaining temperature so i redid the mount with LP again and seems to have fixed it so far - although its really to soon to say. I run chilled liquid (-28c) so my issue is the LP is freezing solid below 8c and becoming a solid and then turning to liquid again above that. Thats a problem most people don't have so i won't go into it here. Suffice to say my CPU mount could be problematic due to phase changing of the tim all the time no load to load etc. That could be ruining its properties , time will tell.

Anyway Results with LP blow away MX4 .. don't even consider MX4 for the die to ihs if using the ihs.. however if you are using the IHS then i would not recommend LP between the IHS and your block . They have to be dead flat as LP puts a very thin layer on. Miles thinner than MX4. So you can get very crappy results if the surface is not dead flat.

Running 5.1 ghz with a 3770k stable in IBT max, albiet high voltage 1.56v. Max temp 32c. I can run 5.0hz with 1.52v. (these are the real measured voltages too).

My rig here http://forums.extremeoverclocking.com/showthread.php?t=363597&page=13