HOW TO REMOVE i7 IHS

[zalbard]

Would that significantly reduce temperatures using air cooling?

[dum]

Would that significantly reduce temperatures using air cooling?

see page 2 up top he answers that question I believe

[Smartidiot89]

Thats not eXtreme.. thats totaly insane. And that video .. wouldnt like to see what could happen if flame met CPU directly.. btw. how can CPU survive that much heat? (Yes I know, since it was soldered, it must been done with some heat too..).

Im tempted to try it too.. but its kinda expensive fun.

Heat isn't what kills a CPU, it's electromigration and for that to happen you need to have it up and running

[deathman20]

Heat isn't what kills a CPU, it's electromigration and for that to happen you need to have it up and running

Well if you melt it, im sure it wouldn't be working as well

[JustCallMeFrank]

You are a very brave and skillful guy! Well done! I don't have the nuts for this!

[WeeMaan]

Truly awesome work I'm impressed you had the guts to do it!

Made a thread at a Swedish forum about this thread just so you know.
http://www.sweclockers.com/forum/sho...hreadid=866224

[rge]

Here is fallwind's link that tried direct die watercooling, may wish to read that as well. It increased load temps by a huge amount in addition to seeing a 16C difference between cores, and he offered a screen shot at stock settings and load. And btw, that is what one would expect.
http://img36.imageshack.us/img36/4885/loaded.jpg
http://www.xtremesystems.org/forums/...d.php?t=225354

And another thread on direct die.
http://www.xtremesystems.org/forums/...d.php?t=227964

[Krautmaster]

+1...which is why passive direct die cooling wont work on i7, I drew a pic back when arguing with someone else ...there is no getting around the very small surface area of die, and .58 W/M*k heat conductance of water is not going to dissipate 180W. I could see IHS/waterblock combo without thermal paste, ie soldered as doing better, but not direct die.

i think you can't argument with the heat conductance of water / copper...
You have to mind that the water comes with high speed and only stays for a very short time directly on that cpu. So take a stick of copper and heat one side of it... Don't you think the waterflow will be faster than the heat conducting through the metal?

The fakt is, that direct die watercooling offers continuously new cold material (water) to dispate the heat. Copper heats on and on, there is no new material coming... heat have to to through the material.

I think direct die cooling only works if the water comes with a quite fast an continuous stream over the die.

Example:

Try to brun a hole in a plastic bottle filled with water. Think you can't...
Try to burn a hole in a bottle filled with copper - or a brick of copper packed in plastic... i think you will know what i'm talking about...

[trueg50]

Don't forget, while Copper provides "666x" faster heat transfer then just the water, the performance will be judged by the poorest component not the best, in this case by either the core ->TIM-> IHS application, or the IHS->TIM-> waterblock.

This simplifies the process from "core->TIM-> IHS-> TIM->copper block -> water" down to just "core->water". Though all our theorizing doesn't take away from the fact that the OP is getting crazy good temperatures that none us water or air cooled guys could dream of.

[rge]

Pull your IHS off a core i7 and prove us wrong (not that hard, I have removed soldered IHS's before) , only shoot a youtube video of it working, if you want to seemingly defy physics and expect belief, you have to show proof. I have youtube videos of a thermocouple in IHS and one directly in the die via drilled holes in IHS and drilled into die, showing actual gradient through IHS and dies, posted in realtemp thread. So you can easily make a video after you pull the IHS off showing both the setup and temps/screen all running and prove it works...screenshots are easy to fake, not that I have even seen a screenshot of any successful direct die cooling.

I am not pulling IHS off on my i7, as I know what will happen when trying to merge 666 lanes of traffic into 1 lane. Yes the cars could all drive incredibly fast...but speeds would be unrealistic. If you think you can compensate by theoretically running water fast enough, I would suggest you start with this pump . As to how to find a waterblock that wont rupture or keep die from cracking under that pressure, I will leave to you.

trueg50, as for simplifying the process down to core>water, you are ignoring surface area.

Direct die = 10,000 cars (heat) in parking lot A drive in single lane to parking lot B (water) at 1 per minute (represents slow .5w/m*k transfer water). That is simplest transfer via shortest route, and the rate limiting step occurs immediately (but low surface area for transfer).

IHS/waterblock = 10,000 cars in parking lot A drive in 160 lanes or 160x fast (160x represents rate tim1/solder vs water) then those speed up to 300 per minute (copper IHS) as lanes widen/go faster, then across tim2 at now still 50X faster than water (surface area increase x tim2 (3-6W/m*K), then through copper water block again 300x fast, then finally through 1 per minute now slow water, but x 1000 lanes (massive surface area of waterblock copper pins) = 1000 per minute to lot B, ie here is where surface area disagrees with your rate limiting arguement being same regardless of where it occurs.

IHS/Waterblock heat (cars) has longer to go to get to parking lot, but guess which parking lot empties faster?

[Demo]

Nice to see that the core is still intact, good job!

[Krautmaster]

i'll bet the heat transport in massive copper is much slower than in a fast water stream... So this direkt di cooling will beat a 5x5x5m cube of copper sitting on your cpu

In your opinion, a double size copper water block shoud be much more effective than a smaler one. I think pure copper limits the heat with could be taken away.
Think of heatpipes, with fluid material inside... This dosn't make sence with your arguments.

So we have a increase in the ability of taking heat away:

pure copper -> copper combined with fluid or gas (moving molecule, atoms, whatever) inside (heatpipe) -> direkt fluid contact at high speeds.

good-> better-> best

And you have to think that the gas / fluid inside a heatpipe is much slower than in water cooling cirquit...

[v0dka]

You forget about surface area, that is needed to transfer heat to the water. Ergo: the most effective way of cooling i7 is with a top performing block (good flow rate, large surface area with lots of pins etc.) directly on the CPU.

Also, you cant compare heatpipes to pure copper or water alone. Heatpipes work on the principe of condensation.

I suggest you read the rather good thread in the watercooling section about this too. But it seems that you simply ignore the test quoted above anyway.

[Krautmaster]

ut it seems that you simply ignore the test quoted above anyway.

and so you ignore the test in post#1?

Why did he get very low temperatures with direct touch? Think Frank should offer a few screenshots and pics from the hole setting...
This is no easy topic. Somewhere, there have to be a maximum of cooling performance with water. The more surface the better temps? Perhaps the minimum of copper compared with the most posible surface is the solution, don't know.

[rge]

Perhaps the minimum of copper compared with the most posible surface is the solution, don't know.

That is pretty much what everyone has been trying to tell you, and that is pretty close to the current solution of IHS/waterblock. If you want to try to better the current solution, solder/attach a copper waterblock directly to die, so bottom of copper waterblock becomes the heat spreader and eliminates tim2. Only post a video of it working this time as might have some positive results.

[zalbard]

So did anyone try removing IHS for air-cooling setup and running it that way? I believe the chip should be able to handle even the normal case's vertical mounting if you attach the heatsink properly.
I am more concerned if one gets better temperatures that way...
http://www.xtremesystems.org/forums/...d.php?t=223542 Someone get a hold of this guy!
I am getting i7 next month, and wondering if it's worth trying.
I don't believe noone tried... Perhaps, if not air, then just a normal (not direct die) water setup? Results should be similar... Anyone?

Edit: one more local linky, direct die cooling and way worse results... http://www.xtremesystems.org/forums/...84&postcount=2 And I believe it would be even worse for normal air or water...

[orig3n]

I understand that direct die most likely wouldn't be the best but surely removing the IHS and mounting a waterblock directly on the core would help alot

[prznar1]

rge i think that you forgot that there is no material that will take 100% of heat and spread also 100% of it. if that thing would exists we didn't have to pay energy bills

more things on the way of heat is worse compared to direct die watercoling. think of it. heat have to go thru IHS, thermal paste, bottom of water block, and what are the results? you are colling a 55* buttom of waterblock (when cores are 80*) with ambient water. what will be better? if your theory about 666x better colling with standard waterblock would be true nobody would care about phase change lizquid nitrogen or whatever, everybody would just sit on AC bc of more copper... nonsense

[k0rnh0li0]

you guys hold up to your name. nice job

[billdavis]

easy work...thanks man

[zalbard]

Let us know your temps!

[poke349]

lol, I just noticed.

The "Intel Core i7 - DO NOT REMOVE IHS!" thread is no longer sticky.

[orig3n]

results please!

[g0dM@n]

So did anyone try removing IHS for air-cooling setup and running it that way? I believe the chip should be able to handle even the normal case's vertical mounting if you attach the heatsink properly.
I am more concerned if one gets better temperatures that way...
http://www.xtremesystems.org/forums/...d.php?t=223542 Someone get a hold of this guy!
I am getting i7 next month, and wondering if it's worth trying.
I don't believe noone tried... Perhaps, if not air, then just a normal (not direct die) water setup? Results should be similar... Anyone?

Edit: one more local linky, direct die cooling and way worse results... http://www.xtremesystems.org/forums/...84&postcount=2 And I believe it would be even worse for normal air or water...

I'm not sure if it's much of a comparison, but I have put naked socket 939 chips on stock coolers and aftermarket coolers without too much trouble.

It doesn't hurt to be TOO careful. I just wouldn't trust any coolers that use the push-pins. Those mount with a ton of uneven pressure since it's a full mount every time you push a pin in. With screws, you can give a small turn to each of the 4 one-by-one... so much safer.

I wouldn't be afraid to mount an aftermarket air cooler... that's if I had the $$ to afford an i7 and then taking the risk to pop off the IHS. I haven't had an abundance of money since the 939 days.

By the way, GREAT TUT!!

[rge]

more things on the way of heat is worse compared to direct die watercoling.

Which is why you should take of your IHS, and do direct die water cooling and prove your theory. No point in arguing what you can prove correct? So look forward to seeing a video of your direct die cooling system with better temps. Because if you honestly believe that, is it very easy to do. Takes about 15 minutes at best to remove IHS on soldered die.

Or are you going to be yet another poster who claims they can better temps with very little surface area of water, but yet fails to prove it.

Funny how all the people that posted any kind of proof after removing the IHS, got worse temps. Also corross says he got better temps, not using water but nanofluids with very high concentration of cnt, so seems to me he thinks water will not work either.

So can you point to even one individual who did get better temps with direct die on 100W plus heat load cpu using water... that even bothered to fake a screen shot, yet alone post a video?

Also regarding your theory of more things in way of heat make it worse, guess I need to stop using tim, it is just more things in the way, especially since tim is only 4-8 W/m*K heat conductance. Also cut fins off aircooling, fins are just more copper in the way.

Or if you want to try something that might work, remove IHS, solder and figure out how to reflow indigo xtreme solder like paste to connect waterblock to die, so bottom of the heat sink serves as IHS, that might actually lower your temps, assuming it does not damage the die.