Removed the IHS from an Ivy Bridge i5 3570k

Crazy Chuckster said:
I lol'ed @ this a bit :D

I was around "back in the day" when cpus did not come with and ihs and to get better temps you slapped a 60mm delta finger ginsu on. As a matter of fact i have a few still floating around here....
Click to expand...

I was digging around in my cupboards the other day and found my old Volcano 7+, maybe I should send it to him. :p
 
Kor said:
I was digging around in my cupboards the other day and found my old Volcano 7+, maybe I should send it to him. :p
Click to expand...

I still have my Volcano 2.. installing that thing was hair raising, the clip was super tight and a screwdriver easily slips out of the little notch and into a trace on the motherboard :p
 
Fixall said:
I wasn't even aware that this stuff existed, lol! Thanks for the continued updates!
Click to expand...

The company website has changed from the packaging.

http://www.indigo-xtreme.com

A two-application pack (for LGA1155/LGA1156) is $19.99 at FrozenCPU.com.

http://www.frozencpu.com/products/1..._LGA_1155_1156_-_2_Pack.html?tl=c127s1605b174

A two-application package for LGA1366 (same price) is also available.

http://www.frozencpu.com/products/9...re_i7_LGA_1366_-_2_Pack.html?tl=c127s1605b174

The main site has PDFs walking the user through the application/reflow process for each CPU type.
 
Crazy Chuckster said:
I lol'ed @ this a bit :D

I was around "back in the day" when cpus did not come with and ihs and to get better temps you slapped a 60mm delta finger ginsu on. As a matter of fact i have a few still floating around here....
Click to expand...

I found one in my box o' parts the other day and plugged it in. It was so loud it made me laugh out loud.
 
OP in your last disassembly pics, that still looks like about 5 times too much TIM to me
 
Subscribed. Still waiting to see if some tweaking with the mount will drop temps some more.
 
mannyman said:
Nah its not as bad as people think, I reapplied it to my old athlon about 5 times now and havent seen any wear.
Click to expand...

I've used it on 3 am3 cpus, and it has scuffed / discolored all of them. Not a huge deal but I still wouldn't put it on a naked cpu.
 
I'm wondering, how did you clean the bare die and the back of the IHS of leftover glue?
Just with cleaning alcohol and that special kind of cloth that doesn't rip?
 
mothman said:
Two perfectly mated surfaces would not require tim. Tim is used to fill imperfections between surfaces and actually impeads the transfer of heat.
Click to expand...
they'd have to be perfect. but they won't be perfect because the finishes aren't perfectly smooth and they're not perfectly level and you can't torque it down perfectly evenly. We had bare die chips back in the AXP days and people were lapping their heatsinks back then, but they still didn't run TIM-less
 
RamonGTP said:
I've got one too. I'll never throw it away, it reminds me of the days I thought it was cool to have a computer that sounded like it was about to take off.

Random Person: "Why is it so loud?"
Me: "because it's fast!"

lmao.
Click to expand...

Weren't those the days? I remember having 15 fans total in my PC at one time.

Also, subscribed. Rooting for you and your success with this project.
 
Activate: AMD said:
they'd have to be perfect. but they won't be perfect because the finishes aren't perfectly smooth and they're not perfectly level and you can't torque it down perfectly evenly. We had bare die chips back in the AXP days and people were lapping their heatsinks back then, but they still didn't run TIM-less
Click to expand...

Exactly. It's funny when people lose the argument they start talking in theory and grasping at straws.
 
Ramon...you should lap the waterblocks base if you want the best contact with the CPU core. WB must look like a mirror:)
 
Also...use Arctic Silver 5 because it is thicker and hard to spread. It was the best thermal compound for Athlon XP for a reson:)
 
kazaakas said:
I'm wondering, how did you clean the bare die and the back of the IHS of leftover glue?
Just with cleaning alcohol and that special kind of cloth that doesn't rip?
Click to expand...

isopropyl alcohol, cotton balls and a little bit o time for the die and chip. I haven't cleaned the IHS up yet. It's a little more difficult as the under carriage is recessed a little bit so the glue is harder to get at.
 
Moogoos said:
isopropyl alcohol, cotton balls and a little bit o time for the die and chip. I haven't cleaned the IHS up yet. It's a little more difficult as the under carriage is recessed a little bit so the glue is harder to get at.
Click to expand...

Do you mean left the alcohol-doused cotton balls on top of it for a while?

I'm thinking of replacing the TIM under my 3770k with some Indigo Xtreme I happen to have lying around for a few months though I'm concerned about both removing the leftover glue and securely re-glueing the IHS back in exactly the right place so it still fits in the mobo socket.
I can get to 4.8Ghz with decent temps, even to 4.9 while staying under 85C, who knows if this mod might makes me reach the legendary number :)
 
kazaakas said:
Do you mean left the alcohol-doused cotton balls on top of it for a while?

I'm thinking of replacing the TIM under my 3770k with some Indigo Xtreme I happen to have lying around for a few months though I'm concerned about both removing the leftover glue and securely re-glueing the IHS back in exactly the right place so it still fits in the mobo socket.
I can get to 4.8Ghz with decent temps, even to 4.9 while staying under 85C, who knows if this mod might makes me reach the legendary number :)
Click to expand...


Somehow I doubt your validity of those numbers stable under stress load, unless you're on some kind of multi loop water.
 
ShuttleLuv said:
Somehow I doubt your validity of those numbers stable under stress load, unless you're on some kind of multi loop water.
Click to expand...

No, the 4,9Ghz is with a thermalright silver arrow at about 1.325v, highest temps it hit was about 83C with IBT and if i recall correctly it barely passed 80C. I have to say that I secured the Silver Arrow extremely tightly though, according to the manual equaling a pressure of 70lbs.

For 5Ghz stable the temps get far into the 90s, (as I need close to 1.4V) which is hotter than I want it to run on
 
Last edited:
kazaakas said:
No, the 4,9Ghz is with a thermalright silver arrow at about 1.325v, highest temps it hit was about 83C with IBT and if i recall correctly it barely passed 80C. I have to say that I secured the Silver Arrow extremely tightly though, according to the manual equaling a pressure of 70lbs.

For 5Ghz stable the temps get far into the 90s, (as I need close to 1.4V) which is hotter than I want it to run on
Click to expand...


IBT could be the old version, also try LinX. And pics?
 
Activate: AMD said:
they'd have to be perfect. but they won't be perfect because the finishes aren't perfectly smooth and they're not perfectly level and you can't torque it down perfectly evenly. We had bare die chips back in the AXP days and people were lapping their heatsinks back then, but they still didn't run TIM-less
Click to expand...

I've run TIM-less on bare die cpu's before. As long as your heatsink and mount are ok the margin of error is extremely small between using TIM and not. If you use a TIM with larger granular properties the temperatures are often worse as it does not sqeeze into the fissures of the heatsink as easily and thus holds the metal of the heatsink off the die.

Metal contact area and pressure are the main priorities for heat transfer capability.


That said, small dab is all you ever need. Any amount that reduces heatsink contact area or pressure to the IHS or DIE is a direct insulator of the conductive energy path.
 
kazaakas said:
No, the 4,9Ghz is with a thermalright silver arrow at about 1.325v, highest temps it hit was about 83C with IBT and if i recall correctly it barely passed 80C. I have to say that I secured the Silver Arrow extremely tightly though, according to the manual equaling a pressure of 70lbs.

For 5Ghz stable the temps get far into the 90s, (as I need close to 1.4V) which is hotter than I want it to run on
Click to expand...

Interesting. My TIM replaced/modified IHS one only hits 80C with H100 at 1.4V (though mine is a bad OC'er and can only do 4.7 at 1.4V) (though I didn't have LLC high enough and voltage dropped to 1.376 under load)

ibt1v47ghz.png


Here is how I run it day to day though. Even IBT hasn't passed 65C with this.

p951hr.png
 
Last edited:
ferzerp said:
Interesting. My TIM replaced/modified IHS one only hits 80C with H100 at 1.4V (though mine is a bad OC'er and can only do 4.7 at 1.4V) (though I didn't have LLC high enough and voltage dropped to 1.376 under load)

ibt1v47ghz.png


Here is how I run it day to day though. Even IBT hasn't passed 65C with this.

p951hr.png
Click to expand...



You are getting good results.

Seems like just replacing TIM is the way to go.
 
SonDa5 said:
You are getting good results.

Seems like just replacing TIM is the way to go.
Click to expand...

In theory, the proper method would allow for the best results with no IHS at all (unless it is soldered on like with SB), but it sure is hard to get a good mount directly against that die.
 
ferzerp said:
Interesting. My TIM replaced/modified IHS one only hits 80C with H100 at 1.4V (though mine is a bad OC'er and can only do 4.7 at 1.4V) (though I didn't have LLC high enough and voltage dropped to 1.376 under load)

*image*

Here is how I run it day to day though. Even IBT hasn't passed 65C with this.

*image*
Click to expand...

I havent replaced the TIM yet, I was merly stating that being able to run 1.4V (and therefore 5Ghz) under decent temps is my drive to do so :)
 
ferzerp said:
In theory, the proper method would allow for the best results with no IHS at all (unless it is soldered on like with SB), but it sure is hard to get a good mount directly against that die.
Click to expand...


The die is so small in comparison to the rest of the chip. I'm thinking that there is a possibility that the IHS on it is able to transfer heat from the entire chip making it more efficient to transfer heat than just having the surface area of the die transfering heat. Just another theory.

Kind of like when you put the lid on boiling water to help the food cook faster and more even. Could be some type of thermodynamic heat release when the IHS is on it. I don't know how else to explain what is happening.
 
Last edited:
SonDa5 said:
The die is so small in comparison to the rest of the chip. I'm thinking that there is a possibility that the IHS on it is able to transfer heat from the entire chip making it more efficient to transfer heat than just having the surface area of the die transfering heat. Just another theory.

Kind of like when you put the lid on boiling water to help the food cook faster and more even. Could be some type of thermodynamic heat release when the IHS is on it. I don't know how else to explain what is happening.
Click to expand...

It is quite possible that the heat from the small die is able to spread over the thinner IHS at a quicker rate than if directly mounted to a heat sink with a much thicker metal block. If this does indeed happen this way then the IHS to heatsink heat transfer will be greater. The only thing stopping me in thinking this is correct is that the IHS, thermal paste, and heat sink essentially become one big block once mated.
 
joderme said:
It is quite possible that the heat from the small die is able to spread over the thinner IHS at a quicker rate than if directly mounted to a heat sink with a much thicker metal block. If this does indeed happen this way then the IHS to heatsink heat transfer will be greater. The only thing stopping me in thinking this is correct is that the IHS, thermal paste, and heat sink essentially become one big block once mated.
Click to expand...

True.. but it depends on what the IHS is made out of. It may be able to soak up heat better than the base of the heatsink. If that is the case, having the IHS in place would indeed help lower the temps compared to having no IHS.

This is going to especially be true of coolers that are made to have the heatpipes directly touching the IHS and coolers that have a very thin layer of metal between the heatpipes and the base of the heatsink.

You have to have enough metal to be able to spread the heat or else you are going to essentially only going to be using at max 3 of the heatpipes in a 5-6 heatpipe cooler.
 
No, the IHS is copper with a very thin nickel or aluminum (not sure which) plating, just like your heatsink will be. The only thing really better to use would be silver.


I don't think that you guys are really thinking this through if you think the IHS can actually help with cooling (unless soldered on). Think about it. If a thin layer and a big block (with two layers of paste overall) could cool better than just a big block (with one layer of paste), then we could sandwich many thin layers together and get even better cooling.

This is, of course, absurd, but if adding an extra physical layer actually assisted cooling, that's what the logical result would be.

If you solder it on with a material that has much less thermal resistance than the paste we typically use and then used regular paste under the heatsink, then yes, it would work as you envision, but as is, it doesn't. (Though SB did)
 
ferzerp said:
No, the IHS is copper with a very thin nickel or aluminum (not sure which) plating, just like your heatsink will be. The only thing really better to use would be silver.


I don't think that you guys are really thinking this through if you think the IHS can actually help with cooling (unless soldered on). Think about it. If a thin layer and a big block (with two layers of paste overall) could cool better than just a big block (with one layer of paste), then we could sandwich many thin layers together and get even better cooling.

This is, of course, absurd, but if adding an extra physical layer actually assisted cooling, that's what the logical result would be.

If you solder it on with a material that has much less thermal resistance than the paste we typically use and then used regular paste under the heatsink, then yes, it would work as you envision, but as is, it doesn't. (Though SB did)
Click to expand...

I didn't realize the IHS was copper.... Are you certain about that?
 
ferzerp said:
I don't think that you guys are really thinking this through if you think the IHS can actually help with cooling (unless soldered on). Think about it. If a thin layer and a big block (with two layers of paste overall) could cool better than just a big block (with one layer of paste), then we could sandwich many thin layers together and get even better cooling.
Click to expand...
yes thats true, I don't think anyone with even a rudimentary knowledge of thermal conductivity would believe the opposite to be true. You're just adding interfaces and lower conductivity paste layers. BUT, you're not taking into account some stuff that people are hypothesizing might be true as well, and I don't have a good handle on whether or not I believe they may be factors, such as what cyclone3d said:

cyclone3d said:
This is going to especially be true of coolers that are made to have the heatpipes directly touching the IHS and coolers that have a very thin layer of metal between the heatpipes and the base of the heatsink.

You have to have enough metal to be able to spread the heat or else you are going to essentially only going to be using at max 3 of the heatpipes in a 5-6 heatpipe cooler.
Click to expand...

heatsinks these days are designed for lidded CPU's, so they may be less efficient with a direct die contact. Its possible that the lowered efficiency more than makes up for the improved efficiency of direct die contact. Still, I suspect a far bigger issue is that getting good contact is very difficult because the whole retention mechanism for the CPU and HSF was designed for a lidded CPU and modding it to work with direct die is problematic
 
M8R-mx8r56 said:
Check out the last page in that thread. It seems the guy managed to get direct die liquid cooling to work. And he managed to overclock his i7 3770k to 5ghz.
Click to expand...

That is not direct die liquid cooling. Direct die liquid cooling involves water directly hitting the die...

He is still using a water block on the naked die.

Those are great results however...5GHz 1.368V peaking at 85C.
 
Activate: AMD said:
heatsinks these days are designed for lidded CPU's, so they may be less efficient with a direct die contact. Its possible that the lowered efficiency more than makes up for the improved efficiency of direct die contact. Still, I suspect a far bigger issue is that getting good contact is very difficult because the whole retention mechanism for the CPU and HSF was designed for a lidded CPU and modding it to work with direct die is problematic
Click to expand...

I think this is a significant part of it. Direct contact heatpipe coolers wil be terrible with a direct die, of course, but I think even regular coolers are not going to work as well - a lot of engineering has gone into designing them to work optimally when mated to a IHS.

zoob said:
That is not direct die liquid cooling. Direct die liquid cooling involves water directly hitting the die...

He is still using a water block on the naked die..
Click to expand...

Why must we argue semantics? Eveyone knew what he meant.
 
Low Temp Solder compounds. I would pick one that goes to 105C before melting. Put it inside the IHS, then put a heat gun on it.

http://www.hpl.hp.com/hpjournal/96aug/aug96a10.pdf

Bismith has a thermal conductivity of .0787 W/cmK
Tin has a thermal conductivity of .666 W/cmK
Indium has a thermal conductivity of .816 W/cmK
Lead has a thermal conductivity of .353 W/cmK

Choosing low temp solders high in Indium would be a good choice.

A decent candidate is 50% Tin 50% Indium compound. With a starting melt point of 118C

50In50Sn That would average out to .741W/cmK. That's 74.1W/mK

Who wants to be the first to try their own version of soldering IHS?

Considering the cost of Indium $$$$, I can see why intel dropped it.
 
Alot of effort though and temps are still pretty damn high on direct die water. Good results though...
 
You must log in or register to reply here.
Back
Top