• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

AMD Athlon Heatspreader!

M

mrman3k

Senior member
I was just browsing through the forums and just got a great idea. Basically for some company to make a heatspreader to put on existing Socket A Athlons. There could be 3 versions, one for T-bird, Palomino, and T-bred. Basically I was thinking, what if they make a heatspreader that goes over the core or the processor and is made out of copper, it would be only a little higher than the core itself, but this would not pose a problem if you are using a heatsink like the Alpha's or Swiftech which has adjustable height. I think this could prove to really work and could give OC for the Athlon, mainly T-bred a little boost.

Please post comments, criticsm, etc.
 
If the goal is to protect the core from breakage, it might help in that way (currently, shims are sold for this purpose, to keep the heatsink from getting angular contact with the core... not that I advocate shims). I've never chipped/cracked a flip-chip core, either P3/Celeron or AMD (knock on wood).

For heat transfer purposes, it would add two layers to the path the heat must travel and would increase thermal resistance as a result, since one of the two layers (namely, the thermal compound) has higher thermal resistance than the materials joined. Look at it this way... there's already a copper heatspreader built right into many of today's better heatsinks. 🙂

I should add that it's rumored that AMD will be adding integrated heat spreaders somewhere down the line, too.
 
My idea was to help dissapate heat, not use it as a shim. So please clarify this for me, isn't an IHS basically a piece of metal (copper plated with nickel or something like that) placed over the flip chip and used as an intermediary to the heatsink? If someone could explain a bit more about the actual design of the IHS that would be great.
 
Originally posted by: mrman3k
My idea was to help dissapate heat, not use it as a shim. So please clarify this for me, isn't an IHS basically a piece of metal (copper plated with nickel or something like that) placed over the flip chip and used as an intermediary to the heatsink? If someone could explain a bit more about the actual design of the IHS that would be great.
Click to expand...
That's about the long and short of it, yep. Since the core's contact area with the IHS is the same as it would be for the heatsink itself, that's the bottleneck... the size of the core's physical contact area with the next layer, whether it be an IHS or the heatsink itself. The heat has to make it through that bottleneck before it can take advantage of the IHS's larger contact area with the heatsink, except for heat that's leaked off the underside of the CPU. So it's the same difference, in a way.

Now, if the interfaces between the IHS and the core, and between the IHS and the heatsink, could be bridged with something that conducted heat better than solid copper (say, if you could fill the gap at the atomic level with monocrystaline diamond or pure silver), then the resistance of the "stack" could be less than if the IHS and the additional interface were not present. Not practicible.

 
Well then please explain why is it better to even have an IHS like the P4 and future Hammer? If the Heat Spreader that was sold fit perfectly around the core and used a bit of AS3, I think it could possibly improve cooling performance. Too bad I can't try to make one and test it, my budget is too small.
 
Look at an AthlonXP with a Palomino core. Surface area: 128 square millimeters. Throw in an Alpha PAL8045 and some AS3.

The heat from the core must come out of those 128 square millimeters to get anywhere. Right? Right. The heat goes from the CPU core to the AS3, which passes it to the Alpha's built-in copper heat spreader, and on to the pins, where it's transferred to the air.

Now let's say we managed to extract the Alpha's copper heatspreader from its base, and attached it to the CPU package, again with AS3 as an interface. Has anything changed? Nope, the contact area between the CPU core and the heatspreader is still 128 square millimeters.

Next, we dose up the top of the extracted heatspreader, now attached to the CPU package, with AS3, and re-fit the rest of the Alpha heatsink to it. Instead of having the copper heatspreader impact-forged into the aluminum, we now have an extra layer of AS3 impeding thermal transfer. This additional layer was not necessary before, but it is now, because the heatspreader and the heatsink have been separated. Thermal resistance goes up.

If you look at AMD's builder's guide, they say to avoid using thermal grease due to pump-out problems as the core and heatsink expand and contract at different rates, gradually pumping thermal grease out of the interface. This is one reason why AMD uses phase-change thermal compound (thermal bubblegum stuff)... it resists pump-out. However, it's good for just one use. The use of a copper heatspreader over the core would solve this problem by

1) making the "end-user's" access point (the junction between IHS and heatsink) large, so low-quality thermal grease will still be able to get the job done and pump-out will take a long time

2) making the end-user's access point a junction between two similar materials (both metal, probably both copper) so the expansion/contraction rate is the same or close, reducing pump-out tendencies greatly.

And of course, the IHS would pretty-much eliminate the possibility of a user pulling the CPU's core right off the substrate due to sticky PCTC adhering it to the heatsink, or cracking the core by exerting force on the heatsink instead of just the clip during installation/removal.

I understand that Intel now recommends thermal grease after confirming that pump-out is not a problem between a P4 IHS and a heatsink, lending some credibility to my speculations. What's between the IHS and the actual CPU core, where the initial heat transfer takes place...? I don't know either, but I'm sure Intel has carefully selected it, whatever it is, for its ability to stay between the core and the IHS for the duration of the warranty period.

Ok, enough of my ramblings... someone else go now 😀
 
So you are basically saying that the IHS lowers the cooling potential? I mean what you said makes perfect sense, instead of going from the die to the heatsink, you are going from the die, through AS3, to heatspreader, to AS3, then to the heatsink. But, then what is the reason for using an IHS, is it simply so people do not accidentally kill the cores when they install heatsinks, or is it like what I thought, to spread the heat over a larger area?
 
Lemme quote myself, thought I sort of addressed that already:

If you look at AMD's builder's guide, they say to avoid using thermal grease due to pump-out problems as the core and heatsink expand and contract at different rates, gradually pumping thermal grease out of the interface. This is one reason why AMD uses phase-change thermal compound (thermal bubblegum stuff)... it resists pump-out. However, it's good for just one use. The use of a copper heatspreader over the core would solve this problem by

1) making the "end-user's" access point (the junction between IHS and heatsink) large, so low-quality thermal grease will still be able to get the job done and pump-out will take a long time

2) making the end-user's access point a junction between two similar materials (both metal, probably both copper) so the expansion/contraction rate is the same or close, reducing pump-out tendencies greatly.

And of course, the IHS would pretty-much eliminate the possibility of a user pulling the CPU's core right off the substrate due to sticky PCTC adhering it to the heatsink, or cracking the core by exerting force on the heatsink instead of just the clip during installation/removal.
Click to expand...
 
I am a bit brain dead this morning, so the main reason to have a heat spreader would be to avoid the "pump-out" effect with the thermal grease?
 
I'd assume it's one possible reason, along with the elimination of cracked/chipped cores, and cores being yanked off the substrate due to sticky PCTC. I'm not saying it's a bad idea overall... it's a step backwards in cooling from a theoretical standpoint, but the other advantages may outweight that, particularly if the new stepping of Thoroughbred really does run cooler as rumored.
 
So really, if we were able to take the IHS off of the P4, we could theoretically get it to run even cooler because the die itself is coming in contact to the heatsink, not through a 3 or 4 layer process.
 
Yep! People used to do that with the old SlotA Athlons, which had a heat spreader fastened over the core. Bolt an Alpha 7125 directly to core, and off it goes 😀
 
I should add that it's rumored that AMD will be adding integrated heat spreaders somewhere down the line, too.
Click to expand...

One major company already had new cpu coolers with this in mind, but had to pull them at the request of AMD. Not sure why and they cannot say.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top