Will we see CPUs with integrated vapor chamber heat spreaders?

[EliteRetard]

The Wraith Spire cooler that AMD gives into the box of Ryzen 7 1700 and Ryzen 5 1500X and 1600 actually: https://www.youtube.com/watch?v=p8qeLXjiVms&feature=youtu.be&t=1m45s
*snip*

Intel did this way back with core 2:

I stocked up on these coolers back when I could get them under/around $20 shipped. Because Intel uses a square mounting system it has been very easy to modify it for newer sockets. I found this cooler to be by far the best compact low profile cooler available when the 4790k was released. I had one of those CPUs in an SG05 build and tried tons of stuff to cool it, ended up sticking with the old beast. By keeping the fan switch in performance mode and using the BIOS to customize a fan curve I could keep the system very quiet at low loads, and keep that crazy CPU cool while stress testing.

I'd be curious to see how these old core 2 coolers compare to that spire...both performance and noise. I've also always wanted to try modifying one of these things, using different and various amounts of liquids inside (water, 90%+ IPA, etc). I actually still have one of these coolers left unused...but I don't have the tools or expertise needed to properly modify and test.

 

[Phynaz]

Intel could replace their 20 year old ATX motherboard layout with something better suited to today's >100W CPUs and >200W GPUs. Something with a fixed CPU socket location and GPU socket location- yes, socket, use MXM or something like it- so that cases can be designed with integrated and efficient cooling.

They did - BTX
https://en.wikipedia.org/wiki/BTX_(form_factor)
http://www.anandtech.com/show/1157

 

[Valantar]

*snip*

*snip*

Hm, now that's intriguing. For some reason I thought vapor chambers only came in "flat" shapes. Is there no wicking material inside? I wonder how that works. Also, might this be cheaper than a simple copper slug, given the price of copper? I also wonder how they're made - are the vapor chambers made separately and pressed into the heatsink like regular copper slugs, or are they put in place, then sealed up afterwards to avoid leaks?

 

[NTMBK]

They did - BTX
https://en.wikipedia.org/wiki/BTX_(form_factor)
http://www.anandtech.com/show/1157

I genuinely think we'd be better off if BTX had succeeded. Having nice clean airflow directly over the CPU makes much more sense than the mess of ATX.

 

[maddie]

Hm, now that's intriguing. For some reason I thought vapor chambers only came in "flat" shapes. Is there no wicking material inside? I wonder how that works. Also, might this be cheaper than a simple copper slug, given the price of copper? I also wonder how they're made - are the vapor chambers made separately and pressed into the heatsink like regular copper slugs, or are they put in place, then sealed up afterwards to avoid leaks?

AFAIK the wicks merely return the liquid to the heat source [base of die], so that cut-away design should work best if mounted upright as the liquid would flow along the sides to the base. The wicks are used for the liquid to go up against gravity or horizontally, instead of settling in the cooler section.

 

[EliteRetard]

Hm, now that's intriguing. For some reason I thought vapor chambers only came in "flat" shapes. Is there no wicking material inside? I wonder how that works. Also, might this be cheaper than a simple copper slug, given the price of copper? I also wonder how they're made - are the vapor chambers made separately and pressed into the heatsink like regular copper slugs, or are they put in place, then sealed up afterwards to avoid leaks?

Based on the similar plug between the two I'd say these are filled afterwards. Once assembled it's pretty easy to squirt some liquid inside, heat the thing up to vapor/boiling point (evacuates all the air), then seal the plug. This creates a vacuum chamber with a liquid that will easily vaporize when heated. Usually there is a wicking material, but I'm guessing these designs depend on condensation to work. It's also why I'm curious to try various other things inside these chambers.

While you are using much less copper material with a vapor chamber, there are other costs involved in both designs. I'm guessing the manufacturing cost offsets the savings in material for AMD, and the Intel cooler is probably more expensive since it has all the expense of soldering on and crimping those tiny fins.

 

[DrMrLordX]

I would think multi-walled CNTs or synthetic industrial diamond would be more cost-effective than miniature vapor chambers. Vapor chambers are pricey, and their overall thermal conductivity isn't THAT great. Synthetic diamond can be grown and cut to order, and it's pretty cheap.

Besides, you're still limited by TIM/solder under the IHS. Even solder isn't that great. Improving the thermal conductivity of the IHS will help spread heat laterally (somewhat reducing the problems from hotspots), but you still need to improve the interface between the die and the IHS.