Ooh-videogames said:
All you have to do is look at the diagram and the image of the real thing.
The liquid metal heats from passing over the core, it then moves on to be cooled by the fan that also has a heatsink. The cooled liquid metal continues on to keep the core from over heating. With the air directed at the back of the case, add a exhaust fan to help pulled that hot air out the console.
This also considering the fact the liquid metal coolant has a boiling point of 2000C.
The 1T-SRAM-Q memory being used is a low power consumption design.
I know how these cooling designs work. The CPU and GPU in PS3 and 360 are going to be dissipating a lot of heat, and that's with them being parred down already. They cut out a lot of fat. Why would the packaging be so large? To get as much surface area to absorb as much heat and then use a large amount of surface area to dump it through an exhaust.
Do you think one of those little ducted fans is gonna be as efficient? You need to turn more rpms, which tends to mean more noise with these kinda of designs. The shroud helps with some of the dBs, but not enough when you're talking about a tiny little case. The Rev is supposed to be even smaller than what we've seen already. You're packing a lot of electronics in a small area, and thus generating a lot of heat to relatively small amount of surface area. Ducted fans or regular fans, the cfms is what's important to cool. Ducted fans need more rpms to achieve the same cfms as standard fans. RPMs are also one of the largest contributors to noise. With ducting, the air turbulence itself creates noise as well. It's not a matter of using a bigger fan either, b/c then all your ducting gets bigger as well.
I'm sure this is the cooling solution Nintendo will be using. A low-profile solution with a heat exchanger. But I have a hard time believing Nintendo can match the performance of Sony and MS's designs without the solutions I pointed out. You couldn't use one of these designs to cool a Cell @ 3.2GHz for instance. Not without cranking some serious rpms on that fan. So it all comes back to how will Nintendo get similar performance from BOTH chips while dissipating less power? Are they goiing to find a GPU design that idles parts of the circuit to save power? I know ATI's Mobility line is supposed to do this, but only when you crank down the graphics. Keep your laptop GPU at full performance (which you'll want) and it dumps heat like a desktop part. There's just no escaping this. And GPUs are highly paralleled parts intended to run at high efficiency. You want most of your circuit churning away.
If they even equal the 360 @ 90nm and that small form factor, then I will be amazed. AMD/Intel comparisons have nothing to do with it. These chips in the 360 and PS3 have been cut down a lot at 90nm and still run hot. As efficient as they are, they still need powerful cooling solutions.
It seems like the different solutions are being combined together here to solve the cooling problem for the form factor, and the performance is assumed to follow through some magic. I don't buy it. Not one bit. At 65nm for both chips, I'll definitely give them the benefit of the doubt, but only then. That's the only way to effectively lower voltage, and still keep transistor logic high as well as clock speeds in a nominal range. Otherwise, something has to give. BTW...heat comes from *both* sides of a PCB. It's not normally a problem in an ATX case where the GPU is the top card in direct airflow of the exhaust fans, but in a case that's supposed to be "3 DVD cases" thick...hmm. Again, it could well be a modern marvel of engineering. :? PEACE.