Re:Startup/Heat Transfer (Score 1)

I've been to Sandia and I've seen the prototype. The heat is drawn away from the chip die by a conventional spreader (could be aluminum or copper), which would likely be coupled to the chip die with thermal paste. The base plate is one half of the air bearing, and the rotor/impeller is the other half. There's a hub in the middle consisting of a stator and rotor. The air bearing is kept in a dynamic stasis by magnets (probably as part of the motor, but I didn't ask about this detail). This ensures that the air bearing gap is maintained within the required close tolerances and also means that the cooler is attitude-insensitive. You can run it upside-down. I do not know how shock-sensitive the air-bearing is, but I would be surprised if it were especially delicate.

As for the efficiency of the air bearing as a thermal couple, I would refer you to the official scientifical wordage at: http://prod.sandia.gov/techlib/access-control.cgi/2010/100258.pdf. The bottom line is that stagnant air (like on a static heat sink) is a pretty effective thermal insulator. Air moving around (as over whirling blades, or in the air bearing itself) is pretty good at conducting heat. If the fluid dynamics and thermal analyses they've done are correct (and "they," meaning Dr. Koplow, are not prone to being wrong about such matters), both the cooling effect and the heat transfer across the air bearing are quite favorable.

I asked Dr. Koplow about the "grinding" effect when the air bearing is not established (powering up and down), and he explained that while this is not a significant issue (both surfaces are pretty smooth and won't be in contact for long, if at all) it is going to be managed in a production context. I can tell you that the solution is obvious and intuitive, but I'm not sure that I'm at liberty to discuss exactly how they licked this problem, so I'm going to have to hold the line at, "ya gotta believe me, it works."

As for the reappearance of this story at present, the news peg is that they've actually licensed the technology now (but aren't saying to whom). I heard rumors of interest from very large players, but note that only two licenses have been granted, apparently for lighting and chip cooling applications. I am dying to find out who.