More later because this discussion is great.  But just a note on Carnot 
efficiency.  The formula is:

eff.  = 1 - (T.cold/T.hot)

These are absolute temperatures so add 460 to deg. F or 273 to deg. C.

Example
Car air temp desired 60 deg F = T.cold
Heat rejected to outside air at 250 deg F = T.hot

eff. = 1 - ((60+460)/(250+460) = 0.27 or 27%

There is no firmly defined "theoretical limit" except it must be less than 1.
The limits result from material capability and other practical limits (i.e. 
nobody wants -30 deg f air and accompanying ice in an A/C).

Gotta go for now.

Charles Vaughan
'84 GL

>Interesting stuff about the Peltier devices and the Rovox Technology.
>
>It seems that the Rovox technology is still based on the Carnot cycle, and, if
>my memory of those horrendous thermodynamics courses is accurate, the
>theoretical limit of efficiency to the Carnot cycle is 25%, i.e., you have to
>put in 4 units of energy for every unit of energy you suck out of the
>chamber that is being cooled.   In general, if I'm not mistaken, the
>efficiency of air conditioners run around 12-15% (my data is probably old
>and misunderstood -- please correct if so)...
>