One part of their FAQ jumps out as being simple, total BS: "Do CoolIT systems require power supplies with greater than normal capacities?
No. While the TECs do use power under heavy load, the system reduces power consumption at the die level due to lower operating temperatures.
" That my friends, is BS. There's no way those TEC's aren't running when the system is on, I'd put money on that without ever even having seen one first hand. (there's no low-resistance thermal connection between the WB's and the heatsink - the TEC's are in the way) And the tiny wattage reduction from running the CPU 10-15° cooler is nothing compared to the (probably) 200W of TEC's they've got.
But having said that.....I'd still like to see one.
In fact, now that I have studied TECs a bit more, I might give some value to their claims.
Lets first define a new term: CoP, coefficient of performance, is used to measure effectiveness of thermal pumps. It's in essence amount of heat pumped (in Watts) per amount of power fed to the thermal pump (in Watts, again). Compressor-based coolers (fridges, freezers, etc) usually have CoP between 2 and 9. TECs traditionally have had CoP of 0.4-0.7 (That is, they produce 2-3 times the heat they transfer).
The catch is that physics behind TECs are non-straightforward, and the CoP is greatly affected by the temperature gradient and voltage. Here you can see interesting image from Electronics Cooling
that shows relation of CoP and voltage (in percent of maximum rated voltage) in certain TEC modules:
As you can see, TECs can operate at CoPs greater than 1, even up to the 4-5 while still providing some useful temperature gradient.
So, to show what I think is the idea behind those marketing claims, I made two systems with Kryotherm TEC calculation software
. These systems are identical in thermal load and heatsink/fan efficiency, the only difference is that the first one uses single TEC, the second four ones. Both are powered with voltage just enough to keep temperature at 20°C or below all the time.
TECs used in this simulation are Kryotherm DRIFT-0.8s, Qmax 172W, Umax 24.6V, Imax 11.3A, ΔTmax 69°C. I aimed at the same TEC cold side temp (TC), as that is what matters here; in Kryotherm simulator they are simulating coolbox, so ignore the inner HSF. I assumed here heat load of 100W from the processor (which should cover most CPUs silencing enthusiast would considre) and K/W (equals C/W in this context) of 0.2 to the heatsink and fan assembly.
Single TEC system:
Voltage fed: 17V
Power usage of TECs: 137W
Quad TEC system:
Voltage fed: 2.8V
Power usage of TECs: 3.9W
Of course those hi-power TECs cost something like 30€ here, so four of them costs 120€ up-front. But as you can see, you definitely can cool single CPU with them with the PC main power supply, as those guys claim.
Looking at their images and these calculations I assume quad-TEC system is exactly what they are using. It costs a fortune up-front, yes, but it really can give you some bizarre CoP.