Strange thinking.......

[Silicon Skum]

OK, here's the deal, I have an idea forming in my head about using Peltiers in a water cooling setup. Now this is going to a slightly different setup compared to most.

What I plan to do is to remove some of the heat from a standard pelt + wateblock setup on a CPU by using a 2 peltier stack. The one in contact with the CPU is hooked up as normal, the second Peltier is going to be PASSIVE, by which I mean there is NO voltage applied to the peltier, instead it will be used in reverse, ie. a thermoelectric generator ( heat from the first TEC will be on one side, the other side cooled by the water block = temp gradient, peltier effect produces a positive ve) The leads from the 2nd TEC will be hooked up to a load, a bulb, fan, lead-acid battery, motor etc.
By using the 2nd TEC as a generating source, SOME of the heat will be converted into an electrical voltage. As long as the TEC is connected to a load there will be a complete circuit, so the TEC *should* still be able to conduct heat through it, so any excess heat not converted into an electrical potential sould pass through the TEC in to the waterblock.

The idea is to convert some of the heat directly into electricity and dump it into a heavy load, thereby I should have LESS heat being pumped into the water block, so cooler water temps = a higer temp diferential = more voltage produced by the TEC to be wasted in a load = removal of heat from the system. Now this is based on the idea of energy conservation (energy cannot be destroyed ONLY CONVERTED) so If I'm converting some of the energy (heat) directly into electricity I should be removing heat from the system as electrical voltage.

Now I know the peltiers are really inifiecient, but this should still be able to lower the temps of both the primary TEC (in contact with CPU) and the water block which dumps the wasted heat in to the rad which then radiated it into the air. Or it should atleast remove heat from the watercooling side of things, alowing for lower water temps and possibly reduced radiator size (more effiecient cooling?).

It seems like a sound Idea in my head, converting phonons (quanta of heat energy) into electrons therby preserving the law of energy conservation (energy conversion not destruction). Of course this is in a best case scenario, ther are probably a LOT of inificiencies in this system, but still it DOES seem to have posibilities.......

OK, guys, does this sound like a reasonable idea or what? Anyone out there who has a better understanding of the thermodynamic principles?

§
_________________
my sig disappeared from the image host (?)
But at least I have a Josh Award!

[Google]

[Ham_fisT]

you're talking about using a tec to convert heat into electricity, kinda like spinning a dc motor by hand generates a current? I didn't know a TEC could be used like this.

if you found the gains to be minimal, you could always hook the passive tec up to some sort of voltmeter, (instead of a light bulb load), and maybe get some visual representation of thermal conduction, provided you still have enough heat being absorbed through the passive tec to keep the CPU cool.

anyway, sounds like a good concept......keep us posted
_________________
Yeah....... ok

[2old2care]

The ineffeciencies will be troubling....for the "passive" tec to conduct thermally worth a diddle the "load" will have to be very low impedence....closer to a short circuit than a load. At least this is the way I'm thinking...if you got a couple tec's sitting around, mock it up and experiment a bit.
_________________
.
Liquid-Cooled Q9450 and an EeePC
.

[[KoG]^weaZel]

the passive TEC would insulate too much and the active TEC would overheat and die. Then the cpu would overheat.

The example that Ham gave of turning an electric motor by hand creates a current. And just like that example its too inefficient to work sufficiently.
_________________
I tweaked and it tweaked back! So I Tweaked some more!

"Barney is like the Michael Jackson of PBS." - James Tybeerious