[Neuromancer]

If you water cool those tecs, that would be sweet. use one quad for the tecs. tecs straight to CPU. 2 quads for a GPUs... If you do add a quad for the cool side of the TEC loop, try it with no fans then low speed fans first

There is always the option of insane TEC cooling as well. Say 200w?

It is $45, but bear in mind it needs a power supply/regulator those uh, 24v maxwells (?) that people started using for GPUs a few years back

http://www.newark.com/multicomp/mcte...00w/dp/08P1466

another consideration, getting some 15V units and just running them at 12v (they dont sell 12v)

Costs a bit more, on tec side but you could actually run them off a PSU... 2x15v 130W TECS at 12v = probably close to 200W cooling without needing separate PSU...

Yeah I know it wont look as cool as the Boreas either.

[Witchdoctor]

I have Zero experience with TEC's ...........

So the biggest hurdel is learning how they work to be able to impliment action. Right now the plan is to take them off the Boreas, use two of the 4 to cool and the other two disapating the heat off the twoi cooling TEC's ... Just sandwitching them if you will to meet surface requirments, if that makes any scence at all, problem is I will cut my capacity in half and in doing so will the water heat disapation make that much of a difference gain back 100% of the performance as that would be the break even to air coupled, TBH I would be happy with that or a little better. Right now the unit does 15C better than ambient water, not sure if you saw my results table a few pages back, but here it is again if you didn't. Being I am not the sharpest knife in the drawer is there any math that can give me the answers I need based on rad surface, air speed across the rad and pump flow rates based on the output wattage of the TEC's and if so does anyone here know what the formula would look like if I could plug numbers into it ?

One thing is for sure Rich, if it works I will make it look good doing it ........