It looks like I have a few new ideas for chilling.

I have a watercooled setup with a 420 Rad and I'm considering peltier cooling, I wanted a water chiller because it meant I could cool other components in the loop, if I later decide to cool them with water...

I see most everybody trying to cool the hot side with external heatsink fans or secondary watercooling kits, but, I guess they're not in my position of already having an immense watercooling setup in the first place...

I also see people having either their peltier setups on the same power supply, or a totally seperate one.

this is the setup that I first came up with and it makes the most sense to me...

The Chiller:

a simple sandwich of two Swiftech Apogee LP's on either side of a 400W TEC, maybe I'm wrong, but the 50mm block is both copper and large enough, I read somewhere that copper is needed on the ceramic plates of a TEC, and that blocks need to cover the whole surface, also, the Apogee LP has the side facing ports I need to mount the setup in a 5.25" drive bay.

the loop:

Pump > TEC - Cold > CPU > (future blocks) > TEC - Hot > Rad > Res > Pump

So as you can see, I'm putting both the cold and hot sides of the TEC in the one loop, I haven't seen anyone actually do that yet, I saw something almost similar, where the water out of the radiator was split and flowed over both sides of the TEC, but this makes more sense to me.

I saw someone have an idea about using a single fan rad as an air conditioner to lower ambient temps in the box, I like that, if I go ahead with this and the water temps out of the CPU block are below box ambient, the next upgrade would be that, I like this, I'd have it in the Side of my HAF-932 in the top right hand corner of the lower grill on the side door, that way it'd blow over my Mach Xtreme 2300Mhz RAM, chipset and intake fans on my graphics cards.

My Aquacomputer Airplex Modularity 420 Radiator is good for reducing 1200W of heat back to ambient, even a heavily overclocked CPU only uses 250 odd watts, so that plus a 400W TEC won't be getting close to the radiators capabilities.

The TEC will however nearly double the power consumption of the whole computer!

The other idea of the setup is with the secondary power supply, I've always believed that the cooling system should be run after shutdown, even if the power to a heat source is switched off, if active dissipation was needed before, undissipated heat buildup could be bad, mostly i like the idea because it's cool. but with a TEC, I'm concerned about two more things, the TEC freezing the water in the block, and the undissipated heat from the hot side damaging the unit. even though the Fluid XP Nano Fluid has a freezing point of -45

The power supply I'm thinking of using is the Thermaktake Toughpower Express 450 in conjunction with a Bitspower X-Station Universal Cooling After Module

Electrically, the setup will be:

Main PSU > Bitspower Aftercooler > Thermaltake Toughpower > Motherboard

With this setup, my Aquacomputer Aquastream XT Ultra water Pump and the TEC will be turned on at startup, but when shutdown, the TEC and Motherboard will power down, but the water pump will keep going for 1-5 minutes, cooling the CPU, warming the cold side of the TEC and cooling the hot side aswell, this will make restarts annoyingly longer, but it would be a cool project none the less.

But what is all this for? you might ask... well, the CPU can't handle 4 sticks of RAM @ 2500 Mhz, so I'm looking at cooling options to achieve this, and if I have enough cooling, I'll probably reach 5.25Ghz on the processor too.

Any critiques? I have questions about this setup...

I see a lot of people putting heatsinks on the cold side and running the water past it through a large box, how much does this improve cooling? because I'm very willing to play with a 5.25 bay resivoir, I love the idea of a Koolance TNK-501 it could be quite a good box for it. Mounting the TEC to the metal lid would act as a good conduit for spreading the cold sides effects, and using a suitable thermal epoxy I could just glue heatsinks to the inside, or just tap and screw them in... anyway, I'm gettig away from myself now...

Having the hot side of the TEC in the loop won't heat the whole loop, the last 3 feet of tubing between the hotside block and the radiator will be warmest yes, but that's not going anywhere but straight into the radiator as it is the last block in the loop, and what does that matter after all the computer components get their fill of the chilled water?, after it reaches the radiator all that heat from the top of the TEC (along with the heat from everything else) will just dissapate.

I think that's what you mean... it sounds a little recursive.

Ah, I see what you're talking about, warmer water in the block cooling the TEC results in higher temps on the cold side, I guess I took it for granted that the water going back over the TEC would be either at or below ambient, and in that situation, a second water cooling loop would either perform either equally or worse.

Provided ofcourse that the water was either at or below ambient.

Any good articles on the math for TEC's? Because I'm not sure about the math I'm using, I'm almost positive that it's missing something...

is it just as simple as saying a 400W peltier removes 400W worth of thermal energy from a medium, and that a 250W heat source would leave that medium in a state of only 150W removed thermal energy?

This article (and Krow, thanks BTW) says it is as simple as that...

So, if I have a CPU, and a low profile 1 fan radiator to use as an air chiller, and nothing else (but the TEC, ofcourse) in the loop, the 250W of the CPU and the 150W of the Radiator will make the water returning to the TEC's hot side be ambient, and essentially no different to using a second cooling loop.

Or, since the water exiting the CPU block is still worth 150W of cooling, if I returned that coolant to the hot side it would help make the TEC a little more efficient.

Right?