Cold plates are the copper plates that sandwich a TEC (or Peltier, or "Pelt") to a waterblock. This is because the waterblock mounting pressure required to make a pelt effective far exceeds the safe mounting pressure on a chip. So you can't just sandwich a pelt between a waterblock and a chip; the chip would get crushed. So instead you sandwich the pelt between a waterblock and a cold plate (tightly), and mount the whole lot onto the chip at the appropriate mounting pressure. So that's a cold plate. It's a square of copper, about 4mm thick. That's all it is. The TEC or "Pelt" is a square piece of ceramic, that, when you pass a high amp current through it, moves heat from one side to the other. The result is that one side gets freezing cold, but the other side gets correspondingly boiling hot. If it is used to cool down a chip, it moves its heat to the other side as well, resulting in a cool chip, but an even hotter hot side. You generally need a special powersupply to run a TEC, and very good cooling for its hot side, as otherwise they overheat and break easily. Lower performance TECs or Pelts can be used with aircooling (i.e. mounted to a heatsink instead of a waterblock), but such models have shown poor performance, and you're actually better off with just a plain heatsink. Better performing TECs have to be used with water cooling, because nothing else comes close to cooling down the TEC sufficiently.
As Highland3r says, Peltiers usually run on 12V to 24V. But that is not the most important consideration; more important is the current (Amps) that they use. Pelts are hungry. Very hungry. So although you could in principle run a small one off a standard PSU, it is not recommended. What you really need is a dedicated Pelt PSU (which looks like a metal box roughly the size and shape of a CD-ROM drive), which will supply it with current commensurate to the temperature it needs to maintain. That means that this PSU is guided by a temp sensor stuck to the CPU die, so it can work out exactly how hot the chip is getting at this moment so it can throttle the Pelt up or down correspondingly (running it at full tilt all the time is rarely advisable for lots of technical reasons I won't go into here). Using Pelts is a very technical, tricky business. You have to consider its special power requirements, mounting and condensation issues, and the limited benefits you'll get out of it for your trouble... do lots of reading up before you consider it further.
Yeah, I'm gonna agree with Nexxo. I ran a pelt before, and I also experimented with a water chiller made with a pelt, and both were resounding failures, basically cause I couldn't cool the hot side enough for it to be efficient. It was a study in pure frustration...
well what about this, instead of put the whole cold plate in the computer instead just use the fan to blow out the hot air on the outside while the cool air stays in the case.
Unfortunately (and correct me if I'm wrong) this is extremely inefficient compared to the normal way of using a pelt. From what I understand, you want to put the cold side of the pelt inside the computer and the warm side out, with a fan blowing on each. However, not only has this been done (both custom-modded and in a physical case, it was some generic case)(also they make 5.25" bay versions), but the cool air would not be enough to substantially cool your processor much. Sorry. *completely and utterly offtopic* I cooled a Molex power supply. It was running really hot, so I plugged in a Pelt in addition to the HD running on it, and just plopped it on. I was so bored that day. */offtopic*
You will need a very large amount of power to reduce the internal case temperature significantly. That power then has to be dissipated elsewhere, which means very loud fans because you have to stop the Peltiers overheating. You can work out how much energy is needed to heat up a particular air volume - assuming your case is 35cm by 50cm by 20cm, that's an a volume inside of 0.035m^3. The density of air is about 1.25 kg/m^3, so you have a total of 40 grams of air inside your case. You need 1005 joules to heat up/cool down a kilogram of air by 1K, so 43 watts are required just to cool down the air by 1°C over a period of one second, and with a continuous-flow system (i.e. cooling air as it flows through the intake), you will need this amount of power. Peltiers are highly inefficient pieces of equipment, too - my calculations are very optimistic in that respect.
*offtopic* knowlegde is power! Forgive me if i'm being rude but seriously can you sure a fpmula, also it seems as though I need that much power than I'll go with some other cooling system e.g. Water cooling
You can indeed use a formula, which is: Q = mcΔθ where Q is the energy required in joules, m is the mass of the material, c is the material's specific heat capacity and Δθ is the temperature differential between start and end. I can redo my above calculations for a more realistic figure if you want, since they were only rough. Assume that you want to cool down the air flowing through a 30CFM fan by 5°C, which sounds reasonable to me. 30CFM is 0.85m^3/m of air, so that's roughly 0.014 cubic metres going in there per second. As I stated before, the density of air is 1.25 kg/m^3, so that means you have approximately 20 grams of air being pushed through per second. This will be in contact with the TECs for a rough period of about 0.5 seconds, so first you can calculate how much energy is required, then work out the power. Q = 0.02 * 1005 * 5 = about 100J. P = E/T so you'd need 100 / 0.5 = 200W Give the Peltier an optimistic efficiency of 40% and that's well over 300W needed
...and then you still need to slap a heatsink and fan on the CPU... Trust Hitman012's logic. His physics calculation prowess cuts through the messy bog of bad ideas like the pure, clean blade of a diamond Katana... Thing is, windwakerfan, that with the advent of increasingly powerful, and therefore increasingly hot hardware, there are many highly qualified engineers all over Asia and the USA breaking their brains over the question on how we can get more quiet, effective and efficient cooling for all that stuff. If it could be done, it would have been done by now. This is not to put down your creative thinking as many an invention has started by thinking out-of-the-box and asking: "Hey, what if...?". But it is harder than you may think.
