Originally Posted by slickwilly
Elloquin what size of rad. are you using?
Also I think the best why to explain why putting a rad in to the warm side of your cold side loop will not give good results is
The returning coolant from the CPU block while warmer than the coolant entering the CPU block is still colder then your ambient temps and thus would pick up heat not dump it, that is unless you northern boys keep your windows open in the winter, it gets frosty in upstate NY
This is what I am using now. http://www.overclock.net/peltiers-te...date-11-a.html
I was using 4 Black Ice GTX 240's with 16 high speed yate loons in push pull and an Iwaki RD-30 pump which is now the pump used in the cold loop.
Very basically once the loop pressurizes and equalizes the chiller is going to hit a max low temp determined by the adequacy of the hotside loop. Lets say -10c into the CPU block from the chiller and 5c out from the cpu block for example. Let's also assume the ambient is 20c (which is pretty close to the lab usually around 19c in there I don't freeze myself by opening the window lol that's what ducts are for
). Once all things eventually become equal (and they will) all you are doing by adding a rad into the mix is turning the rad into a heater. It will not cool below ambient as that is impossible so it will actually raise the cold loop temps back extremely close to ambient or in our example it can't lower the 5c liquid leaving the cpu block's temps and will raise them as close to 20c as possible and it will eventually warm the entire loop. It's not a hot thing it's a cold thing. If all you are trying to accomplish is ambient temps TEC's would be a waste of energy used this way. You would just use a controller which can be made cheaper than buying a rad and fans to control the chillers cold temps by PWM'ing the tecs based off a set CPU temp via software.
ElloquinEdited by Elloquin - 11/8/10 at 3:29pm