|Direct WaterCooled peltier|
Direct Water Cooled peltier :
We don't know if anyone else have done this before - we have checked around but haven't found this idea anywhere else (but we might be wrong -- mail me if you think I stole your or anyone others idea !). Anyway we decided to make a "direct" watercooled a peltier !
Peltierblock - you don't need to make it of copper - iron /aluminum or brass will do just fine. If you don't have access to a milling machine I can't imagine that a mechanical workshop will charge a lot to make a block like this. I made it in half an hour ( I didn't exactly care about the finish since this is just a test !)
The problem: To find a glue that can take the heat the peltier pumps out - this glue is guaranteed to 90'C (194'F). So we decided to give it a try.
Add glue on the edges of the waterblock.
Install the peltier and align it with the waterblock , use a clampvise with a thick shim underneath (we used a piece of aluminum) to get a even pressure on the whole peltiers surface.Thereafter we left it to dry for three days.
Torturing the glue:
A peltiers surface is made of ceramics - which are 100% waterproof so it will never cause leaks nor will moisture be able to pass. The only thing we were afraid of was if the glue could stand the heat : What you see at the picture above is after we removed the cooling lines, as you can see the water in the cooling block is cooking. We didn't turn of the power before all the water was evaporated. Afterwards we attached the cooling lines again and blocked the outlet line with a Eheim 1048 pump running. No leaks whatsoever. Then we tried the sucktest - also known as the Linda Lovelace test :) With the waterblock filled with water and one of the outlet blocked we sucked the open line until we almost fainted. If the glue had failed the torture test airbubbles would have entered into the waterblock and thereafter into the open line which was partly filled with water. ( A very sophisticated test !)
The peltier failed during the torture test : see the corner ( lower right) which is not covered with ice ? This spot is +52'C (125'F) - that a peltier can take partly damage was new for us .
This glue is strong - we had to destroy the peltier to get it of !
Update: we decided to create a waterblock with integrated coldplate to avoid peltier and copper plate sliding around - not to forget the perfect pressure this creates between the peltier and coldplate :
Take note of how we direct the waterstream. - And yes I cared about the finish this time since we now know that DirectWaterCooled peltier is superior to attaching the peltier to a "normal" waterblock !
OcShoot's Direct Cooled water block with cold plate attached !
Go here to see our "neverending" ? Water Cooling project !
Cheaper and easier - Tims watercooled peltier here