Ive heard and read several times that peltier elements need a large amount of even pressure to perform well. Could someone tell me why this is? And quantify this effect in any way.
Is it something to do with peltier internals or is it simply because its important to get a good thermal interface at the ceramic peltier faces?
Thanks
yes on all of what you said
There are several reasons to press the tec. One is to make sure all the bi-metal elements inside have a good thermal connection to the ceramic is good. Second reason is to make sure the ceramic has very very good contact to the water block and cold plate. If you get good contact on all 4 spots you will have a much better transfer of heat from the cpu to the water loop.
Ok, thanks for that. I was wondering (only theoretically) about the performance of a chiller where each side of the peltier had contact with water directly and therefore there was no real pressure on the peltier.
if you go look at one of them little water jug units that give you hot and cold water they have a tec in the middle of the tank (really it is tank cut in half and tec placed between them) and it has the tec squished as well so the hot side gives you hot water while the cold side gives you cold water. The are set on a thermostat so that the tec is only on for a short time.
I see. But in this case the tec doesnt make durect contact with the water? Its 'squished' between soemthign (probably aluminium)?
correct and if you look at a loop in a pc it never makes contact to the water that way either. Tec in water would be like putting your home breaker box in a bucket of water. not a nice site to see LOL
Well you could easily prevent any electrical problems but if it ruins the efficiency of the peltier I'll make sure I don't use this method. Thanks for the replys.
Not if it is potted. You would be fine with direct water to TEC contact if it is a potted TEC. I believe that Swiftech put out a product that did exactly this for a while.Originally Posted by littleowl
the TECs in the Swiftech chiller was never in direct contact with any fluids from what I could tell, they were sandwiched between 2 blocks basically... cooling the fluid on one side and the other side was used to keep the TEC cool.
http://www.swiftnets.com/products/MCWCHILL-452.asp
~NortH~
**---===---===---===---===---===---===---===---===---===---===---**
**---===---===---===---===---===---===---===---===---===---===---**
you are very welcome I hope I have helped.
LOL that is kind of a funny statement! how do you put water directly on the tec? Even if it is potted you still cant do it. IMO you should never run a tec that is not potted no matter what! Also if you go look around on swiftechs site they tell you that you shouldn't run a tec without it being potted.
Look around these forums - there are many people who have made peltier 'blocks' where the water directly comes into contact with the peltier element, and sealed the assembly around the edges with epoxy or similar to prevent leakage. It's certainly not an efficient way of doing it, though.
everyone I have seen here as well as other forums still use a small pice of copper for the water to travel in. If you have seen all these units then please link me to them I would like to look at it.
I've made a couple for experimenting with how the pelt performs with direct water contact. Doesn't work real well. Not enough surface area to move heat off the pelt with direct water contact.
exactly my point I have not seen it done in any way that it could even be mentioned as a cooling solution IMO
Its definately not a common thing, but i've heard of it done in the depths of the past.
I was investigating the idea of a peltier chiller whereby 10 x 4cm peltiers would get direct contact with the water on the cold side. Even though each is only a flat plate, overall you would end up with 160cm squared of peltier in contact with the water which is much better than if you are thinking about a single peltier, though still pretty pathetic.
Well basicly, the optimal pressure is as much as you can apply before the peltier actually cracks... and thats a lot.
For actual use, lets say if you tighten your coldplate to the waterblock using wingnuts, you should tighten as much as you can using your finger strength (ie. without straining yourself) until they won't go any further. its very important to put the pressure on equally, so don't turn one nut all the way and then do the next. Instead tighten one or two rounds max pr nut and htne do another one. work in a cross "X" pattern so that the preassure is equal. Applying lots of preassure to just one side will not only cause a bad mount and bad efficiency, but you also run the risk of cracking your peltier (although it would probably take a little more thna just your finger-strength to do that).
-Stigma
Thanks for the info. Can anyone estimate the thinnest plate of copper you could get away with to apply enough pressure? (ignore the fact that thickness affects heat transfer).
Dunno about thinnest, but here's a thread on optimal:
http://www.ocforums.com/showthread.php?t=370139
Posting Permissions
Reply With Quote
Bookmarks