• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

External Peltier safe water cooling

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ethanzonca

New Member
Joined
Mar 30, 2006
I was wondering if you could easily do peltier water cooling... but in a safer, more external way.

My idea is to have a waterblock on the cpu connected to a pump and reservoir. Also connected in this loop would be another cpu waterblock, OUTSIDE of the computer. On this would be mounted a peltier and a heatsink to chill the water.

I thought this would be safer and maybe better and cheaper than conventional water cooling.

Comments are appreciated! :0
 
Welcome to forums!

Replace "a peltier" with maybe 4 medium wattage ones, at least, a reservoir instead of a second waterblock and better make that heatsink huge...better yet, make it a couple of huge heatsinks.

If you really want to try peltier chilling, this would be the only real way to go about it. You want enough peltiers to be able to handle all the heat that the loop is dumping, and you want to maximize the area that the peltiers are cooling. What you suggested wouldn't effect liquid temperatures much, if at all.
 
although im a noob at water..

i had an idea about external pelts. i was originally thinking of pelting a thin area of a reservoir in which the water would be traveling or pelting the bottom of the resevoir [albeit] with a little distance from the hose connections. theoretically it would form extremely cold spot near the pelt as there is not a thick copper plate but a rather thing one seperating the reservoir from the pelt unlike the cpu which has a huge cold plate.

xp90c [found a use for it XD]
|----------|
Weak Pelt?
IN ------------- OUT
---| |----
---| |----
-------------
^ ^
|| ||
Hot Water Cold Water


And depending on how cold the water is by then. having the radiator before or after it in the loop.

that was my idea. never got around to it. xD
 
I see...

That makes sense...

This really wouldn't be that practical unless you wanted to shell out tons of cash for two 400+ watt peltiers or so... and the peltiers do put out like 2x as much heat as they move from the cold side to the hot side... I suppose if you were really dull you could water-cool the peltiers and send it into a phase-change system...

I was also wondering...

A relative of mine works for GE medical and got his hands on one of the X-ray tube coolant coolers. These supercool a special coolant to negative something. Would it be practical to just hook that thing up to a waterblock? I imagine that you could get some serious overclocks with it. :)
 
ethanzonca said:
I was also wondering...

A relative of mine works for GE medical and got his hands on one of the X-ray tube coolant coolers. These supercool a special coolant to negative something. Would it be practical to just hook that thing up to a waterblock? I imagine that you could get some serious overclocks with it. :)
Lab-grade coolers are very often an excellent choice for liquid chillers, or some times as a good base for a direct die system.
 
Gautam said:
You want enough peltiers to be able to handle all the heat that the loop is dumping

You generaly want enough pelts to be able to handle twice the heat load that they are capable of(using pelts that he max results in a dT of zero, no temp drop).

You would need several pelts run at a fraction of their max voltage to raise the efficiency as well as several quality HSFs to cool them. The end result would not be worth it though.

Grab a maze off the forums for cheap, if you can set that up and enjoy it then think about adding a pelt to it later.
 
The thing about pelts is that the hot side needs serious cooling to get rid of the heat from your system and the extra heat of the pelt. You'd need massive heatsinks mounted on your external pelt blocks. Also, a TEC-boosted CPU block isn't really a good choice for that side of the system because it's designed to pick up heat from a very small area. Why work with such a small area to transfer heat if you don't have to?

OTOH, the large peltier coolers meant for things like portable refrigerators are designed to do the sort of thing you're talking about. Check out Vonkaar's "Chrome Goat" project. Skip down to post 19 if you're not interested in in the chrome case skeleton.
http://www.ocforums.com/showthread.php?t=260669&highlight=chrome+goat
 
Wouldnt it be just easier and more energy efficent to find a frezzer and fill it with water. Then either run the loop through it or just place your rad in it? I havent dont much research but it seems that this would be more enegergy efficent compared to 2x 400+watt pelts?
 
Perium said:
Wouldnt it be just easier and more energy efficent to find a frezzer and fill it with water. Then either run the loop through it or just place your rad in it? I havent dont much research but it seems that this would be more enegergy efficent compared to 2x 400+watt pelts?
Yes, phase change is more energy efficient. OTOH, a thermoelectric chiller has the potential to be nearly silent. It might be the way to go for someone who wants extreme cooling and would be annoyed by the hum of a compressor.
 
Back