PDA

View Full Version : Here's how i made my peltier-block


aasmaukr
10-13-2005, 09:01 AM
On request from another member, i start a thread about how i made my tec waterblock where the water cools the tec directly.

The parts i used when i made it it was:
-A 55x55x6mm plexi block
-A 55x55x1mm copper plate (you might use plexi instead aswelll, but i like soldering better than glue, so i used copper)
-Two 20mm long 12mm copper pipes
And a 50mm tec (in my case, a 172w)

I took the plexi block, driled holes and used a saw to make a hole of almost 50x50mm in the middle of the block. Then i drilled two holes of 12mm and soldered the two copper pipes in it.

Then i glued the plexi parts on the tec, and the copper top on the plexi.
The glue i used was a norwegian glue called "superfix", made by casco if i remember right. I think that it's being sold in a few other countries too, although i am not sure. I think that silicone will do too, if you get a bit pressure on the block (a coldplate on the colde side of the tec).

Here's some pics of the block before it's put together:

http://home.no/aasmaukr/tec_1.jpg

http://home.no/aasmaukr/tec_2.jpg

http://home.no/aasmaukr/tec_3.jpg

Here is a pic of the block in action:

http://home.no/aasmaukr/1_small.jpg (http://home.no/aasmaukr/1.jpg)

I haven't tested it in my computer by now, and i can't do it before about three weeks, i think :(

I hope this is interesting for someone, at least i was a extremely easy block to make :)

Edit: Pics resized

Magnj
10-13-2005, 09:13 AM
wow i can resize them 4 you if you want

Holst
10-13-2005, 09:15 AM
Thanks allot mate, very helpfull :D :D

It was the glue that I was most interested in. Silicone allways seems to leak (for me) over time, i was after something more permanent.

I wonder if marine goop would work.

I have a project in mind for a TEC chiller cluster.

If anybody knows of a reliable glue I can use as above thats avalable in the UK can you let me know please :)

Thanks.

n00b 0f l337
10-13-2005, 09:24 AM
Only problem there is your losing efficiency. You need clamping force on the tec between the coldplate and hotplate. Otherwise your idles will be as bad as your loads.

Holst
10-13-2005, 09:29 AM
Only problem there is your losing efficiency. You need clamping force on the tec between the coldplate and hotplate. Otherwise your idles will be as bad as your loads.

I allready have a work around for this :D

im going to play zombie CS:S now (i love it, playing as a zombie) but ill be back later with some pics of my design using assmaukr's ideas.

This is for a multi TEC chiller, with dual water loops.

aasmaukr
10-13-2005, 10:36 AM
@ Magnj: I've done it now, i did not have time to do it when i posted the first post. But thanks for the offer.

@ Holst: No problem. I am looking forward to see your drawings.

Edit: N00b of 1337, does that mean that my temps will suck arse with a heat load? With a water temp of 27 degrees c, the cold side of the tec went down to -29.

Holst
10-13-2005, 10:51 AM
No, not necicarally.

The main issue with your block is that once you put it onto a CPU/GPU you will struggle to get enough clamping pressure.

TEC prefer to be clamped at 10X what you clamp a CPU at, this way you get the best heat transfer.

I think the design your using with the water in contact with the TEC surface is pretty good, its the other side that needs a little work (as you have not done this bit yet)

n00b 0f l337
10-13-2005, 11:26 AM
Yes, I cant wait to see your project holst, seems me and you are the only ones who really know our tec info.

wittekakker
10-13-2005, 04:35 PM
Last time I did some TEC action was with waterchiller. I just took TEC and colplate, put it under my cpu waterblock and used the waterblock bolts to clamp everything on cpu die. I didn't clamp to much because I didn't want to crush my old atlon xp's core. With no load I got on coldside -37°C.
With CPU idle coldside got -8°C, and with load it went over 0°C.

U really got to give enough pressure to get nice temps.

afireinside
10-13-2005, 05:19 PM
Yes, I cant wait to see your project holst, seems me and you are the only ones who really know our tec info.

Well then why don't you know that with a direct water to TEC hot side block that theres nothing to clamp the hot side to? :lol:

n00b 0f l337
10-13-2005, 06:49 PM
I already said that fire. Dur.

Mr. Tinker
10-13-2005, 07:15 PM
yeah afi, dur.

KaptCrunch
10-16-2005, 05:30 AM
Aasmaukr

best use 100% shrinkfree silicone an can take the ambiant temp range for purpose using it.

aasmaukr
11-13-2005, 12:09 AM
I am testing the block right now, and it idles at -6 degrees celcius with a water temp. of 21,7 degrees, i am gonna test load soon and edit this post. :)

Load: 6 degrees c, water temp was 24,3 degrees c.


Check this post (http://www.xtremesystems.org/forums/showthread.php?p=1133714#post1133714)

Cra$h 0veride
11-24-2005, 07:12 PM
I'm considering doing something similar, but first, what is the general opinion about the clamping power that this block provides, how does it compare to the swiftech blocks which are also water directly on the tec ?