I'm in the process of making a new 3x50mm chiller

[Ultrasonic2]

i thought id share with you guys what im up to0

This is my new chiller the compete unit will fit in a 5.25 drive bay and in fact if you want more cooling like me you can fit another one next to it.

China Chiller.JPG

[KrowniK]

Looks pretty cool! Did you get yer mill back up and working again? Last I heard it was down.

Also, how hard will it be to find high matched TECs?

[Ultrasonic2]

so far

[Ozzfest05]

Did you rifle those holes? If not you may be able to get ext lenght tap. Will help efficiency.
Will be interesting to see as i didnt think TEC was practical at this point..

The first thing that comes to my mind here is to duplicate a second copper heat exchanger.
Identical to the one you have,sandwich mount them, insulate them. One block has coolant running through the other has refrigerant.

A small ss with R22a and a wateecooler compressor would be quiet and sufficient cooling for a chiller.
And way more bad @ss than a pelt. imo

[Ultrasonic2]

Actually now is the best time for TECing in probably the last 8 years. With descent CPU’s only consuming 77watts stock

[Ozzfest05]

Actually now is the best time for TECing in probably the last 8 years. With descent CPU?s only consuming 77watts stock

Intel 3770k at 4.5 ghz draws about 220 watts.

[Ultrasonic2]

Intel 3770k at 4.5 ghz draws about 220 watts.

220 is far more than what the cpu calculators suggest. It doesn't really matter thats still less than my current 1055t at 4.2g that im writing this on that's tec cooled.

[Liam_G]

Glad to see you're still working with TEC's ultra. Which 50mm TEC's are you going to use?

Personally I think there is only 4 TEC's worth playing with from customthermo's list right now:
http://www.customthermoelectric.com/...W_spec_sht.pdf
http://www.customthermoelectric.com/...Y_spec_sht.pdf
http://www.customthermoelectric.com/...W_spec_sht.pdf
http://www.customthermoelectric.com/...W_spec_sht.pdf

My favourite out of those is the 127.7w 59x59mm 34.4v 6a one, lots of surface area and low amps. running 4 of these at up to 24v or 6 at up to 15v works really well for 20*dt. the 127.5w version is basically the same but only 50x50mm.

Also I have modeled up a block in solidworks for 6 of those 127.7w 59x59mm TECs, going for maximum heat transfer at the expense of restriction and cost of manufacture. I have used a large diamond pin matrix, like each pin is 3mmD x 4mmW x 4mmH so it's not super restrictive but it provides plenty of surface area and a decent amount of turbulence. I have also used jet plates at the end of each block to spread the flow equally across the 59mm width. As always with me it is theoretical so I have to commend you for actually getting out there and building your setups, not long now before I can get into the building phase, hopefully early next year.

Here is my 6 TEC block design:

[Ultrasonic2]

i've been doing TEC's for for like 16 years now and i doubt i'll ever stop it just takes me alot of time to come up with some cash for an expensive hobby i have. i've also ordered more arduino parts to play with over the summer

yeah big surface area is of a greater importance then people think

i have 245's. i will have 2 of those chillers for 6 of them the chiller was designed so i could fit 2 of them side by side in a single 5/14 bay mind you i've already changed my mind so im not sure what i'll actually do with them


you picture didn't work can you fix it i'd love to see it..

[Liam_G]

i've been doing TEC's for for like 16 years now and i doubt i'll ever stop it just takes me alot of time to come up with some cash for an expensive hobby i have. i've also ordered more arduino parts to play with over the summer

yeah big surface area is of a greater importance then people think

i have 245's. i will have 2 of those chillers for 6 of them the chiller was designed so i could fit 2 of them side by side in a single 5/14 bay mind you i've already changed my mind so im not sure what i'll actually do with them


you picture didn't work can you fix it i'd love to see it..

yes i've become addicted to TEC's ever since i started reading your thread last year, every time i have a few spare moments in front of the PC i find myself perusing any tec chiller threads i can find from the last decade or so and going over the tec's at custom thermo and then going over the efficiency and cop and heat load it can handle etc, i'm itching to build my set up.

ah yep, i see the 245 you're using. i like those ones i linked because of the low amps, like i would run the 127.7w tec at 15v 2.5amps x 6 tecs for about a 20*dt on 270 odd heat load and only about 225w power consumption. not accurate but close as i can get with just numbers not testing.

thats the fun of this, when you have a great idea you work out all the detail and you can then have fun making up another idea :P

look forward to anything you come up with re the arduino tec controller

i can't seem to get the picture to work but if you right click it and open in new tab it should work. it would have to be machined professionally though i think, so $$$$$$

[Ultrasonic2]

looks pretty good

You may find you'll get better performance if you offset the inlet and out lets. also the plate may be just reducing performance.

However how you test it in solid works can make a huge difference on what's best

[Liam_G]

I've tried quite a few different set ups with and without the jet plate, tried various techniques to get the flow even across the whole block. the diamond pins have a tendency to funnel the flow straight down the row, ie it doesn't like changing the row it first started in, so the offset technique may or may not work ( i don't put much faith in solidworks floxpress so it very well could though). the jet plates definitely give me the most equal flow over the whole block however it does negate the velocity, i have some other designs in play to try and spread the flow without interrupting the flow in such a drastic manner, ie trying to get it to flow straight through rather than into a plenum and then down through the jet plate.

[Ultrasonic2]

i have the full version of flow works.

the problem people have is they apply a pressure to the inlet or a volume both produce wildly in accurate results.

When applying a pressure solid works will jam an unlimited volume though it as there is no restriction on that. and there for the best performing block is a free flowing design

if you apply a volume then the reverse is true and a very restrictive block produces the best performance.

Now with a pump there is both a limit on Volume and Pressure and the relationship it not a straight line. Getting this right is vitally important. in producing a block that actually performs as you thought it would

The other thing to think about is in most cases perfect flow distribution will not proved the best cooling performance as heat distribution is not always perfectly distributed

[Liam_G]

I might have to send this thing to you to run some simulations on it correctly.

I was under the impression that tecs spread heat pretty evenly across their surface with perhaps a hot spot in the middle?? Hence my design to get it even across the whole width of the block, although there will always be slightly more flow through the middle of the block with my inline inlet/oulet setup which might negate that hot spot, kind of a good thing, just ran a simulation with no jet plate and angled plenums to direct and spread the water, seems to work better already :P

[Ultrasonic2]

i wanted to upload some sim pic but am unable to because it says they aren't a valid format i have tried all the formats than they dont work ?

[Liam_G]

oh, XS is stuffed with pics right now, i had this problem the other day. just upload your image to photobucket and link the url, it will just show up like mine did the other day, just a tiny little tab, but if you right click on it and click on "open in new tab" it will show the pic on photobucket. look forward to seeing the results.

[Ultrasonic2]

here we go