Jump to content


Photo

DIY Peltier Chiller for $20-$60


  • Please log in to reply
58 replies to this topic

#26
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005

I'm building a peltier cooler for my end of year engineering project. Using a 118W TEC, old heatsink/fan which was used on an athlon xp 2200+, small pump, and a water cooling block...
I'm having trouble choosing a good cooling block however; copper quite obviously gives the best heat transfer but it's also no good when it's coming into contact with my salt water! Would aluminium be ok or maybe even duralumin? Titanium is an option I guess. Yes I will be keeping corals btw...

Oh the cold side on the peltier will be to the cooling block, and the hot to the heatsink/fan. The tank water will be forced through the cooling block via a pump and then back to the tank. All this will be temperature controlled with a temp probe. Say if it gets too hot the cooling device switches on until the desired temperature is reached.

Edited by shepz, 10 October 2005 - 02:57 AM.


#27
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005
Aluminium will corode in saltwater. You either have to go w/ titanium or you could source an anodized water block from Chris @ nanotuners.com.
Posted Image

#28
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005
It will have to be the titanium then. Just worried about constantly having to anneal a block of titanium when i'm trying to mill it :(

#29
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005

It will have to be the titanium then. Just worried about constantly having to anneal a block of titanium when i'm trying to mill it :(


Ah, I can only dream of being able to do that!

One thing you might consider since you are going to mill the block is to include some type of channeling internalally to 1) increase surface contact area between the water and metal and 2) increase turbulance. Both of these improve the efficiency of heat transfer. Ive seen CPU waterblock designs that use projections or "pins" on the inside. I mean, if you are going to mill titanium, you might as well go all the way! :D
Posted Image

#30
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005

Ive seen CPU waterblock designs that use projections or "pins" on the inside. I mean, if you are going to mill titanium, you might as well go all the way! :D


If I was doing it out of my own spare time I would knowing that there wouldn't be any extra maths/physics involved; however with the course project you have to justify every detail of the final design... with the mathematical evidence and practical evidence :(
I've got 90 hours to do everything in... oh and a 50 college budget! It's going to be real hard, but hopefully fun at the same time :lol:

#31
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005
Oh i've completed the cooling block design now...
What do you think?

Attached Thumbnails

  • finaldesignofexchanger.JPG


#32
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005

Oh i've completed the cooling block design now...
What do you think?

Very nice! :) :)
Posted Image

#33
neanderthalman

neanderthalman

    Two hands working can do more than a thousand clasped in prayer.

  • Members
  • 10,943 posts
  • Joined 11 Mar 2005
  • Pickering, ON
This was a course project? What are the reasons for the "nE" shape?
Posted Image
^^New Coral 2010-03-27^^

Posted Image

I looked down her shirt. *shrug* I ain't gay or nothin', but it's boobs. It's what you do.


#34
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005
The purpose is all to reduce pressure/flow rate. The inner corners of the "N" are not rounded off... this will reduce pressure and also the "E" will kill down on pressure also! You don't want fast water movement nor high pressure because it:
1)won't be as efficient in heat transfer
2)you don't want the fish to suffer from any thermal shock... this way the cold water will mix in with the hot water more effectively before it has a chance to shock inhabitants...

Well that's the idea anyway, I may be wrong! :lol:

#35
Touk

Touk

    Nano Reefer

  • Members
  • 25 posts
  • Joined 03 Jan 2006
Wish to share the photo of a peltier chiller that I made for my 5G freshwater tank. I'm still working on improving the aesthetics.

http://www.arofanati...z/peltiercooler

As I can't find a suitable aluminium water block, I used an anodized PC aluminium heatsink dunk in the tank water to transfer the coldness to the water. It cools the tank water from 85F to 79F. The peltier is running at 9W and I have it on 24hrs as I stay in sunny Singapore and temperature here is warm (80 - 89) and humid year round.

Currently, I am looking to DIY a 10G acrylic tank with an integrated peltier chiller, as I read that arcylic help maintain water temperature better. But I'm having problems getting acrylic adhesive in my country. IPS Weld-on products cost $100+ here and I can't find a store in US that ships international. Any suggestions?

#36
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005

Currently, I am looking to DIY a 10G acrylic tank with an integrated peltier chiller, as I read that arcylic help maintain water temperature better. But I'm having problems getting acrylic adhesive in my country. IPS Weld-on products cost $100+ here and I can't find a store in US that ships international. Any suggestions?

See if you can find MEK (Methyl Ethyl Ketone), That is really all Weld-on is. To make it a bit thicker, all you have to do is dissolve a few chips of acrylic in some of the MEK (takes a few days). Works like a charm.
Posted Image

#37
Touk

Touk

    Nano Reefer

  • Members
  • 25 posts
  • Joined 03 Jan 2006

See if you can find MEK (Methyl Ethyl Ketone), That is really all Weld-on is. To make it a bit thicker, all you have to do is dissolve a few chips of acrylic in some of the MEK (takes a few days). Works like a charm.


Thanks for the info :D
I'll look for MEK then, wish me luck :)

#38
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005
9W at constant??? that's crazy!
I'm using a 118W peltier working on a "thermostat"...

Attached Thumbnails

  • peltiermod.JPG


#39
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005

9W at constant??? that's crazy!
I'm using a 118W peltier working on a "thermostat"...

Could you elaborate for those of us that are schematically challenged?

Thanks!
Posted Image

#40
shao-lin nano

shao-lin nano

    Kung-Fu Reefer

  • Members
  • 1,418 posts
  • Joined 14 Mar 2005
  • So Cal: Rowland Heights
I have a Titan 150 (a thermoelectric chiller). I don't know how large the peltier is on it but the heat sink is enormous, like a little smaller than a brick. I guess the challenge to make it work most efficient is finding a good way to remove heat from the hot side.
my tank knows kung-fu
Far from stock 24g NanoCube

#41
Touk

Touk

    Nano Reefer

  • Members
  • 25 posts
  • Joined 03 Jan 2006

I have a Titan 150 (a thermoelectric chiller). I don't know how large the peltier is on it but the heat sink is enormous, like a little smaller than a brick. I guess the challenge to make it work most efficient is finding a good way to remove heat from the hot side.


I've tried several ways of removing heat from the Peltier before finally settling for having the fan blow away from the heatsink i.e suck. To futher reduce temperature, add another fan on top of it.

There was a guy who posted on a local forum on how he created a 320W Peltier chiller. He ran it at 200W and it is water cooled as it would be impossible to cool that amount of heat efficiently using a fan. He managed to bring a 20G tank from 85F to 66F. Needless to say, if I can get my hands on some water blocks, I'll be trying that out too...

#42
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005

Could you elaborate for those of us that are schematically challenged?


Ok... once the probe picks up the peak temperature (around 29C) it activates the peltier thus the cooling process begins... once the water is cooled to a level below 29C the peltier turns off. Pretty simple concept although you have to give way for the fact that when the peltier turns off it will be heating the water a little; since the heat from the hot side will transfer to the cold side and thus onto the heat exchanging block.

Attached Thumbnails

  • rough_diagram.JPG


#43
doncb

doncb

    11/4/08 Flag of the Union of Socialist States of Amerika

  • Members
  • 1,775 posts
  • Joined 11 Mar 2005

Ok... once the probe picks up the peak temperature (around 29C) it activates the peltier thus the cooling process begins... once the water is cooled to a level below 29C the peltier turns off. Pretty simple concept although you have to give way for the fact that when the peltier turns off it will be heating the water a little; since the heat from the hot side will transfer to the cold side and thus onto the heat exchanging block.

Oops! Sorry. I meant elaborate on the schematic. Component list, wiring diagram, etc.
Posted Image

#44
nanoguru

nanoguru

    Boston, MA

  • Members
  • 145 posts
  • Joined 17 Feb 2006
  • Boston, MA

Oh i've completed the cooling block design now...
What do you think?


u gonna make that in Ti??
Make me one! haha

nano

#45
skripo

skripo

    Nano Reefer

  • Members
  • 172 posts
  • Joined 03 Mar 2004
Why can't we glue the cold side of the plate to the back of the glass rather than drill the tank? Glass is a great heat conducter and I don't understand the need for the probe part.

In fact, I believe it would be MORE efficient because there would be no loss from the peltier device to the probe part and the TEC would be in full contact with the glass.

I would glue it using the same glue that is used to mount rearview mirrors to windshields (essentially high quality super glue ).

Any thoughts?

#46
shepz

shepz

    Nano Reefer

  • Members
  • 20 posts
  • Joined 18 May 2005

Why can't we glue the cold side of the plate to the back of the glass rather than drill the tank? Glass is a great heat conducter and I don't understand the need for the probe part.

In fact, I believe it would be MORE efficient because there would be no loss from the peltier device to the probe part and the TEC would be in full contact with the glass.

I would glue it using the same glue that is used to mount rearview mirrors to windshields (essentially high quality super glue ).

Any thoughts?


Well yeah I do, glass is also a good insulator especially after you've used glue strong enough to take the weight of the peltier and the heatsink lol.
There would be no need to drill the tank, I've made alot of changes to my design since and to the circuit aswell.
First of all the pump will sit in the water, it will then pump water through the heat exchanger back into the tank. The probe is simply a digital thermometer acting as a thermostatic input device for the controls. Once the temperature of the tank gets too high the peltier will activate until the temperature has reached the desired level again.
I'm going to start a new thread once I've started getting the project going besides the maths and everything else! My college still haven't sourced me with my titanium! :angry:

#47
skripo

skripo

    Nano Reefer

  • Members
  • 172 posts
  • Joined 03 Mar 2004
I keep seeing statements about glass being a good insulator yet all my experience with it tends to make me believe the complete opposite.

Found this data:

http://www.shef.ac.u...y001/unit6.html

#48
neanderthalman

neanderthalman

    Two hands working can do more than a thousand clasped in prayer.

  • Members
  • 10,943 posts
  • Joined 11 Mar 2005
  • Pickering, ON
From the link you posted.

6.2 The glass of the window of a room is 5 mm thick and has an area of 4 m2. If heat can only escape through the window, calculate the electric power required to keep the room at a uniform temperature 20C above the external temperature. Repeat the calculation for a double glazed window consisting of two 5 mm thick glass sheets separated by non-convecting air 2.5 mm thick.
[Thermal conductivity of glass is 1.5 W m-1 K-1, of air 21x10-3 W m-1 K-1]

*6.3 A copper block of mass 200 g is suspended in a vacuum by a copper wire 10 cm long and 1 mm in diameter. If the copper is initially 100C above its surroundings, find the rate at which it begins to cool, assuming that radiation can be neglected.
[Specific heat of copper is 380 J kg-1 K-1,
Thermal conductivity of copper is 400 W m-1 K-1.]


Let me summarize.

Thermal conductivity of glass is 1.5 W m-1 K-1
Thermal conductivity of copper is 400 W m-1 K-1

See the difference?
Posted Image
^^New Coral 2010-03-27^^

Posted Image

I looked down her shirt. *shrug* I ain't gay or nothin', but it's boobs. It's what you do.


#49
skripo

skripo

    Nano Reefer

  • Members
  • 172 posts
  • Joined 03 Mar 2004

From the link you posted.
Let me summarize.

Thermal conductivity of glass is 1.5 W m-1 K-1
Thermal conductivity of copper is 400 W m-1 K-1

See the difference?


That was my point. Intutively you would think it is a good conductor but it isn't. Ordered a DIY Ice Probe yesterdaywhen I found that link.

Anyone else want an Ice Probe for $75.00, power supply included?

#50
bigbabich

bigbabich

    Nano Reefer

  • Members
  • 628 posts
  • Joined 25 Jan 2006
  • Newburyport, MA
I pay my own electrical bill.
I'm thinking a real chiller will be in my best interest price wise.
I've been thinking of using a peltier, as I have one laying around from when I was trying to go fanless/overclocking on my game machine. But considering the price per gallon in terms of electricity it's not making lots of sense for my aquapod 24.
Cool project though.