Peltier/Thermoelectric Drinks Cooler

Discussion in 'The Projects Forum' started by Recon, Oct 16, 2009.

  1. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    I am planning to build a drinks cooler with individual tubes for each can (8 in total). This is how I was thinking of doing it:
    [​IMG]
    The aluminium tube is just big enough for the can to slide inside. The copper block at the bottom fits snugly inside the tube.

    All the TEC examples I've seen on the internet require some kind of temperature control circuit. Would this be necessary for this project?

    I have no real experience with electronics, and was wondering what the best way to power these TECs would be. I was also planning to have each tube individually switched to minimise wasted energy. What would be the best way to construct the circuit?

    Any input into the best way to achieve this would be appreciated! How effective do you think this cooler would be?
     
  2. blueroomelectronics

    AAC Fanatic!

    Jul 22, 2007
    1,758
    98
    TEC's are pretty inefficent. You're going to need a few amps of current and some fans on a rather large heatsink to keep 8 of them cool. A simple PTC fuse could probably be used as a crude regulator.
     
  3. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    Thanks for the reply. I thought that I might need to have a couple of fans blowing over the heatsinks. :)

    I was planning to use 12V 45W TECs (360W total). Would I be able to use a standard PC power supply? Or are there cheaper/better solutions?

    Sorry for my ignorance (I really don't know much about electronics), but why would I require a PTC fuse? I just found out that they 'trip' to break the circuit if the current drawn is too great, but what would happen if I just connected the TECs to the power supply directly?
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You will need a power supply capable of supplying at least 22.5 Amperes at 12v.

    Marlin P. Jones & Assoc has an ATX supply capable of that current for under $15:
    http://www.mpja.com/prodinfo.asp?number=17512+PS
    You could convert it by plans available on the 'net by Googling for "ATX bench power supply"
     
  5. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    Did you mean 30 Amperes? I want to run 8 of them.

    I will look into converting an ATX PSU into a bench power supply, thanks.
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    OK, then you'll need at least 30A.

    Converting a couple of ATX supplies will be your least expensive route; so many of them have been made and competition is so fierce that you can get them for much less than a typical switching supply would be.
     
  7. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    Once I've converted the power supply, is it as simple as just connecting the TECs to the +12v and ground (in parallel) and putting a switch next to each? Do you think a PTC fuse (as mentioned earlier) or something similar would be required?

    Remember I'm new to this electronics stuff! :confused:
     
  8. Andrew Leigh

    Active Member

    Sep 8, 2008
    85
    0
    Hi,

    know nothing about peltiers but a couple of practical thoughts.

    1.) That sounds like a massive amount of energy for 8 cans. I am working on my caravan fridge at present, 180l capacity and it uses a 180W element at 12V, I can have 24 cans in there plus plenty other stuff. Are peltiers innefficient?

    2.) Are you planning on rapid cooling the cans or maintaining temperature? A fridge cycles somewhere between 1° and 5°C. You would want your cans at 1° to 3°C for best beer :cool: drinking temperature.

    Have you made provision for thermal insulation?

    Just a thought, as cold air falls and hot air rises, you may wish to invert your arrangement as currently you will have slightly opposing thermal "forces". So perhaps a box? and in the lid you have the peltiers and heat sink above, when you close the lid the peltiers make contact with the tube.

    Someone clever has just designed me an electronic thermosat for my fridge with an automotive relay to switch the heavy current. I will ask him if I can share the information. I have competed the board design but it is not yet made.

    The most critical thing is Δt. You have to get heat away from the peltier as efficiently and rapidly as possible so. We regularly camp in 40°C daytime weather so Δt become a problem for us, I think you will find the same for your unit, the hotter is gets the less the temperature differential and the less the cooling. For this reason I built an automatic fan controller that increases in speed as the ambient lifts, it uses a 90mm fan (can drive a bigger load). The reason for variable speed is that the fans are remarkably noisy in the dead of night, so when the ambient is cooler they reduce in speed and are not heard.

    I am curious about 8 cans .... heard of a six pack :D.

    Cheers
    Andrew
     
  9. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    I think they are quite inefficient. I do want the cans to be cooled as quickly as possible really, but it would be good if the temperature could be maintained at 1° to 3°C. Maybe I could get away with less powerful TECs, but I'm not sure.

    The reason for 8 cans (and not 6!) is that I'm going to convert a V8 engine into a coffee table (like below -- although this isn't mine) and have each cylinder as a cooling 'bay' for a single can. It isn't possible to invert the coolers.

    [​IMG]

    The 'tunnel' in the engine (where the crankshaft used to be) will have fans blowing air through across the heatsinks.
    [​IMG]
    I considered having a lid on each cylinder, but would prefer it if it wasn't necessary.
     
  10. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I have a few thoughts on this:
    Are you sure that the cooling the cans by convection through the small ring on the bottom of a typical can will have low enough thermal resistance, and will be even capable of cooling the cans in a reasonable time?

    Maybe it´s better to have the inner diameter of the cooler slightly bigger than the can, and blow the cold air around the can upwards.

    Other thing is, that the motor block is such a big piece of metal, that you don´t need to use heatsink on the TEC, and just by using a thick alu bar you can thermally connect it to the motor block, which will surely dissipate the heat.
     
  11. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    I plan to use the heatsinks as the faster I can move heat away from the TEC, the more effectively it can cool. Using the engine block on it's own probably would work, but I really need to dissipate the heat as efficiently as possible.

    I'm not sure how well this would cool to be honest. I did plan to tightly mount the copper block inside the aluminium tube (with an interference fit) to maximise heat transfer from the side of the can.

    It would be good if I could fill the air gaps around the can with a thermally conductive but 'squishy' material -- but I don't know of such a material.

    Thanks for the input. :cool:
     
  12. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    I suggest you try it on one of the tubes, and see how cool you can get it. You will see how it performs, if the heatsinking is enough etc.
     
  13. Andrew Leigh

    Active Member

    Sep 8, 2008
    85
    0
    Hi

    I think you will be up against it with that amount of steel. It will absorb 360W then stabilise at a temperature. The cooling I thought was based on the rate of heat transfer from the peltier to the heatsink. Cooling that block could take something special.

    It is an automotive theme, how about a small radiator for your cooling?

    Cheers
    Andrew

    PS: I would have the aluminium tubes thermally insulated from the block or else you will end up having to cool the entire block which you ain't going to to without huge amounts of power. Fridges have in the order of 50mm of thermal insulation to stop losses, and they are compressor driven, probably more efficient.
     
  14. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    I'm not certain what kind of temperature differential you'll get, but expect somewhere around 30°F-40°F from the cold side to the hot side of the Peltier device - that's if you have good thermal isolation between the two sides.

    Water has very good thermal conductivity; about 1.4 times that of copper.
    Copper is one of the best metals for thermal conductivity.
    Aluminum is roughly 60% as good as copper.
    Iron and steel are comparatively poor thermal conductors, at around 11% that of copper.

    Trying to use an engine block as a basis for a cooler would be rather frustrating, I'm afraid. You would have to thermally isolate the block from the beverage containers.

    In a similar vein; have you seen the "Jet Powered Beer Cooler"?
    http://www.asciimation.co.nz/beer/
    I'm in no way suggesting that building one of those things would be a good idea (hazardous on several levels), but it IS humorous and entertaining.
     
  15. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    Yes, I was going to thermally insulate the whole assembly from the engine block. The blue part of my diagram will wrap around the aluminium cylinder to separate it from the engine block. Additionally, the cooler will be mounted on a perspex sheet (not directly onto the engine block).

    [​IMG]

    I plan to get some aluminium tube and a copper block to fit inside it to test it out (I already have a TEC and heatsink).
     
  16. Andrew Leigh

    Active Member

    Sep 8, 2008
    85
    0
    Have you see those old Pentium IV heatsinks, those with the raised copper portion in the middle. If suitable those would work a treat.
     
  17. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    I'm not sure what type of heatsink you mean, but any CPU heatsink would probably do the job. I was thinking of something like this:

    [​IMG]
     
  18. KitCarlson

    Member

    Sep 27, 2009
    13
    0
    I think you might want a control system to regulate the temperature and control power consumption.
     
  19. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,535
    I liked the water idea, though it is a little messy. Matching metal to metal on the bottom is going to be next to impossible. Maybe water in a plastic bag or a small water balloon?

    My brother had a similar unit in his van that worked well. It used a heat sink similar to a CPU type, but the metal block was machined to match the curvature of the side of the can, and when it wasn't in use it had a light clamp to force contact. I figure it drew around an amp.
     
  20. Recon

    Thread Starter Member

    Oct 16, 2009
    30
    0
    I found these:

    I'd be interested to see how they work. I assume both use TECs. I guess the cylinders for the can is just precisely machined for a close fit.

    Is this necessary? Would it still work if I just connected the TECs in parallel (with individual switches) to the power supply?

    That sounds like a decent solution, but I'm not sure how durable it'd be. Plus it could freeze flat if the cooler was on without a can in it.
     
Loading...