Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Peltier Beer

  1. Oct 11, 2004 #1


    User Avatar

    Staff: Mentor

    Just to grab your attention, check http://www.stud.ntnu.no/~arnesen/peltierbeer/ [Broken] out. This man is my new god.

    Anyway, the real purpose of this thread is to discuss how to play with a peltier device without burning my apartment down. I have ordered a 40mm, 110w (15V/10A) peltier from Ebay and a bunch of accessories (heat sink, fan, copper plate, etc). I may do the peltier beer thing, use it to cool my computer cpu, make a cooler, or use it to cool a CCD camera for my telescope (more on that one later...), or just mess around with it. What I'm concerned about is how to power it and control it.

    From what I understand (not much), a peltier acts a little like a slow inductor electrically - when you first turn it on the resistance is low and it gradually increases as the device gets up (down) to temperature. So what I'm most concerned with is frying a power supply by drawing too much amperage when I first turn it on.

    So my first question is - is the initial resistance low enough to cause an amperage spike? If so, how can I avoid that? Resistors (I probably want a rheostat anyway to vary the output)? Capacitors?

    Next (related) - if I plug a basic 6 or 9V DC power supply rated at 1-2A, will it supply the rated power or supply as much as the peltier can suck from it until it burns out? That would be a simple, low-cost (low output) alternative to a real power supply if it can work. Again, if I need to add something to keep the amperage down, I can do that.

    Any input would be greatly appreciated.
    Last edited by a moderator: May 1, 2017
  2. jcsd
  3. Oct 11, 2004 #2


    User Avatar
    Science Advisor
    Homework Helper

    Its been a really long time since I have looked at these.
    Heard there have been some minor improvements, don't know.

    Current is high 7 to 15 amps, Input voltage is low. Efficiency is low.
    No peak/spike that I know of, but think there is a minimum required current.
    Seem to recall that a constant current power source was required so it did not self destruct.
    Maximum hot side temp is fairly low about ?120c?. They melt easily.

    You need GOOD heat sinks, probably with forced air.

    They can be used as power sources as well, with heat input.

    Get the spec sheet! Have fun :smile:
  4. Oct 11, 2004 #3
    It’s been eons since I played with Peltier cells but for what it is worth. I’d use a 120vac to 12vac step-down transformer with a triac on the primary side. A timing circuit incorporating a fixed resistor, to limit maximum current, and a thermister to alter the pulse repetition rate, triggers the triac. The secondary side needs a hefty bridge rectifier but no filtering. The thermister is mated to the surface to be cooled. The aim is to maintain the desired temperature, not a constant current. Componants should be sized to 1.5 times expected load current draw.

    (Edit) The power supply I described will draw a large amount of current from the mains (50 to 70 amps?) but for a short period of time during each cycle so the main panel breaker will not trip. It will generate a lot of noise, so use a different circuit for computer power.

    (Edit 2) A device similar to a light dimmer, but for an inductive load might be modified to substitute for the triac and timing circuit.

    Last edited by a moderator: Oct 11, 2004
  5. Oct 12, 2004 #4


    User Avatar

    Staff: Mentor

    Thanks, NoTime - good info for a start. Geniere, I gotta be honest, I took one semester of EE and about half the words you used went right over my head. What's a "triac" for starters? Also, isn't AC going to cause both sides to generate heat as the voltage keeps flipping? And that seems a lot more complicated than a simple 12vdc power supply...

    In any case, this gives me an idea - I'd like to learn more about electronics, so maybe I can avail myself of the expertise here and build a controller for this thing (if I can get most of the components at Radio Shack). I bought one of THESE heat sink/fans (I want to play with heat-pipes too) and various other accoutrements. What I'd like to do for starters is make a controller that can take a 12VDC input and vary power to both the fan and the peltier with one rheostat (what's the difference between a pot and a rheostat?). Another pot/rheostat would vary the ratio between the peltier and fan voltages so I could "tune" it to specific d-T's and cooling outputs. And maybe I could add a high-temp cutoff and a reversing switch.

    I could also (maybe) use/adapt something like THIS to do most of that, plus add actual temperature monitoring (if it can handle the amperage?).

    edit, looks like that one only monitors, doesn't control. There are plenty that do control - like THIS. It looks digital and perhaps analog would be better for me to play with (ie, modify a 3-fan controller to control the fan and peltier and the ratio).

    Ideas, suggestions, help! would be appreciated. Play-time starts in about a week when the peltier arrives....
    Last edited: Oct 12, 2004
  6. Oct 12, 2004 #5


    User Avatar
    Science Advisor

    Russ, you could always rectify the AC to make it DC (which would be done inside a 12VDC power supply). But on the topic I'd say a 300W computer power supply you can pick up on sale for $15 would be by far the cheapest route aside from salvage stuff. A small wall-wart power supply for running a cordless phone or something would simply supply an incredibly tiny amount of power (9VDC .2A is less than 2W by my math) and burn up pretty quickly if you try to draw a bunch of current.


    A rheostat is just a variable resistor, a pot is that but it has a tap on the other end of the variable resistance so you can use it as a voltage divider or change the wiring of it to have it work the same with the opposite turn of the knob. Plus it seems rheostats can handle big current and voltage like a light dimmer but pots seem to be typically only for smaller signals (as far as I know, which isn't too much).

    A light dimmer seems the easiest choice, next would be a linear regulator based on a transistor (darlington to minimize components?) controlled by a pot, or a full blown variable power supply ($$$$). Dimmer seems cheap, quick, easy and might even have its own heatsink!

    Neat tempeature monitoring device link, might have to pick one of those up!


    P.S. I'm thinking of one of those extra cheap light dimmers that would be a basic rheostat. A fancy one would likely not work.
  7. Oct 12, 2004 #6


    User Avatar
    Science Advisor
    Homework Helper

    Forget about controlling the fans. Too much cool is never going to be a problem here.
    All the input power + whatever actual cooling gets done = heat. Lots of it.

    GENIERE's device is crudely speaking a modified light dimmer.
    That bridge rectifier thing changes AC to DC.
    The thermister is a temp sensor. Feedback from this turns the input power down to prevent overheating the peltier.

    Not sure what effect pulse mode operation has on a peltier device.
    I have never seen any doc on that.
    It might be a truly worthwhile idea.

    Unfortunately, building a suitable high current power supply is not a trivial task.
    Pulse mode could make it cheap though.
    Buying one will really cut into your beer money.
  8. Oct 12, 2004 #7
    Russ – If you’re not concerned about temperature regulation at this point in time, just use a light dimmer to feed the primary of a 120vac to 12vac step-down transformer. The secondary side is connected to a bridge rectifier to provide the DC current. You might look at ceiling fan speed controllers instead of light dimmers as I think (not sure) they are designed for an inductive load. I think I saw one at Home Depot for about $15.00 (Nutone brand?). Basically for manual control, you’ll need to buy only three devices, transformer, bridge rectifier, and a dimmer or speed controller. I did use a conventional light dimmer to power a ceiling fan when I lived in Texas, it was still going strong when I left 5 years later but the fan motor made a peculiar noise as it started when preset for low RPM. Later you might want to modify the dimmer to advantage feedback from a thermister to directly control the temperature. In lieu of a transformer, you might look at motorcycle battery charger which has rectification built in, but I don’t know if it’s a full wave or a half wave output.

    I think it would be difficult to modify a computer power supply to change the 12vdc output. To my knowledge the Peltier cells are very hardy beasts and do not need TLC. If you’re willing to risk damaging a Peltier cell and sniffing some smoke, you could try just using the light dimmer and a single rectifier (half wave DC output) to feed the cell. Start with minimum setting! I think these dimmer thingies use either Triacs (almost full wave output) or SCR’s (almost half wave output).

    Radio Shack is nice because they’re all over the place but if it’s not a spur of the moment thing, try to get hold of a Mouser Electronics Catalog.

    As I’m writing this I’ve recalled that motor homes often use Peltier cells for small refrigerators. Maybe it would be cost effective to buy one and gut it for the necessary ingredients.

    Good luck!
  9. Oct 12, 2004 #8


    User Avatar
    Science Advisor
    Homework Helper

    Won't work on the output side. 12V is not high enough to avalanche the diac inside the light dimmer.
  10. Oct 12, 2004 #9
    Right you are!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook