TEC / Peltier Advice.

Discussion in 'The Projects Forum' started by roblloydwales, Aug 20, 2009.

  1. roblloydwales

    Thread Starter New Member

    Aug 20, 2009
    Hi all, Found this forum whilst looking for info on TEC / Peltier devices.
    I'm hoping to build a Fuel filter pre-heater so I can run my diesel engine on vegy oil.
    All I need is way of pre-heating the oil in the filter until the engine has reached running temp.
    I was wondering if a simple 12 Volt circuit with a thermostat connected to the TEC would work.
    I just dismantled a 12V cooler box that uses TEC's and all they have is a thermostat that cuts the power when a certain temp is reached.

    Can it really be that simple, I have read about control circuits being needed, but it seems these are for systems that need temp control, am I correct in assuming this ?.
    Also, I have read that switching these devices on and off repeatedly can damage them, would it be OK to use a thermostat, where switching would be limited to once or twice a minute ?.

    Sorry for all the questions, I,ve only just discovered TEC's, was amazed to find out they've been around for over a century.

    Hope someone can help, Thanx for reading.

  2. someonesdad

    Senior Member

    Jul 7, 2009
    What is the starting temperature and ending temperature, the specific heat, and the volume of material to heat? That will give you an idea of the power needed. That, in turn, will tell you something about the load on your battery.

    I can't imagine why you couldn't just use a cheap resistive heating device along with a thermostat.
  3. rspuzio

    Active Member

    Jan 19, 2009
    Thermocouples are quite inefficient as heat pumps, so you likely are
    better off just running the electricity through a resistive heating element as suggested above. Also, most Peltier devices have a
    maximum operating temperature of 80° C or so, which is likely
    too low for your application.
  4. roblloydwales

    Thread Starter New Member

    Aug 20, 2009
    Hi, thanx for the quick reply.

    First off, Starting temp will vary due to time of year ie winter/summer.
    end temp would be no more that 90 C (vege oil viscosity is ok at this point).
    Volume of vege oil approx 1/2 litre enclosed in a mild steel (1.5mm wall) fuel filter.

    Have attached diagram if it helps.

    As for the suggested resistive heating, I have tried several setups and found that it radiates the heat in all directions, therefore i would have to build an enclosure to direct the heat, not a problem with TEC's I believe.
    Also I have never used a TEC, would be nice to experiment with one.

    Again thanx for the reply.

    • TEC.JPG
      File size:
      27.3 KB
  5. bertus


    Apr 5, 2008
  6. roblloydwales

    Thread Starter New Member

    Aug 20, 2009
    Hi, thanx guys, you have me thinking now, perhaps I need to go back to the resistive heating element and try and find a way of directing the heat.
    ThanX for your help and suggestions.

  7. rspuzio

    Active Member

    Jan 19, 2009
    Quick back-of-the-envelope calculation: Oil has a specific gravity
    of around 0.9, so the oil in your tank will weigh around a pound
    (450 g). The specific heat of oil is around 1.7 J/g°, so it will take
    around 750 J to heat it a degree. (Note that 753 J/sec is a
    horsepower for reference.) Assuming a winter case where you
    need to heat the oil 100°, that works out to 75 kJ. At 12V, that
    requires running 6 kC of charge through the heater. Converting
    units, that's around 2 A-hr, so all is well (a typical car battery has
    something like 80 A-hr).

    As for power, from the comment earlier, if we want to raise the temperature of the oil a degree per second, that will require 750W,
    which means 60A, which is plausible.

    What about dunking the heater inside the oil tank, much as is done
    with water heaters for plumbing rather than trying to heat the
    tank from the side? Since the oil is non-conductive and non-corrosive,
    I don't see any problem. For instance, you might use a loop of
    NiCr wire as your heater an put a thermistor near the outside of
    the tank to detect when the oil is hot enough to shut off the
    heater and start the engine.

    As for directing heat from the outside, maybe wind
    heating wire around the filter and then wrap it in
    glass wool or some such insulating material to keep
    the heat from escaping.
    Last edited: Aug 20, 2009
  8. roblloydwales

    Thread Starter New Member

    Aug 20, 2009
    Hi rspuzio, ThanX for the info, you've inspired me, I remember buying some Nichrome wire back last year, dont like the idea of putting it in the tank tho.
    Would it be feasible to coil the NiCr so I could fit it underneath the fuel filter, See attachment, then insulate it so that all the heat rises up into the body of the fuel filter.
    I do realise that coiling it would increase resistance, therefore increasing the power needed.

    ThanX again for everyones input,
  9. rspuzio

    Active Member

    Jan 19, 2009
    What is the gauge of your wire? From that information, we
    can figure out what length is needed.

    That sounds like a reasonable plan to me.
  10. roblloydwales

    Thread Starter New Member

    Aug 20, 2009
    Have 32 SWG NiCr, says on package 30mm = 1 Ohm.

    Really greatfull for your help.

    By the way, I've been doing a little experimenting with TEC device, interesting stuff, think I'll make a pint cooler lol.
  11. steveb

    Senior Member

    Jul 3, 2008
    Is that a safe thing to do? What if the heater becomes defective and a spark is created? I realize that oil is not ignited as easily as gasoline, but personally I would not be comfortable with this.

    Also, I agree with previous comments. TECs are all wrong for this application for a number of reasons.

    What about inductive heating of the metal casing?
  12. rspuzio

    Active Member

    Jan 19, 2009
    That's quite thin for the application at hand so you'd have to
    twist many strands to make a cable or, equivalently, use many
    strands in parallel to deliver the power needed to heat the oil
    in a reasonable time. As for how many, let's do some more figuring.

    According to the estimate above, we want to have around 60 A
    flowing through the heater. To keep it safe, we don't want to
    have the wire get hot enough to burn anything --- since the oil
    only needs to be heated to 90° C, it would be good if the
    heater would not get too much hotter. As some guide to what is
    safe, we could start with a current/temperature table for NiCr
    wire, such as the one found here:


    To be sure, these numbers won't apply exactly because they're
    for straight wires or coils in air, but they at should at least give
    a rough indication of what to expect. For instance, if we dump
    all 60 A through a gauge 10 wire, the table says 800° C, so
    that would make the wire red hot, which is no good. The lowest
    temperature in the table is 200° C, it says to divide by 2 for
    coiled elements, so let's see what that gives for different gauges.
    Using the table on the bottom of the web page, we can then
    figure out how long a wire would be needed:

    10 AWG --- 8 parallel strands of 25 ft. each
    16 AWG --- 20 strands, 16 ft.
    22 AWG --- 40 strands, 8 ft.
    28 AWG --- 100 strands, 6 ft.
    32 AWG --- 175 strands, 4 ft.
    40 AWG --- 500 strands, 1.5 ft.

    This makes it look like wire is not a practical option for the heater.
    (So much for my earlier thought of a loop of wire.) Maybe instead
    make the heater as a ribbon of sheet metal --- a lot of fine wires
    running in parallel looks like a sheet. For instance, you might
    wrap a sheet of metal around your filter as a heating element.

    At this point, the proposal sounds a lot like Steve's idea with the
    induction heating, only the current would come directly from the
    battery as opposed to having to first make the A.C. needed for
    induction heating out of it. (But chopping the D.C. up with a switch
    might be all that's needed, so definitely worth considering further.)

    Although it's a moot point now that the original poster is not
    interested in doing this, but, out of curiosity, I looked up the
    flash point of vegetable oil. The figure I found was around 300° C
    (as opposed to, say -40° C for gasoline) so it looks like, at the
    sorts of temperatures discussed here, (as opposed to, say a deep
    fryer) there should not be too much to worry about from sparks.