Discussion in 'General Electronics Chat' started by subhash.vnr, Nov 29, 2010.

  1. subhash.vnr

    Thread Starter New Member

    Nov 29, 2010
    can a thermocouple be fabricated manmade? i.e at my home
  2. Wendy


    Mar 24, 2008
    I used to make them using a version of an arc welder designed for it. If you have a welder you might be able to make one, you need to fuse the two wires (both different metals) into a molten ball. When it cools you have a thermocouple.
  3. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    We use them in lots of equipment we manufacture, so we manufacture lots of them ourselves.

    You need some thermocouple wire. (We use J and K type exclusively, but not for any specific reason I can gather). Then you need a way to weld it. We use a spot welder. After welding and cleaning we try to pick the weld apart with an ice pick. 10-20% of them fail this test. We clip the welded tip off and 'do it again' :)

    any two different metals will form a thermocouple, but if you want to use the prolific millivolt tables vs temp you need to use the same metals those charts reference, which means using the thermocouple wire.

    a capacitor discharge welder might do it if the wires are small enough gauge.
  4. JDT

    Well-Known Member

    Feb 12, 2009
    Actually, just twisting the wires together will work - but not very reliably!
  5. KMoffett

    AAC Fanatic!

    Dec 19, 2007
    Agree. You can also use a propane or MAPP torch to melt the twisted wires together. Spot welding or TIG welding quickly melt the wires, without melting the wires' insulation.

  6. someonesdad

    Senior Member

    Jul 7, 2009
    Yes, you can make them at home with simple equipment. The simplest way is with two pairs of pliers. Clamp the two bare thermocouple wires together, then use the other pliers to twist the wires together. Some judicious smashing of the twisted joint with a hammer also doesn't hurt. The idea is to get a good mechanical joint that holds the two metals in contact.

    The usual approach is to weld the two ends together because this provides a strong mechanical joint. This just takes a moment in the flame of an oxy-acetylene torch. In industry, spot welding is often used.

    However, the above is not necessary. You can use a small block of metal (for example, copper or brass) and just screw the two thermocouple wires close together. If the block is isothermal, then you'll still be able to get an acceptable thermocouple reading. The effects of the intermediate metals cancel out.

    In practice, using thermocouples is fairly easy as long as you have a high-impedance voltmeter to measure the thermal EMF. You also need to realize that thermocouple EMFs are typically small voltages; here's part of a table for a type K thermocouple:

    Code ( (Unknown Language)):
    2.  ITS-90 Table for type K thermocouple
    3.  °C      0      1      2      3      4      5      6      7      8      9     10  
    4.                                Thermoelectric Voltage in mV
    6.    0  0.000  0.039  0.079  0.119  0.158  0.198  0.238  0.277  0.317  0.357  0.397
    7.   10  0.397  0.437  0.477  0.517  0.557  0.597  0.637  0.677  0.718  0.758  0.798
    8.   20  0.798  0.838  0.879  0.919  0.960  1.000  1.041  1.081  1.122  1.163  1.203
    9.   30  1.203  1.244  1.285  1.326  1.366  1.407  1.448  1.489  1.530  1.571  1.612
    10.   40  1.612  1.653  1.694  1.735  1.776  1.817  1.858  1.899  1.941  1.982  2.023
    Note that the difference between e.g. 40 and 41 °C is 40 μV. Thus, to resolve to 1 °C around temperatures of 40 °C, you'll need to be able to measure voltages of 1.5 mV with (0.04/1.5)100 = 2.5%. If you're building your own equipment, this means a carefully-made DC amplifier.

    A thorny part of using thermocouples is getting the reference voltage. This is done in the lab by inserting another thermocouple junction (of the same type) into an ice water bath and is called, not surprisingly, the ice-point reference. This is usually inconvenient, however, so there are various electronic ice point references that can be purchased or made.

    Note for crude measurements, you don't really need an ice point reference. I once was in a rush on a hot summer day to haul our trailer to another state to meet my wife and daughter. The vehicle I was using had been having occasional overheating problems and the temperature meter had recently quit working. Not a good thing. My "emergency" solution was to use a hose clamp to clamp a type K thermocouple junction to the cast aluminum cover of the thermostat and run the wire through the firewall to the driver's seat. I then just measured the thermal EMF on my Fluke digital multimeter and used a thermometer to make a correction for room temperature. I had made up a hand-written table to convert EMF to temperature. It was crude, but it let me make the drive safely without overheating.

    If you measure the thermocouple voltages with a computer, another method of correcting for the ice point is to use e.g. a thermistor to measure the temperature of an isothermal reference block and use this as the reference. This works well and has been used for decades in general-purpose scanning devices.

    I've attached some python routines I wrote to convert back and forth between thermocouple EMFs and temperature using the NIST polynomials. It wouldn't be too hard to convert this stuff to another programming language.