Connect with us

Calibrating a Thermocouple

Discussion in 'General Electronics Discussion' started by TR34TM3NT, Aug 15, 2012.

  1. TR34TM3NT

    TR34TM3NT

    3
    0
    Aug 15, 2012
    I have a beverage cooler that I'm trying to modify to make the internal temperature colder.
    The internal digital thermostat only allows me to set the temperature down to 39degF, and I'd like to have the temps about 5-10deg cooler.

    I know the compressor is capable of pushing temps below 39degF by wiring in a completely new thermostat, but I'd like to try a simpler/cheaper approach.

    I came up with the idea of adding a Potentiometer to the thermostat's thermocouple, in hopes it would allow me to adjust/calibrate the Ohms coming back to the circuit board... making the thermostat think the internal air temperature is warmer than it actually is, and forcing the compressor to continue cooling past the 39degF mark.

    I made a basic diagram that shows what i was thinking, but that is as far as my elementary knowledge of electronics has taken me. I am unaware of the potentiometer type I would need, or if this would even work.

    Hoping someone can lend some advice or push me in the right direction regarding this project! Thanks!

    [​IMG]
     
  2. john monks

    john monks

    693
    1
    Mar 9, 2012
    I believe you would have better luck with a thermistor.
    You can start out with one designed for refrigerators and and ohmmeter.
    The high end can be done with boiling water and ice water (water with ice cubes in it).
    Then a simple comparator circuit with a relay can be set up to run the freon pump.

    Thermocouples are a real pain to calibrate.
     
    Last edited: Aug 15, 2012
  3. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,191
    2,693
    Jan 21, 2010
    The potentiometer will load the thermocouple so much that the fridge will probably thing the internal temperature is way, way, way below freezing and hence never turn on the compressor.

    (That's assuming it really is a thermocouple)

    I would be looking at the board to see if there wasn't some obvious way to alter the minimum temperature (can you provide a photo?).

    You may also end up butting heads with the fact that the fridge has been designed for a certain temperature range and may not have the capacity to reduce the temperature a lot further (especially with high ambient temperatures).
     
  4. Harald Kapp

    Harald Kapp Moderator Moderator

    9,284
    1,881
    Nov 17, 2011
    "Thermocouples measure the temperature difference between two points, not absolute temperature."

    If you really have a thermocouple (not a PT100, NTC or similar sensor), you can try to locate the cold junction (usually the terminal) in your cooler. If you cool this cold junction (e.g using a fan that blows cold air from the cooler over the terminals) you'd trick the cooler into "thinking" that the temperature difference is less than it actually is. This could trick it into cooling a bit more.

    Where did you get that information? Are you sure the compressor is capable of operating harder than it does? Probably another model of cooler has not only another thermostat but also another compressor. By tricking the existing compressor into working harder, it may blow up.
     
  5. (*steve*)

    (*steve*) ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd Moderator

    25,191
    2,693
    Jan 21, 2010
    The "cold junction", if implemented using a third metal (say, copper) doesn't need to be cold. They (because there's 2 of them) just have to be at the same temperature.
     
  6. Harald Kapp

    Harald Kapp Moderator Moderator

    9,284
    1,881
    Nov 17, 2011
    I think you're not completely correct. The cold junction has two tasks:
    1) maintain the same temperature for both connections at the cold junction.
    2) give a reference temperature.
    The reference temperature is required because the thermocouple's output is a function of the temperature difference between the sensor head and the cold junction.
    You're right insofar as the cold junction doesn't have to be cold, but the temperature has to be known. It depends on where on the controller the "real" ambient temperature is measured.
    If it is measured directly at the terminals, my idea is moot. right.
    If the temperature is measured somewhere on the PCB, assuming thatt the terminals have the same temperature, you could trick the controller by cooling the cold junction. The measured voltage will then be lower as expected, forcing the compressor to do more work.
     
  7. shrtrnd

    shrtrnd

    3,670
    453
    Jan 15, 2010
    You also have to consider what type of thermocouple it is, which I don't think was
    given. Thermocouples operate linearly only in a specific range of their characteristic
    curves, and the type (probably J or K for this range), matters for the circuit.
    I think john monks thermistor is a better pick for this job.
     
  8. TR34TM3NT

    TR34TM3NT

    3
    0
    Aug 15, 2012
    Thanks for the replies!

    A couple beer forums show people using this exact model and just swapping out the thermostat to get it to 35.

    I can't remember what made me assume it was a thermocouple, I think I may have found out from the manufacture's customer service line (but i guess that still isn't a reliable source, lol).

    Here is an image of the board and the wiring diagram:

    [​IMG]

    [​IMG]
     
  9. GreenGiant

    GreenGiant

    830
    6
    Feb 9, 2012
    In the wiring diagram here you can see that it is a thermistor, not a thermocouple
    you would either need to put a resistor in parallel/series with it or replace it with a different one with different ratings
     
  10. Electrobrains

    Electrobrains

    259
    5
    Jan 2, 2012
    In general, I think your simple idea could work! You can offset the temperature sensor by either (as you intend) put a resistor or potentiometer parallel to the sensor, or in series with the sensor, depending on if the thermistor is a ptc or ntc type.

    The "high tech" way of doing that would be to measure the resistance of the thermistor with an ohm-meter and then calculate what parallel/serial resistor to add. I suppose the thermistor value is in the range 100 Ohm to 10k Ohm

    If you want to make it quick and easy, I think you can just test it.
    Some important hints:
    1. If you try the parallel potentiometer approach, then use a high value type (eg. 100k Ohm) in 2-wire connection (use the middle wiper of the poti and one end position, connected directly over the sensor, the other end position can be left open or connected to the wiper).
    2. Start with the highest value and slowly turn it till the compressor starts. Never turn it totally down to zero (as security you could add 100 Ohm in series to protect against short circuit).
    3. If the compressor doesn't start, you could try the serial approach. Put a low-value potentiometer (eg. 100 Ohm or 1k) in series to the sensor (2-wire configuration: wiper and one end side of the poti).
    4. Start with the the lowest value (0 Ohm) and slowly turn it till the compressor starts
    5. Adjust the potentiometer so that the compressor is switched off at the right temperature. Be aware of that it's not good to let the compressor run all the time and probably you will reduce the life expectancy of your cooler by doing this modification (also, all warranty will be lost)

    (everything above written under the condition, that the compressor and other parts of the cooler has the ability to carry the extra load and strain without overheating or other damage)
     

    Attached Files:

    Last edited: Aug 18, 2012
  11. TR34TM3NT

    TR34TM3NT

    3
    0
    Aug 15, 2012
    Yes, apparently it is a NTC thermistor, thank you all!

    I went ahead and tried adding the potentiometer as shown above (just without the extra resistor).

    Got the knob to a point where the thermostat thought it was 10degF higher than the actual temp! However, I noticed it would still turn the compressor off when the actual internal temp got around 38 or 37. I imagine it's doing this due to auto-defrost thermistor (white).

    Having to add another potentiometer to the defrost sensor, while trying to determine a resistance amount without visual feedback from the thermostat might be more trouble than it's worth. :\

    I guess my next step will be replacing the thermostat with a PID controller and SSR.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-