My apologies if this is too localized, but I am really in a bind.

I am attempting to amplify the voltage difference (so to be read on an arduino) between two thermocouple's copper wires, as pictured here:

thermocouples as sap flow http://www.cens.ucla.edu/pub/EnvironmentalSolutions/SFimages/install_schematic.png

The temperature difference between these two probes is very small, no more than 20 degrees C

I have had little success using both an INA122 and LT1920 on a breadboard to amplify this signal, following the thermocouple schematics in the datasheets.

Is there a simpler amplifier to use for this signal?


For those who might have my problem in the future:

The easiest method of reading these small voltages on an Arduino for someone who might not be adept with analog circuitry is to use a separate ADC.

A great example of interfacing one with an arduino can be found here: http://forums.adafruit.com/viewtopic.php?f=31&t=12269&p=58824

  • \$\begingroup\$ I'm not familiar with those amplifiers, but there is no way I would attempt to build a high gain, low drift amplifier on a breadboard (assuming you mean one of those things with holes every 0.1"). It's an exercise in futility. Get a piece of bare PC board and deadbug the circuit. \$\endgroup\$
    – lyndon
    Nov 7 '13 at 0:50

As long as you need the difference, you don't need a special thermocouple amplifier with zero point compensation.

The sensitivity of the T type thermocouple (copper-constantan) is 43uV/K, so for 20K difference you will have 860uV or 0.86mV range. If you have an ADC with 5V range it means you need a gain of 5814.

This value is pretty high and you will need a very precise OpAmp in order to get acceptable temperature drift.

But here comes the question of what accuracy you need? If you assume 0.1K accuracy and for example 20K difference in the device working temperature, you will need an OpAmp drift of 4uV/20K = 200nV/K or less, which means Auto-zero opamp.

Something like AD8676 (200nV/K) or AD8538 (30nV/K).

Don't forget about the input offset compensation, AC noise filtering and the quality of the passive elements (resistors) that can ruin the accuracy of such small signal amplifiers.

  • \$\begingroup\$ The AD8628 in 2nV/K drift: analog.com/static/imported-files/data_sheets/… \$\endgroup\$
    – Andy aka
    Nov 6 '13 at 21:13
  • \$\begingroup\$ @Andyaka - yes, there are many of them. Lower drift means higher accuracy here. \$\endgroup\$
    – johnfound
    Nov 6 '13 at 21:16
  • \$\begingroup\$ On further consideration, I'm going to go with an ADC converter, as the processes are more familiar to me. \$\endgroup\$
    – camdenl
    Nov 6 '13 at 22:18
  • \$\begingroup\$ @camdenl I didn't understand what does ADC here? You still need to amplify the thermocouple signal, before passing it to the ADC. \$\endgroup\$
    – johnfound
    Nov 6 '13 at 22:23
  • \$\begingroup\$ I'm sorry I meant use an external ADC, 18 bit or higher. It would have enough resolution to read a low voltage accurately. \$\endgroup\$
    – camdenl
    Nov 7 '13 at 1:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.