I understand that most people use a bridge circuit for RTD and possibly thermistor temperature sensing. I'm trying not to do that for several reasons.

I'd like to use an opamp in the classic constant current configuration, with the temperature sensing resistor as the load resistor. I sense the voltage change with an INA.

This seems to work so far, on the bench. Has anyone else had experiance with this? Are there any gotcha's I should be aware of when I put it in the field?

  • \$\begingroup\$ The devil is in the details. Please post your circuit. What thermistor are you using? Did you calculate the power (and self-heating) at maximum resistance? \$\endgroup\$ – markrages Feb 25 '11 at 21:11
  • \$\begingroup\$ The current I'm running is less than 10uA. I don't care exactly what the value is, so long as I can measure what it is, which I can. What I'm not confident about is how stable that will be, as BarsMonster mentioned, over temperature and time. \$\endgroup\$ – user3164 Feb 26 '11 at 1:45
  • \$\begingroup\$ A bridge is good at canceling resistor tempco, where a current source is effectively comparing the varing resistor to the one in the current source. But with a good thin-film resistor that may not be a problem. Details are important. Please post your circuit. \$\endgroup\$ – markrages Feb 26 '11 at 18:44

As you usually need just few mA, the only concern is temperature stability of your opamp & passive components around it. Bridge circuit & precision refV source of ADC usually handle that, but in your case it's much much harder.

  • \$\begingroup\$ Am I worried about ambient temp (which I can measure and use a look up table to correct for) or die temp which would be much harder. I'm running this very low current, so I don't expect power dissipation by the opamp itself to be a driving factor in heat buildup. \$\endgroup\$ – user3164 Feb 26 '11 at 1:46
  • \$\begingroup\$ Well, if you can digitally correct it you must be fine. Selfheating on such currents shouldn't cause much problems (as long as you are not shooting for 0.001C accuracy :-) ) \$\endgroup\$ – BarsMonster Feb 26 '11 at 23:24

A diode can be used as a temperature sensor to get a bigger signal. Most resistors not meant for temperature sensing are designed to not to change resistance with temperature.


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.