• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 50 subscribers
  • Views 1221 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

INA333: Thermistor Signal Conditioning

tedex45
Prodigy 190 points
Part Number: INA333
Other Parts Discussed in Thread: INA828, INA826, INA330

Hi all, 

I have a general question regarding the signal conditioning of a 10kOhm Bead Thermistor. What is the recommended option for best accuracy to interface to a standard ADC within a micro controller. Would it involve a wheatstone bridge and then an instrumentation amplifier? 

Thanks in advance!

  • 0
    TI__Mastermind 23050 points
    Hi Tedex45,

    While there are certainly other ways to measure a thermistor's resistance, placing it in a bridge configuration in front of an instrumentation amplifier is the most common and provided you balance your bridge properly and use precision resistors with low temperature drift, then you can get very good accuracy out of this solution. Interfacing with the ADC can be a challenge in itself depending on what the ADC's sample capacitance, resolution, and acquisition time are. Most INA's will struggle to drive an ADC standalone and typically require a higher bandwidth buffer on the output, but for a microcontroller ADC you may be able to get away with it. I would use something with a little higher bandwidth though, like the INA826 or INA828 just to be on the safe side.

    Alternatively, we also have the INA330, which is specifically designed for 10kOhm thermistor applications and only requires a single precision resistor. It also includes an integrated buffer for the output, though it was not designed with driving an ADC in mind so I'm uncertain how it would perform in this regard; you may still require a higher bandwidth output buffer.
  • 0 in reply to Zak Kaye
    Prodigy 190 points

    Thank you for the quick reply. One question, what is the reason for the need for a higher bandwidth buffer? Does this have to do with the ADC itself?

  • 0 in reply to tedex45
    TI__Mastermind 23050 points
    In a way yes. It's related to the load transient response time of the op-amp, its output impedance and the characteristics of the converter. The sampling capacitor on the ADC switches in and out to take a sample and when it switches in the op-amp must source a certain amount of charge to bring this capacitor up to the output voltage of the op-amp. This initial charge dump causes the output voltage of the op-amp to droop a little bit, and you want it to recover before the end of the acquisition time so that the voltage reading is accurate. The output impedance of the op-amp and the capacitance of the ADC sample and hold capacitor effectively form a time constant. Greater bandwidth translates to more loop gain, which lowers the op-amp's output impedance. A lower output impedance means a lower time constant, or faster charging of the sample and hold capacitor. A carefully selected RC filter between the op-amp and ADC can significantly reduce the bandwidth you need to achieve the desired settling performance.

    The ADCs in microcontrollers tend to have very low sample and hold capacitances (typically around 5pF), whereas standalone higher resolution, higher SNR converters tend to have around 55pF. That's why I said you may be able to get away with a lower bandwidth device. If you are interested in learning more about this, we have a TI precision labs series on how to drive SAR converters that you can find here: training.ti.com/ti-precision-labs-adcs
  • 0 in reply to Zak Kaye
    Prodigy 190 points

    Thank you very much! Appreciate the support. 

  • 0 in reply to Zak Kaye
    Prodigy 190 points
    When using a wheatstone bridge in conjunction with the INA, the output will still be non-linear with respect to temperature. I guess, the calibration would then be done in software to account for that or is it usually linearized in hardware itself?
  • 0 in reply to tedex45
    TI__Mastermind 23050 points
    Hi Tedex45,

    You can perform the linearization with either hardware or software. I believe most people tend to do the calibration in software because it allows you to avoid adding components to your front end. While I have never performed a temperature calibration in software, it is my understanding that this can be a difficult function to implement, so if your experience is more on the side of analog then you may find a hardware implementation more straightforward.
  • 0 in reply to Zak Kaye
    TI__Genius 11135 points
    Hi Tedex45

    We haven't heard back from you so we assume this resolved your issue. If not, post a reply below, or create a new thread if this one is timed-out.

    Thanks
    Dennis