1
\$\begingroup\$

I am considering using a digital potmeter with, e.g., I2C interface as the resistor in the feedback loop of my transimpedance amplifier. I am concerned about this. Will this add any additional noise compared to a normal resistor or somehow limit the transimpedance amplifier performance in any way?

EDIT: Conversion range needed: 10nA - 10mA

|improve this question
\$\endgroup\$
  • 2
    \$\begingroup\$ My main worry would be about bandwidth. The digital pots I know about are all really slow. And the TIA applications I know are all generally trying to maximize bandwidth. \$\endgroup\$ – The Photon Oct 20 '13 at 2:24
  • \$\begingroup\$ @ThePhoton: You are right. I was looking at some of the graphs provided in the Analog Devices digital potentiometers data sheets and the bandwidth performance is poor (at least for high resistor values). So this does not seem like a viable option. \$\endgroup\$ – iQt Oct 20 '13 at 11:24
2
\$\begingroup\$

Digital pot could have more noise compared to a discrete resistor.

The amplifier can be split into two (2) stages.
1st stage is a fixed gain transimpedance amplifier. (One could call it a pre-amp.)
2nd stage is a variable gain amplifier with a digital pot. The 2nd stage shall have a strong input signal from the 1st stage, so the noise from the digital pot would have only a small impact.

There is also a concern that a digital communication signals from which the digital pot receives its setting (e.g. I2C, SPI, etc) can crosstalk into the analog path of the pot. I've seen that happen with D/A converters. Check this with an oscilloscope: trigger the scope from the digital lines, observe the analog signal. If you see this problem, it can be addressed by creating a quiet period without digital communication, during which the measurement would be made.

|improve this answer
\$\endgroup\$
  • \$\begingroup\$ You also need to be careful about adding too much capacitance onto dpots in more complicated filters. Be sure to read the datasheet \$\endgroup\$ – Napthali Oct 20 '13 at 3:38
  • \$\begingroup\$ @NickAlexeev: The reason I do not want a fixed resistor is that I must be able to convert 10 nA to 10 mA. So I was thinking about having a digital potmeter in both the transimpedance feedback as well as in one or two cascaded gain stages. Any suggestions on how to solve this with fixed resistor in transimpedance amplifier feedback? \$\endgroup\$ – iQt Oct 20 '13 at 10:24

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.