Circuit #1: enter image description here

Circuit #2 (R is small, and Vout is amplified using the same op amp in the above circuit, which is not included in the image below): enter image description here

My voltmeter (an ADC) has a low input impedance (10k Ohms).

  • \$\begingroup\$ Both are likely to be quite noisy. The second one your gain is limited by your meter resistance. \$\endgroup\$ – The Photon Jun 18 '15 at 16:38

The first one (transimpedance amplifier) will have a much better high frequency response. If you want to reduce the high frequency response you can add a parallel capacitor across the feedback resistor.

The second will have frequency response limited by the photodiode capacitance.

Here is a TI application note on calculating noise in a transimpedance amplifier.

In the second case, you'll have an amplifier with voltage and current noise spectra and Johnson-Nyquist noise in the resistor.

There is also shot noise in the photodiode.

  • \$\begingroup\$ I've used transimpedance amps for very precise measurements of laser absorption spectra. Depending on what he means by "light sensing". Transimpedance amplifiers may need either LP or notch filters at flourescent switching frequency to kill light polution noise. Both of them will need 60hz notch filters for line frequency flicker in incandescents and tv screens. If you know the noise you can reduce it with a targeted filters easily. Another strategy is to use a second photodiode and an instrumentation amp to get an error signal depending on measurement set up. \$\endgroup\$ – crasic Jun 18 '15 at 17:33
  • \$\begingroup\$ @crasic Good points all. The OP's question is too vague to be worthy of covering all possible situations. \$\endgroup\$ – Spehro Pefhany Jun 18 '15 at 17:36
  • \$\begingroup\$ However, I am using this for dc light sensing, so frequency response is irrelevant. \$\endgroup\$ – Dave Jun 19 '15 at 0:00
  • 1
    \$\begingroup\$ @Dave if this is under normal room conditions you will want a lp filter to cut out serious flicker noise from ambient lighting. You can also average in software. Transimpedance amps make it easier to hook directly into adcs and provide some signal buffering if you are noise sensitive. \$\endgroup\$ – crasic Jun 19 '15 at 6:04

Circuit 1 with modification: You don't need to bias the photodiode with a -5V. Instead, connect it to ground. This will reduce dark current. Biasing the photodiode is only beneficial when you want higher bandwidths as it reduces the capacitance of the photodiode.


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.