0

I have 100 IR photodiodes that I would like to multiplex to one (or the fewest achievable) transimpedance amplifier(s), both for space and price constraints.

I am aware of the capacitance and noise issues due to the distance between the photodiodes and the TIA, explained in this answer.

I am looking for a way to increase the signal to noise ratio. As I am in control of the IR sources, I am wondering if modulating the source as explained here can help filtering some of the noise after the TIA.

However as I am already multiplexing the photodiodes at 100Khz (each photodiode should be read every millisecond), I am not sure of which modulation frequency could be used to help reduce noise, or if it would be achievable at all given that the TIA would need to handle the choosen frequency (its Gain BandWidth Product, and slew rate should be rated accordingly).

Do you think that modulation can be benefical in this case, at which minimal frequency ?

littlegreengeek
  • 73
  • 6
  • 1
    How much photocurrent are you getting from the PD? I don't think this is a good system to multiplex before the TIA. Especially if you want to increase the SNR. – Kevin White Mar 30 '21 at 18:12
  • I know it is far from ideal, I am trying to make the most out of the limited space/budget I have. Some multiplexers have very low biased current though. – littlegreengeek Mar 30 '21 at 19:00
  • 1
    The capacitance of a multiplexing tree will kill your SNR though. Even worse if you are already looking to improve your SNR. I'm afraid the best solution would probably be to try and keep the TIA as simple/cheap as possible and mux after the TIA. You didn't reply to my question about how much photocurrent you have. That is a critical parameter. – Kevin White Mar 30 '21 at 19:11
  • Suggest testing your setup to get specs on the reverse light current. Found a reference that states same thing about mux usage depending on photocurrent and mux effecting bandwidth and feedback capacitor [link](https://e2e.ti.com/support/amplifiers/f/amplifiers-forum/550698/multiplex-multiple-photodiodes-to-one-transimpedance-amp) I would add to use guard traces carefully to avoid leakage paths [link](https://www.analog.com/en/technical-articles/optimizing-precision-photodiode-sensor-circuit-design.html) – Ernesto Mar 30 '21 at 19:33
  • Thanks for your suggestions and advices. To add clarity, I am not currently experiencing bad SNR (for now), but my current prototype has a single multiplexer and does not operate in real world conditions. I can't answer how much photocurrent I will have, as I will need to measure in real world conditions. As it is critical I am going to focus on that for now. I thought I could mitigate multiplexer capacitance using an increased feedback capacitor, as explained [here](https://e2e.ti.com/support/amplifiers/f/amplifiers-forum/550698/multiplex-multiple-photodiodes-to-one-transimpedance-amp) ? – littlegreengeek Mar 30 '21 at 22:35
  • What information are you trying to extract from your photodiodes? A simple light or dark measurement? Or multiple levels? Or a precise radiometric measurement? How closely spaced are the photodiodes? – Matt Mar 30 '21 at 22:48
  • The photocurrent increases/decreases almost lineary as a function of time (during a 0.5 second to 2 seconds) interval. I need to measure the duration between 2 peaks. Linear interpolation works nicely for now. – littlegreengeek Mar 30 '21 at 23:04

0 Answers0