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.

AC Coupled Photodiode

I have been looking at this circuit posted by Bruce Trump and am trying to use it as a method reject ambient light at DC, 60Hz, and 120 Hz.  My application is very low bandwidth, I am trying to pass a signal at 625 Hz with a bandwidth of 50 Hz.. 

There are a few things I am trying to clarify, first is the frequency response, it first appears that the cutoff frequency would be 16Hz. However according to another post

it is actually 6.5 kHz because the transfer function of the transistor. Does anyone have an equation or can explain this better?  Because I would like to adjust it to 350Hz.

This circuit also seems to work well in simulation with small DC currents 0 to 20uA but becomes unstable for higher currents. 

I assume VE of T1 needs to stay below VB and this can be adjusted by lowering R4 to accommodate higher ambient current.   I imagine there is a limit to how low R4 can reasonably go, any ideas?

In my application, I am not interested in bandwidth but highly interested in noise. Any ideas on how this servo loop and transistor affect the noise model or have recomendations to keep the noise low?

Thanks for the help,


  • Hi Chris,

    Be aware that you are not going to get much attenuation of 60 and 120 Hz relative to the desired response at 625 Hz. The single pole response of this feedback approach, at best, will only attenuate directly according to the ratio of the interfering frequency to the signal frequency. That's not much attenuation in your case. This approach is most successful when the signal frequency is much higher.

    The low frequency cutoff of the transfer function actually varies with the DC component of the input signal it is rejecting. It's much lower with low interfering current. This is due to the varying transconductance of T1 as its current varies. As current increases enough to produce approximately 200mV (degeneration voltage) across R4, the cutoff frequency will flatten out to a nearly constant value. This variation in cutoff is not significant in many applications where the desired signal frequency is much higher.

    It might be possible to control this cutoff variation by connecting the feedback of C1 to R4 but I expect this could lead to instability of U2. I have not investigated this possibility. 

    You can scale the cutoff relative to present values by scaling C1. Cutoff is inversely proportional to C1. Virtually all the other circuit values also affect the cutoff frequency but adjustment is probably best made with C1.

    Note that I am now retired and contribute on the forum only sporadically. I try to look out for questions related to my old postings when I was employed at TI. All my old postings now show me as a forum member rather than an employee.

    Regards, Bruce.

  • Hi Bruce,

    Could you use a passive low pass filter in as an input to the negative side of U2, instead of using the integrator and rely on the op amp to handle the variable transconductance of T1? Would this result in a more consistent frequency response for varying ambient currents?

  • Chris,

    Additional filtering in series with R2 will cause instability or oscillations in the servo feedback loop. If you must have such a low desired signal frequency, the best approach is probably an additional high-pass filter in the signal path after U1 (not in the servo feedback loop).

    I'm not sure I understand your other possible change. I tried a quick simulation connecting C1 to the emitter of T1. This may be a viable approach to controlling the cutoff frequency.

    Since little 120 Hz attenuation can be achieved relative to 625 Hz with the servo loop, it may be better to just make this cutoff at a very low frequency to handle very slow changes in background photo level. Then use a sharp cutoff high-pass filter at 500 Hz after U2 to deal with the 60/120 Hz issue. Then you don't need to be so concerned with variation in the servo loop bandwidth.

    Have you done all you can to optimize the S/N of the optics? If your desired signal is from a laser or LED, the proper optical filter can make a dramatic reduction in optical noise from ambient light.

    Regards,  Bruce

  • Ok, I will stick with the integrator feedback loop to maintain stability.  I am not too concerned with eliminating 60 Hz and 120 Hz since the amplitudes of that light will be much lower than the DC component (this will be used mostly outdoors).   Like you said those power line harmonics can be attenuated with a low pass filter after this TIA stage.   My main concern is just to keep the TIA out of saturation in the presence of ambient light.  I will have a pretty large gain (500 kOhm) and a low voltage rail (1.8V)

    Would there be any benefit of using a FET instead of a BJT, for T1 would the frequency response be more predictable?

  • Chris,

    I think there is little fundamental improvement to be gained with a FET. There may some finessing to come from the selection of a specific transistor due to the junction capacitance. This would be more significant with high bandwidth applications and could only be explored by including detailed modeling of your photodiode.