Hello, I've got a quadrant photodiode operating in photovoltaic mode. I am trying to determine what the minimum detectable power for the device at frequencies higher than DC. At this point, using a reference diode and power meter, I have studied the relationship between optical power input and voltage output. It matches the scale factor suggested by responsivity and transimpedance gain of the diode. There is an offset that seems to correspond to the dark voltage -- that is, when I measure the voltage in dark conditions, it is about the same as my offset. However, I am really interested in the minimum detectable power for non-zero frequencies. My question is the following: If I studied the relationship between optical power input and voltage output at higher frequencies (by modulating a beam at some frequency and changing modulation amplitude), should I expect a different offset? Say, something equivalent to the noise floor away from the zero-frequency component of the frequency spectra? I am asking this because I've actually done this, but I my fit results in an offset close to the dark voltage value. I've attached an image illustrating what I'm talking about. What is the behavior of dark voltage at higher frequencies? Thanks!