I took a transfer function of RFPD S/N 02 tonight with the Jenne Laser. It was much easier to calibrate this one because I didn't have to worry about what the Cougar was doing (since I didn't put one on this board). I also noticed that for some reason the power leaving each port of the beamsplitter was not equal---there was about 2x the power going to the reference PD as to the DUT---this explains why I was getting less gain than expected by about this factor with the last PD.

Here is a transfer function:

Some notes:

• With the resonant notch (aka "aLIGO style") readout, the easiest way to tune the readout notch using an optical transfer function like this is to actually tune the parasitic parallel resonance that occurs nearby to the frequency predicted by a model such as LISO. Before I tuned RFPD01, I found that for a 33 MHz notch and a 200-pF diode capacitance, this resonance occurred at ~35 MHz. As I learned after doing that, the diode capacitance with a 5V bias is more like 100 pF (which explains why the AC coupling resonance of the first PD did not match with the model). I had "35 MHz" stuck in my head today when I tuned the second PD, which means that the readout notch is actually at a lower frequency than 33 MHz. I adjusted the readout notch capacitor in the model so that the TF matched the one I took and it gives a value of about 31 MHz. I will retune it to the proper value (and do the same for the first PD).
  • Another, perhaps easier, way of making sure the readout notch is at the right frequency is to just look at the voltage above the readout capacitor (instead of at the output, which is an amplified version of the voltage between the capacitor and inductor). Here, the readout notch looks just like the rejection notches and can be tuned directly. I think I'll just do this rather than rely on unknown or sketchily known parameters.
• Besides slight differences in frequency and Q, the measured and predicted transfer functions match up very well, as opposed to the case with the Cougar involved (see link in last bullet). For this reason and more importantly for reasons that will be highlighted in my next post, I propose that we do away with the Cougar altogether for the time being. You can read ahead or just remember that the noise figure of the AP1053 is >3.5 dB at 33 MHz; It's not worth it for this application.
• I took a noise measurement of this PD, but I forgot to turn off the light source before doing so. The noise is a factor of ~1.5-2 worse than predicted, which could either be a non-ideality in the circuit or else it comes from noise on the light amplified by the PD. New measurement after the retuning.