Notes on 1.5um avalanche diode design
From conversations between Alex and Lee:
but like many of their diodes, it is very expensive. It gets sub pW/rtHz noise floor that we can potentially achieve for much cheaper.
6fA/rtHz current noise is crazy. Voltage noise is also very very good
so that with a 100k trans-impedance resistor would be mightly low noise, with a gain 50 APD, then you could get fW level sensitivity. Probably would end up dark-current limited on the diode though
those amps are crazy. 100mA output and nearly rail-rail
those should be the same diode, but one is only 80$ a pop, though a pain to mount. If you can find any kind of test board to solder that to, it would be great to not buy the more expensive one
or buy fewer of them, since we are likely to burn a couple out with the high bias voltage
the datasheet indicates that the 80um version has 350fA/rtHz dark current noise. So the op-amp is overkill a bit. It indicates that it has an APD gain of 20. It has a nice low capacitance of 2pF. So if you use a 100k resistor, you won't get much current noise from the resistor. That give 350e-15 A/rtHz / 20 ~ 20fW/rtHz noise floor. That would be an exceptionally low noise photodiode
the 100k resistor on 2pF gives a photodiode bandwidth of 1 / (100e3 * 2e-12 * np.pi * 2) → 800kHz. That's pretty good! You could possibly use as low as a 10kOhm resistor and still get decent noise if we want >1MHz bandwidth