I moved the Hartmut Green PD to the Jenne laser bench to try to determine if the response at RF was reasonable or somehow very much smaller than it should be. It was set up as shown in the attached diagram. The first pass at this was by comparing the ratio of the RF photocurrent of the green PD to the RF photocurrent of the New Focus 1611 InGaAs PD. That ratio (at a sufficiently low frequency) should be the same as the ratio the DC photocurrents of the two PDs.
Using the network analyzer I measured the ratio of the voltages of the two RF signals (and then scaled each of these by the respective transimpedances of the PDs: 700 Ohms for the 1611 and 240 Ohms for the Harmut PD). The resulting ratio is shown in the attached plot.
I measured the DC voltages from each PD and scaled those by the transimpedances to get the photocurrent (10 kOhm for the 1611 and 80 Ohm effective for the Harmut PD). The ratio of the DC photocurrents was 0.37. This is roughly 3x the ratio of the RF photocurrents at 500kHz (=0.14). This discrepancy is uncomfortably large.
The full set of measurements is given in the table below:
Measurement 
Value 
DC voltage from Hartmut PD 
6.5mV (checked by turning laser on and off and measuring the difference) 
DC voltage from 1611 InGaAs PD 
2.20V 
Transimpedance of Harmut PD at DC 
80 Ohm (effective) 
Transimpedance of Harmut PD at RF 
240 Ohm 
Transimpedance of 1611 InGaAs at DC 
10 KOhm 
Transimpedance of 1611 InGaAs at RF 
700 Ohm 
Incident Power on Hartmut PD (100% on PD area) 
0.28mW (measured by Ophir power meter) 
Incident Power on 1611 InGaAs (<100% on PD area) 
0.64mW 
Responsivity of Silicon PD at 1064nm 
0.02 A/W (estimate) 
Responsivity of 1611 New Focus PD at 1064nm 
~0.8 A/W 


There is one other troubling point: using the estimate of responsivity on the Harmut PD * incident power * transimpedance at DC = (0.02A/W) * (0.28mW) * (80 V/A) = 0.45 mV.
But the measured DC voltage is 6.5mV = inconsistent. 