From March 21-26, I conducted some measurements of the response non-linearity of some mock-up DC readout photodetectors. The detectors are simple:
Vbias ---
|
PD
|-------- output
resistor
|
---
-
This is a description of the final measurement.
The laser current modulation input was given a 47Hz sine wave at 20mV. A constant small fraction of the beam was shown onto the reference detector, and a beam that was varied in DC power level was incident on the test detector. Spectra were taken from both detectors at the same time, 0.25Hz bandwidth, over 100 averages.
At each incident power level on the test detector, the Vpk in all multiples of the modulation frequency were measured (ie. V[i*w]). The difference between the 2f/1f ratio in the test and reference was then calculated, ie:
V_test[2*w]/V_test[1*w] - V_ref[2*w]/V_ref[1*w]
This is the solid black line in the plot ("t21-r21_v_power.png").
The response of a simulated non-linear detector was also calculated based on the Vpk measured at each harmonic in the reference detector, assuming that the reference detector had a purely linear response, ie:
V_nl[beta,2*w]/V_nl[beta,1*w] - V_l[2*w]/V_l[1*w]
these are the dashed colored lines in the plot ("t21-r21_v_power.png").
The result of the measurement seems to indicate that the non-linearity in the test detector is less than beta=-1.
The setup that was on the big optics table south of the laser, adjacent to the mode cleaner, is no longer needed. |