Set current to 2.3 amps. This corresponds to 3.96 W out of the amplifier. The beam goes through a PBS, 50:50, 90:10, and PBS. The waveplates before the PBS can be optimized to give <1% loss in the S pol. Assuming minimal loss at the 50:50 and 90:10 this gives 1.75 W into the fiber. With 85% efficiency through the fiber my best guess is we have ~1.48W going into the SHG. The readout on the power meter shows 38 mW. This can be adjusted by the ratios in responsivities of the power meter at 775 and 830 so 38mW * 207.1/146.77 = 54 mW of 775 nm light. If we believe the covesion data sheet it predicts ~350 mW for similiar inputs. Could our SHG be that much more inefficient? Based on the power we ARE getting out of the SHG, according to the data sheet, you would think were feeding in approximately 500 mW, which I don't think is the case. Don't want to stick the power meter in there though.
2.3 A on the amplifier was the previously calculated acceptable operating point to not damage the faraday isolator at the amplifier output. I'm assuming 50 mW of 775 nm light is not sufficient for all of gquest, although I don't have a good sense for this number.
Lee suggested to check we are at the correct temperature. According to this test the peak temp is 50.423 C, whereas the manual suggests 50.461. There is a slight uncertainty since we don't have a way to perfectly sync the time scales of the moku and temperature controller logger. Even still though the efficiency as a function of temperature is fairly flat near the peak. I don't think this explains the big discepancy in efficiency of the SHG.