The measured finesse of IMC was 540 +/- 70.
This is consistent with the mirror reflectivity measurements by LMA.

What we did
 1. Decreased the modulation depth by -12.5dB since PDH signal from IMC REFL PD was saturated. This was done by changing the RF oscillator output from 0dBm to -12.5dBm

 2. Did the mode scan with the same way described in klog #613, but with higher data acquisition rate (sampling frequency of 16 kHz).

Results
 1. The time series data we have used is attached (modescantimeseries20160117.png). The blue circles show the timestamps for the TEM00. It was a six-FSR scan.

 2. Got t_FSR = 0.74 +/- 0.02 sec from the fitting (fsrfit20160117.png).

 3. Got t_FWHM = 0.0014 +/- 0.0002 sec from the Lorentzian fitting of the transmitted beam peak. One example plot of the result of the fitting is attached (peakfit201601173.png).

 4. By t_FSR/t_FWHM, we got IMC finesse to be 540 +/- 70.

Discussion
 1. Assuming MCe loss to be very small, and MCi and MCo loss to be equal, finesse of 540 +/- 70 gives MCi/o loss to be 0.59 +/- 0.06 %. LMA reported the mirror reflectivity of MCi/o to be 99.37 %, so this result is consistent with their value. This means that the mirrors didn't get too dirty during the installation work!

 2. Assuming the incident beam power to be 1 W and perfect mode-matching, finesse of 540 gives intra cavity power to be 170 W. MCe transmission of 5ppm gives MCe transmitted power to be 0.8 mW. Half of this power is injected into IMC TRANS DC PD (Thorlabs PDA100A). PDA100A has the gain of 0.75e5 V/A for 40dB setting, and has the responsivity of 0.28 A/W. Thus, 0.8 mW/2 should give 8.4 V output. However, the output measured when locked was 2.8 V. About 70 % of the beam is lost (possibly from the modematching, the viewports, and PD). More investigation needed.

Next
 1. Measure modematching ratio with mode scan (and also the transverse mode spacing to derive the RoC of MCe).
 2. Measure the power of the MCe transmitted beam.