T. Yamamoto, Ushiba, Akutsu; Follow-up of 18000.
Abstract
We, by ourselves, roughly estimated the IMC absolute length with sweeping the thermal actuator of the NPRO source, and also estimated the IMC finesse. To do so, while keep the thermal actuator sweeping, we independently picked up data with ndscopes and independently did analysis.
Absolute length
Here I would describe the overview of our method, and I hope each of us will report independently. With awggui, we excited K1:IMC-SERVO_NPRO_TEMP_BIAS_EXC in 0.01 Hz triagular, amplitude = 0.001, Gain =1 (and Ramp 5 sec), and looked at the behaviors of K1:IMC-CAV_TRANS_OUT. The fundermental manner is the same as those in 17989 or maybe in 613 or 618.
Then, in my data and from my cursors manipulation in the ndscope, the time separation between carrier-carrier: 7.402 sec, 7.683 sec, 7.429 sec, 7.663 sec, respectively, so the mean is 7.5443 sec. And, the time separation between carrier and the corresponding upper RF sideband (13.78 MHz away from the carrier) is 18.400 sec, 18.724 sec, 18.539 sec, so the mean is 18.5543 sec. Then, 13.78MHz/18.5543 sec * 7.5543 sec = 5.603 MHz, which is my estimation of FSR. So the estimated absolute length is c/5.603 MHz = 53.50 m, roughly speaking. Seems consistent with the designed value, 0.5+26.4*2 = 53.3 m, according to the wiki. But note that this error is unrealistic --- 20 cm in the absolute length (10cm in one way) would be unrealistically large.
Finesse
Also with the cursors in the ndscope, we estimated the FWHM of the peaks. Unlike "clean" peak shapes when sweeping MCe, in our sweeping of the thermal actuator, the transmission peaks seemed relatively less cleaner (see the attached figure), and there was uncertainty of the FWHM measurements! Actually in the figure the diffrence in the time cursors show about 5 msec, while you might think this might be about 10 msec! maybe due to the double-peak-like structure. Anyway, most of the peaks looked like this, and let's assume ~10 msec as the FWHM. Then, according to the result above, FSR would correspond to 7.5443 sec, so 7.5/10e-3 ~ 750, which seems higher than the design... why?? Strange... If we adopt ~ 5msec case, the finesse would be double...
Next
Well, leave this, and would like to try the method in 616.
