Naoki, Vicky, Dhruva
With the IFO being down for the last few days, we decided to recable the squeezer to lock the filter cavity on the homodyne and take some characterization measurements.
We took a few ADF sweeps of the FC, at nominal detuning and at a high detuning. Based on fitting m_p inferred from the high detuning antisqueezing sweeps, I was able to infer the following parameters -
Round trip loss : 24ppm
Mode-matching : 97%
I used the vendor quote for FC1's transmissisivity (909 ppm).
In order to see the sensitivity of the measurement to mode-matching and loss estimates, I forced the mode-matching to 100% and ended up getting 54ppm as the best-fit value for the round-trip loss. The difference is very small but it looks like the lower mode-matching and round trip losses are better fits for the data. If I were to play with the input mirror transmissivity, it would be possible to get a better fit for a 100% mode-matched filter cavity.
With the obtained FC parameters, I tried to fit the nominal detuning sweeps to figure out what the actual detuning of the FC is during operation (RLF@104.645kHz). The data seems to align well (but not perfectly) with an estimated detuning of around -36 degrees (+/- 1-2 degrees also agrees to some extent).
We took some more ringdown measurements of the FC using the IR Refl and Trans PDs, see 66403. The finesse estimates from Trans seem to be too high (or too low) to be reasonable but the data makes it look like the quality of the fits aren't great. The Refl looks a bit more consistent but it looks like some of the fits might be over-estimating the storage time. The lower ones are around 6400-6600. If we assume an input mirror transmissivity of 909ppm, this might put the round trip loss closer to 20-30 ppm which agrees with the ADF.
Attached FC mode scan is consistent with ~2% SQZ-FC mismatch and misalignment. Scans FCGS VCO freq, while FC is IR frequency-locked w/feedback to sqz laser vco.
Power not in TEM00 ~ (4.5+2.1+2.5)/(2.5+4.5+2.1+444) ~ 2%. Mode marked by cursors as 2nd-order HOM actually is double-peaked hom (heights of 4.5, 2.1). This 2% SQZ-FC mismatch also conservatively counts the little bits of misalignment modes seen in the scan. We checked ZM2 PSAMS optimization 2 ways: 1) while FC-ASC was running to optimize the WFS_A_Q dip (opo locked_clf_dual, fc locked w/sqz laser vco), and 2) while sitting on this higher 2nd order HOM peak to minimize it (opo dither locked on seed, fc locked w/sqz laser vco). Overall ZM2 PSAMS @ 90V still looked best, leaving it here.
HD spectrum with ADF inferred parameters