Attached a quick plot showing dB of squeeze we get at various CLF power. 1X CLF power being 0.05 mW sent into the coupler, 1uW transmitted. We were able to gain a little more just by decreasing the power (at least down to 100Hz). This might be an indication that there is something to be gain just by investigating CLF noise alone. More noise budget to follow. 

The second plot shows DARM at these power.The mysterious peaks that get better with higher power are between 200Hz-500Hz (well, not so mysterious in a sense that we know they correlate with acoustic from SQZT6, but how does it get into DARM?). I kept the UGF of CLF and LO the same at all these measurement. I also checked the alignment.

Another useful thing I found that at the same CLF power, the optimal squeezing phase stay the same between the two injections (on Feb 19 and Feb 21). 30.4deg on spare phase delay slide bar, 69.21 everywhere else, CLF sign positive is the optimal place for 0.05mW CLF launch. Note that the green power is kept stabilized at 20mW.