I expect an entry will be made about the details of the work today to put the B4 QD2 6Mhz signal to zero. It happened today between 12 UTC and 15 UTC, and was kept for the following lock.

Figure 1 shows a spectrogram of darm between 200Hz and 1kHz. During the change in alignment working point, the spectrogram became much more blue and stayed that way. The broadband noise level in DARM has been reduced.

Figure 2 shows the coupling of the laser frequency lines (SSFS at 3345Hz, FMOD at 1111Hz and IM at 33Hz), all three of them improved by about a factor 2, and very low frequency fluctuations (time scale of hours) stopped, with faster fluctuations remaining around zero.

Figure 3 shows that SSFS coupling fluctuation are very coherent with the RF OMC alignment vertical drift control signal that is currently not in loop. The coherence is over 90% up to 200mHz. The double demodulated OMC AA signal is also in good agreement, but with lower coherence, probably as it is more noisy. So the coupling is very consistent with the beam arriving on the OMC having a large low frequency jitter. I haven't been able so far to find an obvious source of the jitter. COMMp might be an explanation at 10mHz, DIFFp at 200mHz.

The right solution is to continue improving the working point of the interferometer, and also to reduce the angular fluctuations of the various interferometer angular degrees of freedom.

A wrong solution, which still might give good results, is to make the OMC (and SDB1 bench) chase the beam as it wanders at the output of the intereferometer. For that one would need to put back the RF signal in loop, and then blend with the SDB1 B5 QD2 above the loop UGF of the loop to avoid injecting too much noise above 10Hz (the RF signal (double demodulation at 56MHz and at 12kHz) has 1000 times more noise than the SDB1 B5 QD2 signal)

Figure 4 shows many things, but the most important are the 4 last plots. The TX and TY direction of SDB1 are controlled in different ways, with opposite flags of control. If I remember the logic correctly, in TX the high-bandwidth conditionally stable control is being disabled, as the fluctuations in alignment are above a safety threshold. This could explain why TX is so predominant in the SSFS coupling coherence, and maybe just enabling the SDB1 high bandwidth control by changing the safety thresold would resolve the issue.