The afternoon we monitored the powers sent to the detection while the CITF was being aligned. This corresponds to the situation shown on Fig.1 between 18:44 and 19:02 utc, with the end mirrors misaligned. One can see that the maximum power on B1p_PD2 is 1.25 mW, corresponding to a power of 5.2W on the SDB1 bench.
Then the CITF was aligned with the arm cavities locked on the green beam. This corresponds to the situation between 19:11 and 19:15 utc. In this configuration the maximum power reached on B1p_PD2 is 1.5 mW, corresponding to 6.2 W on SDB1.
In case of accidental misalignment of the SR mirror (which should not happen with the new safety), the ghost beam that could hit the B1/B5 diaphragm would be lower than 4W, and could therefore only marginally exceed the safe threshold value (~3W). This statement is valid with 25W of injected power.
It is however crucial to protect the OMC and the B1 photodiodes during the CITF alignment. The procedure should to be to always keep the OMC slow shutter closed during the CITF alignment and the lock acquisition until the ITF is locked at dark fringe.
Fig.2 shows the fluctuations of the SDB1 B5 QD2 signals and of the angular corrections sent to the marionetta. One can see that at the beginning of the CITF alignment we had left the SDB1 drift control closed and large fluctuations were observed in the TY correction. Therefore, we decided to open the drift control (after updating the set points to the current values) before aligning the PR mirror. The procedure to follow concerning the bench angular control remains to be refined.