The issue is actually visible earlier on in CARM NULL 1F, with a mis-balance of 56MHz sidebands on B1p and a higher 12MHz mag on B1p.

Figure 1 shows the situation from Aug 10, before the mainteance with a factor 3 in power between B1p 56MHz  USB and LSB, and B1p 12MHz mag at ~4uW when the OMC shutter opens (B1s goes from zero to a few tens of mW).

Figure 2 shows the situation after the maintenance of Aug 13, the B1p 56MHz USB/LSB ratio has increased to factor ~6, and B1p 12MHz mag to ~10uW.

A similar issue was present during the horizontal shift of WI DAS

Figure 3 shows when the WI DAS was shifted horizontally which created too much power on B1s, the B1p 56MHz USB/LSB has a ratio of ~4, and B1p 12MHz mag is ~6uW

Figure 4 shows after the WI DAS 1.5cm horizontal shift was reverted, the B1p 56MHz USB/LSB ratio has decreased to ~3, and B1p 12MHz mag is ~4uW

All of these cases are much worse than before the July installation.

Figure 5 and 6 shows two examples. The B1p 56MHz USB/LSB ratio was around ~1.5, and the B1p 12MHz mag ~2uW.

Last year we already had issue with lock acquistiion to DC read-out when B1p 12MHz mag was large (I remember we couldn't lock when it was above ~5uW), and the B1p 56MHz LSB/USB powers misbalanced. A longitudinal offset on MICH can create a misbalance of 56MHz sidebands on B1p. And sometimes last year, but not always, adding a MICH offset in CARM NULL 1F was able to compensate for this misbalance.