Today we looked at the data from last night, where we left the ITF in auto-relock with the SSFS; there were a number of unlocks, a few of them seemed due to a stepped ramping of the SSFS correction, which eventually saturated at +-10V. In Figure 1 there is some cleaner trend of a few days ago, in Figure 2 a single instance and in Figure 3 its spectrogram.
We investigated this issue further, and what is seems is that there could be something related to how the correction is computed, maybe in the digital noise correction, given the discrete nature of such trend? We see similar trends also in the channels related to the Digital Noise, but not on the DC power of B4 (see Figure 4).
During a single instance, we could also see a subsequent increase of some broadband noise at each step of the stepped ramp (Figure 5).