Hrec peaks at 101 Hz, 107 Hz, 157Hz that we know are seen by SSFS, are seen also in RFC error signal also at times when the ITF is not locked and SSFS is not engaged (Figure 1, blue line around 4:00 on June 5, corresponds to RFC on and SSFS off, as shown in Figure 4). Figure 2 shows these peaks and the coherence with RFC error signal, that when SSFS is engaged is the same as SSFS.

Paolo's suggestion is that these might be modes of SDB1 or something on it (diedron?). It might be worth trying to inject noise on the bench to check.

Also the noise path is not clear. Back in November 2017 (40120) the amplitude of these peaks in hrec was clearly affected by the SSFS gain (reducing the gain the hrec peaks amplitude increased), we repeated the test recently on May 3, 41630 but we saw no clear effect  (Figure 2). The effect might have been masked in Hrec by the timing noise,  It is maybe worth to repeat the test now with cleaner hrec.

Figure 3 shows that at these peaks there is good coherence between SSFS and MICH (much less with PRCL).