I have looked back at O3 compared to O4 commissioning at the quadrature of the SSFS error signal, to try to understand if the issue is new of excess noise in the SSFS sensing is new or not.

Figure 1 shows the SSFS error signal quadrature in blue O3, in red a week ago, and in yellow November last year. The amplitude of the O3 signal has been adjusted so that the SSFS in-phase signal has the same height for the 1111Hz line. It seems that we have significantly improved the sensing, by using the B4 6MHz signal, instead of B4 56MHz without signal recycling. This seems reasonable, the 56MHz had an optical gain of ~8 during O3, while the 6MHz has an optical gain of ~35. Also the 6MHz is not cut-off by the PD capacitance that has a pole at ~30MHz.

It is not clear if the excess noise we have is new or not compared to O3. If that calibration is really correct, we should expect to see a ~30% change in level in the O3 data below 300Hz compared to 1kHz, but that does not seem the case. Which would suggest the problem is new.

FIgure 2 shows the SSFS in loop signal, that have after adjustment the 1111Hz line roughly at the same height, also the background around it is at a similar level. It is not a perfect way of calibrating the signals, as the height probably also depend on the SSFS loop gain, but it should be good within a factor of 2.

Figure 3, what is clearly worse is the B4 DC RIN compared to O3, with an extra ~1/sqrt(f) like noise. I don't have an explanation why that would be the case.