In order to understand the TF between the H_rec and the RIN, and disantangle the frequency response of DARM (differential double cavity pole) with the filtering of the RIN by the CARM pole (or commoin double cavity pole), I've plotted the TF between the B1_DC and the Pstab signal. The TF is relativelt flat and it seems that there is no sign of this CARM filtering. This can be an indication that the RIN is transmitted to B1_DC mainly by the high-order modes in the power-recycling cavity, rather than the TEM00. 

There is an issue with the PSTAB noise projection figure. The frequency axis looks wrong by about a factor 2, as the violin modes are around 250Hz. There might be the wrong sampling frequency used in the PSD computation, or some similar typo.

I corrected the typo pointed out by Michal in the computation of the RIN projection.

The correct one is shown in Pic.1 (also compared with O3 sensitivity). Pic. 2 and Pic. 3 show respectively the low and high frequency TF from RIN to H_rec.

The PSTAB lines used to monitor the PSTAB  coupling with DARM have been put permanently at the ISYSnoise server level:

• PSTAB0 with 1e-1V@43Hz
• PSTAB1 with 5e-3V@236Hz
• PSTAB2 with 5e-3V@1501Hz

As the PSTAB_HF_CORR  channel is too noisy,  the PSTAB_PDa_AC_MONIT channel is used to monitor the injected lines, and the amplitudes of the lines extracted on this channel are used as normalize factor to build  the LSC_DARM_PSTAB{0,1,2}_COUPLING channels (see plot)

We ran a Bruco with a subset of channels ENV,NN,Clock,SBE,DAQ:
https://scientists.virgo-gw.eu/DataAnalysis/DetCharDev/users/fiori/bruco/EDB_B1s_QD2_RF_56MHz_I_OMC_X_I_50Hz_1364221818/

The top coherent channels are the following:

SBE_SWEB_diff_speed_bench_MIR_z
ENV_CEB_RF_6MHz_Q_FS
ENV_CEB_RF_81MHz_I

Attached plots are a screenshot of the Bruco table and the spectra and coherence of these channels with LSC_DARM and EDB_B1s_QD2_RF_56MHZ_I_OMC_X_I.

For the laser intensity noise coupling there are several additional aspects to take into consideration, which are included in the simulink noise budget.

The coupling is changing in time with the CMRF, so to do an up to date projection one needs to to update the projection based on the height of the permanent PSTAB lines in the sensitivity.

Figure 1 shows a projection from this morning. PSTAB is the domdinant part of the LSC line in that curve. If one looks at the PSTAB line at 1501Hz, the projection (cyan curve), matches well the calibrated DARM spectrum (blue curve), but is a factor 5 above the line in h(t) (black dashed line). I expect the reason is that the PSTAB line also appears in SIB2 B2 8MHz I, and Hrec sees the coherence, and removes the line. We really need a working h(t) without noise subtractions to avoid this type of confusion.

Figure 2 shows the corresponding noise projection of just the laser RIN. The green line is the contribution of the shot noise on the PSTAB photodiodes, the black is the contribution of the sidebands RAM, as about 3% of the power on the PSTAB photodiodes is due to sidebands. And the spectrum of the RAM is taken from VIR-1225B-19. Below 500Hz the sideband RAM contribution is dominant for the RIN, higher than PSTAB photodiode shot noise. Installing the RAM servo on the 6MHz and 56MHz would decrease that contribution by a factor 2 or 3 (there would still be the contribution of the 8MHz and 81MHz RAM).

Figure 3 comparing the time with the PSTAB noise injection (purple) and some more quiet looking time before that (blue curve), there is some coupling to the out-of-loop monitors of the SSFS error signal (on B2 6MHz I), but this looks minor compared to the noise present there at normal times. I think this explains why Hrec partially removes the 1.5kHz PSTAB line from h(t), but that the coupling path of PSTAB to DARM doesn't actually go through the SSFS.

On Michal's request the ISYSnoise server has been updated to provide the LSC_DARM_PSTAB{0,1.2}_{I,Q} channels in order to use them as error signals if needed . Theirs demodulation phases can be tuned  using the DARM_PSTAB{0,1,2}_phi0  the channels.  The  id  0,1, and 2 refers to the  lines at 43Hz, 236Hz and 1501Hz .

The new channels are not yet sent to the ISC rtpcs  to be used in some control loops