Adding half-finished analysis here for now.

I plotted some relevant channels during this test, including POP18 and 90 I and Q.

Oddly, it seems like the total amount of POP18 stays the same no matter how the RF9 is moved.
Similarly, the POP90 moves up the most for a reduction in RF9.

The only effect I can see in POP18 from the RF9 reduction is a phase rotation.
A similar phase rotation is visible for POP90 from the RF45 reduction.

I can't explain why this, only that it is not what I expected from the sideband power witnesses.
The carrier channels (especially the arm power channels) seem to rise and fall as expected.
More quantitative analysis to come.

I repeated the sideband power to total power analysis from Daniel's alog alog 59058.
The purpose is to re-evaluate how much carrier, 9 MHz, and 45 MHz we have in each port since May 2021 last year (from the first locks after the ITMY was replaced).
This is particularly interesting analysis for the low OMC transmission we've witnessed.

The principle of the measurement is based on the idea that we know how much we are changing the modulation index in radians,
so if the DC power at a port changes (or does not change) we can use the measured DC power ratio between the modulation index changes to determine the sideband power there.
This was done for both 9 and 45 MHz.


Results

Sideband to Total Power ratios
Channels
9 MHz
45 MHz
Carrier
H1:IMC-PWR_IN_OUT16
0.013
0.015
0.972
H1:IMC-IM4_TRANS_NSUM_OUT16
0.013
0.015
0.972
H1:LSC-REFL_A_LF_OUT16
0.056
0.045
0.899
H1:LSC-REFL_B_LF_OUT16
0.054
0.044
0.903
H1:LSC-POP_A_LF_OUT16
0.020
0.009
0.970
H1:ASC-POP_A_NSUM_OUT16
0.020
0.010
0.970
H1:ASC-POP_B_NSUM_OUT16
0.020
0.010
0.970
H1:ASC-AS_C_NSUM_OUT16
0.122
0.526
0.352
H1:ASC-OMC_A_NSUM_OUT16
0.123
0.580
0.296
H1:ASC-OMC_B_NSUM_OUT16
0.123
0.629
0.248
H1:ASC-X_TR_A_NSUM_OUT16
0.005
0.008
0.987
H1:ASC-X_TR_B_NSUM_OUT16
0.005
0.008
0.987
H1:ASC-Y_TR_A_NSUM_OUT16
0.005
0.008
0.987
H1:ASC-Y_TR_B_NSUM_OUT16
0.005
0.008
0.987

At first glance, it seems that our ratios are largely the same as they were in May 2021, except there is way more carrier and 9 MHz (and less 45 MHz) at the AS port.
The AS PDs seem to be telling us the same story regarding 9 MHz content, but not for 45 MHz content, unclear why.
But there appears to be around 58% 45 MHz light in the AS port, as opposed to 85% in 2021.

Otherwise, there still seems to be 90% carrier in REFL, 98% carrier in POP, and 100% carrier in the arms.



Code

The code lives in /ligo/gitcommon/labutils/mod_depth_up_down_test/code/mod_depth_analysis.py
or https://git.ligo.org/aligo_commissioning/labutils/-/tree/master/mod_depth_up_down_test