J. Kissel

I'm are trying to figure out the best metrics for showing off the improvements to the OSEM PD's satellite amplifier's whitening improvements.
Thus far, Oli's been using the input to the damping loops as the metric, using a regression of the corresponding ISI's GS13s to subract out a fit of how much of that sensor signal is seismic noise, and dividing out the loop suppression -- see LHO:86149 for the most recent examples comparing before vs. after the sat amp upgrade. 
Without the presence of any other noise or control signals, that should be a fair comparison of the OSEM PD's sensor noise improvement.
However, for a lot of these comparisons ISC control signals are complicating the comparison -- usually at low frequency where ISC control is typically distributed to the top masses.

I use this aLOG as an example of how to better understand this contribution breakdown for a relatively simple suspension -- H1SUSMC1 -- which only has P and Y ISC control from the IMC WFS. (Longitudinal control for IMC L is fed to MC2). This will also be interesting in the future 
    :: in the context of how SPI and other sensors may improve the cavity motion, 
    :: in terms of what DOFs and loop's worth of control drive at which frequencies -- important for discussions along the lines of "DOF [blah] is dominating the control signal, and the actuator cross-coupling for M1 drive of DOF [blah] to M3 optic DOF [blorp] is large, so let's reduce the DOF [blah] drive," and
    :: in terms of whether/where implementing ISI GS13 estimator feedforward will improve things.

To understand how much of the damping loop *error* signal is composed of ISC *control* signal, I look compare the 
    - the ISC control signal,
    - the DAMP *control* signal, against the
    - the MASTER total control request,
all calibrated to the same point in the control system -- where the control output is summed and in the OSEM basis; just down-stream of the EUL2OSEM matrix, and just upstream of the COILOUTF filters which compensate for the coil driver frequency response (uninteresting for this study).

Pitch -- the T1T2T3 actuators
    (3) Attachment 3 Pitch Noise Comparison excerpt from Oli's LHO:86149. These are times when the IMC was LOCKED, so there should be ISC control. But, see the expected factors of 2x-to3x improvement in the OSEM noise below ~5 Hz. So, maybe the ISC control is so low in bandwidth that its affect isn't impacting this study. But, we can see that there's clearly some other loop suppression that has not been accounted for, so maybe it *is* high bandwidth? Let's find out.
    (1) Attachment 1 Comparison of ISC pitch, DAMP pitch, as well as the other DAMP DOFs that use the T1, T2, and T3 actuators -- Vertical and Roll -- control signals.
    Here, we can clearly see that the damping loops are dominating the T2 (and thus T3) control signal above ~ 0.5 Hz, or conversely, the IMC WFS DC coupled control is dominating below 0.5 Hz.
    (2) Attachment 2 shows that the T2 and T3 sensors receive identical request (mostly an out-of-phase combination of Pitch and Roll damping request, as expected from the EUL2OSEM matrix), and T1 drives mostly Roll damping request. The vertical drive request is subdominant at all frequencies.
    (3) Attachment 4 shows the open loop gain and loop suppression TF magnitudes for pitch. The loop suppression here looks very much like the inverse of the shape of the ASD left in the pitch regression, making me worried that Oli's automated regime for removing the loop suppression isn't perfect... I'll ask.

Yaw -- the LFRT actuators
    (7) Attachment 7 The before vs. after comparison of OSEM noise
    (5) Attachment 5 Similar comparison of ISC vs. relevant DAMP control -- showing IMC WFS control dominating only below ~0.2 Hz.
    (6) Attachment 6 As expected from the EUL2OSEM matrix, the LF and RT actuators receive the same control.
    (8) Attachment 8 Shows the open loop gain and loop suppression TF magnitudes for the Yaw damping loop.