Displaying report 1-1 of 1.
Reports until 12:51, Tuesday 07 February 2023
H1 SUS (ISC)
jeffrey.kissel@LIGO.ORG - posted 12:51, Tuesday 07 February 2023 - last comment - 13:50, Thursday 16 February 2023(67266)
Gathered PRM and SRM Global Actuation Transfer Functions 
J. Kissel, E. Capote

Elenna and I characterized the "global actuation" transfer functions, i.e. each actuation stage drive to optic response, or M3 to M3, M2 to M3, and M1 to M3, for two HSTSs today, PRM and SRM. We gathered these TFs with the former damping loops (which we've again reverted to as of Jan 26 2023; LHO:67031), and with the "level 2" damping loops (installed, commissioned, and tried last on Jan 18 2023; LHO:66859) in hopes to better understand and either prove or disprove to ourselves that the damping loop change impacts the ISC loops in L, P, and Y, that are wrapped around them.

It's too much data to process coherently, and we don't yet have a complete set, so here I only document that we performed the measurement, and where the data is saved. 
For future reference, 
    - PRM was ALIGNED, with alignment offsets ON
    - SRM was MISALIGNED with *misalignement* offsets ON
    - For both SUS, we had the M1, M2, and M3 coil drivers in state 1, 2, and 2, respectively.

The data has been committed to the SVN here:
PRM M1:
    /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/PRM/SAGM1/Data/
        2023-02-07_1824_H1SUSPRM_M1_M3_CDBIOState_1_L_0p03to50Hz_new_and_old.xml
        2023-02-07_1824_H1SUSPRM_M1_M3_CDBIOState_1_P_0p03to50Hz_new_and_old.xml
        2023-02-07_1824_H1SUSPRM_M1_M3_CDBIOState_1_Y_0p03to50Hz_new_and_old.xml
    (REF traces are with old M1 DAMP damping filters, live trace is with new Level 2 filters)

PRM M2:
    /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/PRM/SAGM2/Data/
        2023-02-07_1940_H1SUSPRM_M2_M3_CDBIOState_2_L_0p03to50Hz_new_damping.xml
        2023-02-07_1940_H1SUSPRM_M2_M3_CDBIOState_2_P_0p03to50Hz_new_damping.xml
        2023-02-07_1940_H1SUSPRM_M2_M3_CDBIOState_2_Y_0p03to50Hz_new_damping.xml
    (REF traces are with old M1 DAMP damping filters, live trace is with new Level 2 filters)

PRM M3:
    /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/PRM/SAGM2/Data/
        2023-01-24_1840_H1SUSPRM_M3_WhiteNoise_CDBIOState_1_L_0p03to50Hz.xml   # actually in State 1, unlike what's said above
        2023-01-24_1840_H1SUSPRM_M3_WhiteNoise_CDBIOState_1_P_0p03to50Hz.xml
        2023-01-24_1840_H1SUSPRM_M3_WhiteNoise_CDBIOState_1_Y_0p03to50Hz.xml
    (REF traces are with old M1 DAMP damping filters, live trace is with new Level 2 filters)

SRM M3:
    /ligo/svncommon/SusSVN/sus/trunk/HSTS/H1/SRM/SAGM3/Data/
       2023-02-07_1828_H1SUSSRM_M3_WhiteNoise_L_0p03to50Hz_ODF.xml
       2023-02-07_1828_H1SUSSRM_M3_WhiteNoise_P_0p03to50Hz_ODF.xml
       2023-02-07_1828_H1SUSSRM_M3_WhiteNoise_Y_0p03to50Hz_ODF.xml
    (Blue ref trace is undamped, Black ref trace is old filters, Red live trace is Level 2 filters)


We'll get more data and begin to process it in the coming weeks.
Comments related to this report
craig.cahillane@LIGO.ORG - 18:50, Wednesday 15 February 2023 (67457)
Link 

Status update on suspensions.py.

I've got something working for the HSTS model for applying the damping loops.
It uses filter_module_state_extractor.py to get the damping loops from foton files,
then the Richard Murray python control library to form the closed loop with M1 drive -> damp loop -> M1 disp.

With help from Lee to get Jenne's old damping loop code running, and Jeff explaining the measurements, I analyzed these measurements taken here.

The first three plots show the M3 -> M3 disp witness / drive measurements for the SRM.

The next nine plots show the M? -> M3 disp witness / drive for the PRM.

The model is not perfect. 
Somehow the damping applied does not perfectly match up with the damping measured.
I applied some by-eye damping loop gain adjustments to get good agreement with the M3 -> M3 measurements, sometimes with the cost of more imperfect M1 -> M3 and M2 -> M3 measurements.

The damping loop gain fudges are reported here:

SRM: {"L": 0.59, "T": 1.0, "V": 1.0, "R": 1.0, "P": 1.0, "Y": 0.7}
PRM: {"L": 0.3, "T": 1.0, "V": 1.0, "R": 1.0, "P": 1.1, "Y": 0.3}


For ASC (P and Y dofs), we probably care more about the accuracy of the M2 -> M3 damped suspension response.
For LSC (L dof), we care mostly about the M3 -> M3 damped suspension response.


This code lives in 
/ligo/home/craig.cahillane/Git/aligo_controls
To run the PRM test suite, for instance do the following in a fresh terminal:

$ conda activate aligo_controls
or, if that does not work for you straight off,
$ conda activate /ligo/home/craig.cahillane/.conda/envs/aligo_controls

$ cd /ligo/home/craig.cahillane/Git/aligo_controls/
$ pytest-watch

This will automatically run the unit tests in the tests/ directory.  Those will produce some plots in tests/figures/ after some five minutes or so.

Work continues to get the quads going.
Non-image files attached to this comment
jeffrey.kissel@LIGO.ORG - 13:50, Thursday 16 February 2023 (67467)CSWG, ISC, SYS
Link 

Tagging CSWG in light of Craig's work developing python infrastructure for comparing SUS models to measurement.
Displaying report 1-1 of 1.