Displaying report 1-1 of 1.
Reports until 16:10, Tuesday 02 May 2023
H1 CAL
jeffrey.kissel@LIGO.ORG - posted 16:10, Tuesday 02 May 2023 (69265)
Reverted all settings for Front-end Computed, Demodulated Calibration Line Coherence-Based Uncertainty
J. Kissel

As we re-learned last Friday (LHO:69175), the coherence-based uncertainty estimate on the calibration-line-to-DARM_ERR transfer functions depends on the choice of the "Number of Averages" BUFFER_SIZE variables, and one can't independently change the DEMOD low-pass filters because their corner frequency (say, 0.1 Hz) should be kept in direct correspondence with the "FFT LENGTH" COH_STRIDE variables.

As of this morning's model changes (described in LHO:69226), we now have split out the BUFFSIZE (Navgs) and COH_STRIDE (FFT LENGTH) from a single set of values for the entire collection of calibration lines, to having an independent set of numbers for each line.

As a part of filling out these new values, I've performed two of the actions from LHO:69175,
    - Revert all DEMOD the band-passes to have a pass band that's +/- 0.1 Hz wide (what we had in O3) >> except for the 410.3 Hz line, processed by the PCAL_LINE2 demod -- see below
    - Revert all DEMOD I & Q low passes to a 10 second time constant, or 0.1 Hz corner frequency

The PCAL team as still trying to suck the last bit of juice out of having a PCALX calibration line within 0.1 Hz of the 410.3 Hz PCALY line. The narrow band pass is necessary while the PCALX line is so close at 410.2 Hz, because of now-known contamination of the optical gain and cavity pole frequency "TDCFs" LHO:68479. So we can't revert this just yet. At least for now, we can independently set the Navgs and FFT length appropriately. As such, for the 410.3 Hz, PCAL_LINE2 infrastructure, we're still using
    - the NBP410.3, with +/-0.03 Hz wide band pass
    - the LP40s I & Q lowpass filters with a corner frequency of 0.025 Hz
    - the now independent N Averages of 3, (H1:CAL-CS_TDEP_PCAL_LINE2_COH_BUFFSIZE = 3)
    - the now independent FFT length of 40 seconds, (H1:CAL-CS_TDEP_PCAL_LINE2_COH_STRIDE = 40)
    
All the changes have been accepted in the SDF system.

The reversion process was easy with a little bit of command line BASH code (after confirming that there are LP10s in FM1 of all the DEMOD banks):
$ export ACT=PCAL; export LINE=LINE1; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_COH_BUFFSIZE 13; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_COH_STRIDE 10; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_GRM_MDN_ARRAYSIZE 2049; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_GRM_AVG_ARRAYSIZE 160; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_GRM_STRIDE_SEC 0.0625; \
caput ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_SUB_UNC_THRESH 0.09; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_${ACT}_DEMOD_I FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_${ACT}_DEMOD_Q FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_${ACT}_DEMOD_I FM2 OFF; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_${ACT}_DEMOD_Q FM2 OFF; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_ERR_DEMOD_I FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_ERR_DEMOD_Q FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_ERR_DEMOD_I FM2 OFF; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_ERR_DEMOD_Q FM2 OFF; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_EXT_DEMOD_I FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_EXT_DEMOD_Q FM1 ON; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_EXT_DEMOD_I FM2 OFF; \
z switch ${IFO}:CAL-CS_TDEP_${ACT}_${LINE}_EXT_DEMOD_Q FM2 OFF;
where to move on to the next demod, I up-arrowed, and made "ACT" either PCAL or SUS, and changed "LINE" to go from LINE1 through LINE10.
Displaying report 1-1 of 1.