J. Kissel

Since updates to the CAL-CS model have been a little sporadic over the past few days, I recap below what happened when, as logged by the filter archive.
What Changed                                                                    Local Time (PDT)    Corresponding Archive of Change    Corresponding aLOG

SUS dynamical model update is complete                                          Thu Sep 10 13:50    H1CALCS_1125953445.txt             LHO aLOG 21322

Cavity pole frequency changed from 389 to 341 [Hz] (+optical gain name change)  Thu Sep 10 14:03    H1CALCS_1125954209.txt
Optical gain is renamed (no substantial change)                                 Thu Sep 10 14:03    H1CALCS_1125954234.txt   
DARM_ERR Optical gain is tuned                                                  Thu Sep 10 17:49    H1CALCS_1125967809.txt             LHO aLOG 21385

Installed inverse actuation filter                                              Sun Sep 13 22:35    H1CALCS_1126244167.txt             LHO aLOG 21487

Kiwamu changes L1 and L2 violin modes                                           Mon Sep 14 09:03    H1CALCS_1126281835.txt             LHO aLOG 21500

Copied new inverse actuation filter over to blind injection                     Mon Sep 14 09:08    H1CALCS_1126282148.txt             
Changed the names of the blind injection filter banks (no substantial change)   Mon Sep 14 09:09    H1CALCS_1126282176.txt             LHO aLOG 21499
(Note the filter files are tagged with the GPS time that they're created. so they're more accurate (to the second) than the local time I quote.)

Given that the DARM calibration was tuned at Sep 10 2015 17:49:00 PDT = Sep 11 2015 00:49:00 UTC, and Kiwamu started the PCAL to CAL-DELTAL_EXTERNAL transfer function at Sep 10 2015 18:08:12 PDT = Sep 11 2015 01:08:12 UTC, I would conlcude that 
the front-end calibration has been stable for LHO since the first lock stretch that was placed in observation mode after that measurement, which starts at 
     1125972087 = Sep 11 2015 02:01:10 UTC = Sep 10 2015 19:01:10 PDT
until the last observation stretch -- which ended at 
     1126281754 = Sep 14 2015 16:02:17 UTC = Sep 14 2015 09:02:17 PDT
-- before Kiwamu's change at Sep 14 2014 09:03:00 PDT = Sep 14 2015 16:03:00 UTC. 

I say the front end calibration has been "stable" instead of "valid" because because we know that Kiwamu's work this morning was to correct a systematic error from a bug that had been found.

We only briefly took the IFO out of observation mode (though the IFO stayed happily locked) while Kiwamu loaded new SUS filters and I the copy of inverse actuation filters. We resumed a new observation segment Sep 14 2015 16:10:43 UTC.

As noted in Kiwamu's most recent aLOG about updating the calibration this morning (LHO aLOG 21500), his update changed the UIM and PUM stages, and for features well-above the cross-over frequencies of those stages (TST/ PUM ~30-40 [Hz], PUM / UIM ~ 2 [Hz]). The update was a bug fix, and it means that the front is now more accurately representing the violin modes that were already correct in the DARM model. We all suspect this is a small change in the overall actuation function, but the recent discovery of how poorly the PUM has been rolled off (LHO aLOG 21419), means we have to verify. 

After these messages, we'll be riiiight back. *whistle glitch*

P.S. Since Matt and I performed all of our inverse actuation filter design on the original matlab model of the DARM actuation function, the inverse actuation filter design and implementation is unaffected by Kiwamu's change, nor was it impacted by the systematic error that he found.