Displaying report 1-1 of 1.
Reports until 13:47, Thursday 27 July 2017
H1 SUS (DetChar, ISC, OpsInfo)
jeffrey.kissel@LIGO.ORG - posted 13:47, Thursday 27 July 2017 - last comment - 18:49, Thursday 27 July 2017(37818)
ETMX Violin Modes Get Rung Up After Post-EQ Lock Acqisition; MODE4 Settings Updated
J. Kissel, J. Warner

We've lost lock due to a non-exciting earthquake, re-acquired up to DC Readout Transition watching the violin modes carefully, but we lost lock there (while sill controlling DARM with ETMX). Upon the *next* re-acquisition, though, the ETMX violin modes were incredibly rung up. 

As such, we stopped at Shutter ALS and baby-sat the ETMX modes until the DCPD RMS was reduced from 1e4 to 2e3. We used Ed Merilh's as our babysitter configuration for the now-problematic ETMX MODE 4 at 505.805 Hz, namely driving in Length, Pitch, and Yaw, ~75 gain, FMs 1 (505.805), 2 (-60 deg), and 4 (100dB).

This was successful, so we've now programmed this into the guardian.

We highly recommend operators hold any re-acquisition from now on at SHUTTER_ALS, wait for the OMC to lock, then grab a ~0.005 mHz spectra, and opening the outputs of ETMX first 5 violin MODE filters to make sure that violin modes are not rung up. Then, if violins are OK (i.e. RMS of DCPDs is less that 2e3 [ct]), you can go to the first Violin Mode Damping state but be prepared to turn the damping filters off before things get terrible!

Only one more month of this, team -- we can do it!
Comments related to this report
jeffrey.kissel@LIGO.ORG - 18:44, Thursday 27 July 2017 (37832)OpsInfo
It may be that the violins are so rung up that you're saturating AS_C, which makes SRC2 loop unstable -- so you'll see pitch go into slow oscillation.

To prevent this, after every lock loss (for now), park the IFO in ENGAGE_REFL_POP_WFS, and run a DTT spectra with the following channels.
H1:OMC-DCPD_A_IN1
H1:OMC-DCPD_B_IN1

H1:ASC-AS_C_SEG1_IN1
H1:ASC-AS_C_SEG2_IN1
H1:ASC-AS_C_SEG3_IN1
H1:ASC-AS_C_SEG4_IN1

To move past ENGAGE_REFL_POP_WFS, you need the AS_C segments to have an RMS of ~2000 [ct] or less below the 500 Hz violin mode cluster, and all peaks should be below 20000 [ct/rtHz].

Once that's true, you can move on the turn on the automatic violin mode damping as normal, so select VIOLIN_MODE_DAMPING_1.
To move past that, it's the same criteria: and RMS of ~2000 [ct] or less below the 500 Hz violin mode cluster, and all peaks should be below 20000 [ct/rtHz].

It won't be until you get an RMS of ~100 [ct] RMS that you'll stop seeing huge shoulders around the lines, so *keep baby sitting*.
jeffrey.kissel@LIGO.ORG - 18:49, Thursday 27 July 2017 (37833)
Templates can't be attached 'cause they're too big.

Look in 
/ligo/home/jeffrey.kissel/Templates/
H1ASC_AS_C_forViolins.xml
H1DCPDs_forViolins.xml
Displaying report 1-1 of 1.