I have tried to use again the signal on B1p QD1 112MHz for controling SDB1 instead of B5, repeating the test of Dec 11, and preparing more completely the automation of it.
The test was successful in the sense that all the lock acquisition could be done, with a few changes still done by hand that could be easily automated (change of filters at times when loop is open, and closing a galvo loop in LN2).
However, the bench has been moving much more in TY, with a peak at 100mHz. I don't know if this is due to a higher ground motion, something that I have forgotten to do this time. There is 1.5 hours of lock in LN2/LN3 with this signal in loop.
I have tried to revert back to the nominal configuration, but I haven't been able to relock afterwards. The oscillation on SDB1 TY have still been big, looking back at past locks it may actually be normal for some of the steps of the lock acquisition. Reverting back the SDB1 B5 QD2 offset to what they were earlier today seemed to have helped in reducing the oscillation when reaching CARM NULL 1F, and putting the offsets closer to zero helped even more. I have left these smaller offsets, which give a larger range that is linear on the quadrant around the set point. It may be problematic to lock the OMC or DC READOUT with that smaller offset.
This however did not allow to reach CARM NULL 1F, with unlocks in the first few minutes of CARM NULL persisting. It is unclear to me if it is the problem for the past two days coming back, or if there is a change in the SDB1 control that I forgot to undo, and that proves to be critical. The sea activity is also increasing with the sea activity flag in red, and the IP in yellow, which is definitely not helping.
------------------------
Adjusted SDB1 B5 QD2 offsets to values seen in LN2 earlier today
Changed error signal for SDB1 in SDB1_LC using external file
Moved SDB1 drift control from DRMI_LOCK state 50
(LOCKED_DRMI_3F_ASC_FULL) to ITF_LOCK state 69
SETTING_CARM_OFFSET_MID, and disable B1p QD1 galvo in ITF LOCK state 69
Change SDB1 set points to position seen in LN2 earlier today
Couldn't start SDB2 pico process. So did not zero the galvo correction
on B1p QD1. Signals are relatively small, below 2V, so decided to try
to continue like this.
Closing the loops as on Dec 11. There is a slow oscillation at ~100mHz
on TY, +/-3urad. In the time series by eye it looks correlated with
the vertical position of SDB1 and SDB2 (LC_Y). The oscillation in TY
is present both for low and high bandwidth control of SDB1, and
changing the gain by a factor 2 doesn't remove it. Moreover it is not
present on the correction of B1p QD2, which is out of loop for the SDB1
control.
It is likely a contamination of the B1p 112MHz QD signal. With the
SDB1 closed on that error signal, in the out of loop quadrant B1p QD2
the H oscillation in the correction of the galvo are smaller, than
earlier today where the ~100MHz oscillation was clearly present in the
horizontal correction.
The sea activity (0.1-1Hz ground motion RMS) was 4 times higher today
than during the last test of Dec 11. https://logbook.virgo-gw.eu/virgo/?r=68361
16:38:45 A few minutes after reaching LN2 close the B1p QD1 galvo with
low gain, takes about two minutes to zero the error signal offset put
in by the OMC drift control
17:30:40 reseting B1p QD1 galvo filter to reduce its gain by a factor 10
Unlocked manually at 18:08 UTC to restore standard configuration for SDB1 control
------- Notes for automation ----------------
DRMI_LOCK ACQUIR DRMI 1F ASC drift enable B1p QD1 galvo
DRMI_LOCK ACQUIRE DRMI 1F disables SDB1 B5 drift control
DRMI_LOCK DRMI_3F_ASC_FULL enables SDB1 B5 drift control, corresponds to state 64 in ITF_LOCK
ITF_LOCK state 69 CARM OFFSET MID, turn off B1p QD1 galvo
Use normal GalvoFlt_B1p_QD1_H and GalvoFlt_B1p_QD1_V filters in DRMI
Use GalvoFltLowGainH and GalvoFltLowGainV filters in LN2
