Virgo Logbook

AdV-COM (AdV commissioning (1st part) )

bersanetti, allocca, chiummo - 23:32 Monday 25 June 2018 (41986)

ITF Recovery: back in LOW_NOISE_3, max BNS Range > 35 Mpc

During the afternoon shift we continued the recovery of the ITF:

we tuned the last demodulation phase, the B5_QD2 one for the PR/BS automatic alignment loops;
no other major tuning was needed but the gain of the SSFS for the most part of the lock acquisition;
the 100 Hz line was very high tonight, especially so on DARM and PRCL; on DARM it gets quite reduced when moving from RF to DC sensing; since the DARM spectrum was quite low at this point, we reduced the UGF line amplitude of a factor 6;
once in LOW_NOISE_3, after the BS loop switches to full bandwidth, we engaged by hand the new filter for MICH (again, removing the notches for, probably, filter allocation reasons);
that made both the sensitivity and the range get better, especially at low frequency, resulting in a BNS Range floating around 35 Mpc, reaching 36 Mpc for a very brief time!
a stretch of clean data is from 19:16:00 UTC to 19:39:00 UTC;
after this period of time, we had a sudden drop in the sensitivity, with the bucket often rising; it may be interesting to look for correlations during this phase;
however, the ITF and/or the IMC are a bit shaky after an unlock (natural or induced) and the first trial usually fails at CARM2MC;
at the end we played a bit with the setpoint of the BS_TY loop, with the usual change in the CMRF but no effect on the sensitivity;
we did a few more trials to engage the full MICH filter, increasing both its order and the one of the LSC_NONE 'placeholder', without success; in the attached log there is one attempt, where one concerning thing is the "15/12" relative to the order, where without the 3 notches the log says "12/12";
a very weird unlock, triggered by the 100 Hz on DARM but with lots of lines and glitches on the spectrum, occurred around 21:12:11 UTC.

We left the ITF with target state LOW_NOISE_3, with the old MICH filter and in the AUTORELOCK mode.

The last working configuration for Metatron was checked in as revision r73511.

Non-image files attached to this report

filterlog.txt

Comments to this report:

masserot - 6:06 Tuesday 26 June 2018 (41988)

In order to perform some checks I put the following lines:

in LSC_MICH.cfg line80: ACL_FILTER_CH_OPTION MICH_OUT -1 "sin_inout" "linear" 0 10 0 50
in LSC_DARM.cfg: line57: ACL_FILTER_CH_OPTION DARM_OUT -1 "sin_inout" "linear" 0 10 0 50

where the last field (50) means that the filter computation is done using the low noise pole filter implementation as we may have pole @0.06Hz while the input filter channel sampling frequency is 10KHz .

After these modifications, the ITF relocked at LOW_NOISE_3

mwas - 11:33 Tuesday 26 June 2018 (41991)

The demodulation noise has improved significantly.
Figure 1 shows the projection which is now more than a factor 10 below the noise curve at most frequency, and with lines reaching the noise floor.
This is likely due to the swap of the GPS receiver.

Figure 2 and 3 show the improvement in the demodulation phase spectra. The SPRB 10MHz clock and the B4 112MHz phase spectrum is a factor 10 lower. Note, the atomic clock comparison is not useful as there was some 2*pi wrapping jumps in the reference time. This also explains the disappearance of several bumps in the PRCL spectrum between 50Hz and 150Hz, that where caused by the SSFS loop error signal being spoiled by the demodulation phase.

Images attached to this comment

masserot - 18:38 Tuesday 26 June 2018 (41998)

I made some tests , in fact the use of the low noise filter implemenation doesn't improve the situation with the "MICH_control6*MICH_main*PRCL_violin" filter.

Investigations inprogress