Virgo Logbook

Virgo Runs (C8)

ruggi, genin - 18:36 Monday 08 May 2017 (37501)

Coupling of SDB1_LC to DARM

The fluctuations of the sensitivity are dominated by the power fluctuations of B1p, which are due to the fact that the ITF alignment is not yet controlled by the quadrants (fig 1).

When the noise is high a lot of structures grow, which resamble to mechanical resonances (fig 2).

A good candidate to be the source of that noise is the angular control of SDB1. Fig 3 is the transfer function between SDB1_LC_TX_corr and DARM, when the noise is maximally coupled.

Images attached to this report

Comments to this report:

mwas - 12:52 Tuesday 09 May 2017 (37515)

Good catch! The roll-off on SDB1_LC_TX_corr (and TZ) is missing.
For comparison in the middle bottom pane is the SDB1 and SNEB TX feedbacks. The bench suspension are completely different, so the two are not really comparable. But there is clearly a gap of a few orders of magnitude that should be easy to fill.

As a side note. In the upper left plot is shown the signal from the bench local controls SDB1_LC_TX, and from the quadrant SDB1_B5_QD2_TX_err. The two are added to generate the error signal for the SDB1 TX control, SDB1_LC_TX_err. Around 0.5Hz the drift control signal from B5 becomes significant. Above a few Hz it is completely negligible. And will state that way as the roll-off filter is applied after the sum.

Images attached to this comment

bonnand, mwas - 18:50 Tuesday 09 May 2017 (37532)

We added a roll-off filter to the TX (and TZ) control of SDB1.

It seems that the coherence between error signals of SDB1 in TX and DARM has been killed, see figure attached (purple is during C8, blue is after adding a roll-off).

We are not sure the filter used to control SDB1_TX is optimal, let's see how things are with the roll-off, it might be a good idea at some point to inject noise and find a better filter for the control of SDB1 in TX

Images attached to this comment