Virgo Logbook

Detector Characterisation (Broadband noise)

mwas - 9:25 Tuesday 20 September 2022 (57110)

A check on input beam 56MHz correlated noise

Alain reminded me that DAQ FMOD 56MHz is coherent with many length control signals at CARM NULL. It is not new, it was already the case a few months ago.. This signal looks at the beam after the input mode cleaner, demodulated at 56MHz.

Figure 1 summarizes the situation, the purple line is a time at CARM null and the blue line at a time with just the arms locked. When at CARM null a shelf of noise up to 60Hz appears in DAQ FMOD 56MHz phi/mag, but also the 150Hz line becomes very hight and the 157Hz bump appears.

Back during O3 there was a large bump at 18Hz in DAQ FMOD 56MHz, it was at a few 1e-4 on phi channel, so would be clearly visible now if still present. The 18Hz bump was related to scattered light between the in air injection bench and the laser bench. It does not seem to be present at the moment.

Figure 2 shows a spectrogram of DAQ FMOD 56MHz mag, The excess noise at 150Hz line, 157Hz bump and between 30Hz and 50Hz is visible. And it is clearly turning on and off in coincidence with CARM null lock. That night the interferometer was locked during the first half of the night, then unlocked and relocked for a brief time, and then stayed unlocked for the rest of the night. So all of this excess noise is brought back onto DAQ FMOD 56MHz by the interferometer control (most likely the SSFS, but I haven't tried to investigate if it happens at the exact time the SSFS is turned on).

Do we know if the sideband are balanced on the input beam? Maybe scanning the RFC to confirm that the lower and upper 6MHz and 56MHz are balanced at the input would be useful.

Images attached to this report

57110_1663657461_Screenshot from 2022-09-20 07-38-42.png

57110_1663657945_Screenshot from 2022-09-20 07-51-21.png

Comments to this report:

fiori - 14:41 Saturday 24 September 2022 (57182)

A peak at 157 Hz is mentioned in elog 43978 (O3 commissioning) and disappeared when closing SDB1 FI shutter. That peak was thought to belong to SIB1 (maybe a bench mode? see for example: 48615 ).

direnzo - 11:55 Thursday 29 September 2022 (57236)

To characterize this noise also from the point of view of glitches, I attach three plots of the time-frequency distribution of Omicron glitch triggers in V1: DAQ_FMOD_56MHz_row_mag_FS in the considered time interval.

Fig. 1 is the classical "glitchgram", where every dot represents the central frequency of the corresponding glitch trigger identified by Omicron. The color and the size represent its SNR.

Fig. 2 gives an alternative picture of the previous glitchgram, where the rectangles represent the duration and bandwidth of each glitch.

Fig. 3 glitchgram provides additional information on the class of the various triggers identified by Omicron. The classification has been obtained as follows. Scattered Light glitchs in V1: DAQ_FMOD_56MHz_row_mag_FS look, in the spectrograms, "approximately similar" to those in hrec. So, I used a pre-trained classifier, trained on O3 hrec Virgo glitches, to classify V1: DAQ_FMOD_56MHz_row_mag_FS triggers in this time interval. These are shown as red circles. The classification of the other glitch classes is not as accurate (for many hrec classes there is no equivalent in DAG_FMOD_56MHz) so I have collectively labelled them as "Others".

Ideas about further analyses: other time intervals can be inspected. The rate of Scattered Light glitches can be correlated with auxiliary sensors, such as seismometers, as done in the ExtENV paper.

Images attached to this comment