Virgo Logbook

A few things that we have learned this week

A change in the state of the interferometer started on Monday Oct 20 around 6:20 UTC https://logbook.virgo-gw.eu/virgo/?r=67996
That change resulted in a decrease in the sideband recylcing gain, and an unbalance of the 56MHz on B1p, with the USB decreasing in power and the LSB increasing in power
Increasing the DAS correction of the inner and outer rings of NI and WI by ~20%, and in addition increasing the NI inner ring to reach an increase of a factor 2. Allows to restore the sideband gain, the sideband balancing. The dark fringe is then at ~0.03mW on B1p PD2, which low enough to attempt locks to LN3, but not in a reliable way as on occasion the power crosses the 0.04mW threshold that triggers the detection system safety.
Instead of acting on the DAS, reducing the PR CHROCC by 1V, which reduce the temperature by ~20 degrees (figure 1) also a manages to restore a good sideband gain, and sideband balancing. But the dark fringe is brighter than when acting on the DAS, around 0.04mW on B1p PD2. Too high to even try to lock in LN3. Steps of 0.1V are too small to have a noticeable effect, so larger steps are needed if one wants to do a tuning using the PR CHROCC.
The dark fringe power is affected by the balancing of the sidebands when the misbalance is very large. In the sense that adding a MICH offset which restores the sideband balance, but does not change the sideband gain, also reduces noticeably the B1p power. For example on Monday Oct 20 around 17:30 UTC from 0.04mW to 0.03mW.

The 1V change in PR CHROCC correction correspond to a ~15m change in effective PR radius of curvature, according to the CHROCC calibration measurements (VIR-0830A-22), taking into account that the analog electronics gain has been reduced by a factor 9.4 since those calibration measurements. So currently 1V of chang correspond to 0.1V of change in the presentation from 2022. 15m of change in radius of curvature correspond to ~7e-6 diopters, and from what I have been told the 20% increase in DAS correction also corresponds to ~7e-6 diopters. So we can restore the sideband gain and balancing by putting a negative lens on the CP or on PR, which is the expected results as the sidebands, especially the 6MHz, do not care at which location inside the PRC the lens is added.

The changes in PR CHROCC temperature as a function of voltage look comparable to past tests in 2023 (VIR-1079A-23). And given that a 1V change was needed to restore a situation similar to normal it is unlikely a change in behavior of the PR CHROCC is the cause of the issue. All the DAS and CH CO2 laser powers have been checked and appear normal, so it makes also unlikely that a change in DAS is at the origin of the issue. This means we need to expand the list of possible origins of the problem:

The absorption of one or both of the input mirrors could have changed for some reason. If it is on the HR side it should be visible in the position of the order 2 mode in LN2 on B1 which is at ~10.6kHz. The best way to see it is to look at the transfer function between SSFS error signal and B1 in LN2 (avaiable for ~1 minute at each lock acquisition)
The input beam mode matching has changed, for example due to some falling dust getting burned into an optic. Making a measurement of the input beam mode matching to the arms could check that hypothesis.

However it is not necessary to understand the origin of the problem to attempt to fix it. Looking back at the various tuning steps of this week, the best I could find in terms of dark fringe was on Wednesday morning (https://logbook.virgo-gw.eu/virgo/?r=68015), before any of the tuning steps that were done on that day. The power on B1p PD2 at ~10 UTC was 0.028mW in CARM NULL 1F, increasing to 0.033mW when the dark fringe offset is added. The issue at the time was that the sideband gain was low, increasing the NI and WI DAS correction had managed to increase the sideband gain, but it probably degraded the dark fringe, it is harder to judge as that tuning was done in LN3, the power in LN3 on B1p did increase, but we cannot be certain if that would have also been the case in CARM NULL 1F.

Hence, the TCS actuation state from Wed Oct 22 in the early morning could be a good starting point to restart the TCS actuation tuning. And then do the tuning in CARM NULL 1F, trying to improve sideband powers without degrading B1p, using some combination of PR CHROCC and the two DAS. One thing to keep in mind is the weather, we do expect B1p to increase when the weather is bad due to larger motion of all of the suspensions, so that can be biasing our analysis when comparing the interferometer state between days with very different weather.