Virgo Logbook

AdV-DET (Photodiodes + demodulation boards (longitudinal control))

mwas - 12:31 Saturday 11 October 2025 (67915)

B1 PD2 noise

I have tried to close B1 PD2 shutter with the interferometer locked to investigate the extra noise on B1 PD2. However there was an unlock a few seconds after closing B1 PD2 shutter, I suspect some part of the automation has seen the power on B1 PD2 disappear and requested the unlock. I haven't check if that supposition is true.

Figure 1 shows in blue yesterday data without extra noise, in purple this morning with extra noise, in red this morning in single bounce with several mW on the photodiode (side of carrier TEM00 mode). A similar looking extra noise can be measured in single bounce, and it disappears when there is no light on the photoiode.

06:06:10 (1min) only B1 PD1 in loop

06:07:15 (1min) OMC controlled with B1 PD3

06:08:50 B1 PD2 shutter closed, unlocked after a few seconds
the vbias turned off automatically

06:15 Vbias enabled - no excess noise on photodiodes with shutters closed

06:39 UTC single bounce, both B1 PD opened on the side of the TEM00,
0.2mW of power
06:45:40 fluctuating around a few mW
around 06:51 fast shutter on SDB1 closes a few times because of 6mW on B1 PD1/PD2

07:03:10 (1min) B1 PD1/PD2 shutter closed
07:04:30 (1min) B1 PD1/PD2 shutter open

07:06:20 UTC moving OMC away from resonance

--------------------------------

After relocking attempted to investigate the issue in another way. I have reduced the DARM offset to have 1mW per photodiode instead of 3mW, this reducing the temperature of the photodiode by a few degrees. Which doesn't seem to have a direct clear impact. What seem to have an impact is that the level of noise decreases with the power on the photodiode. So at the end the BNS range improves, even though it gets only up to ~50Mpc instead of 55Mpc. Still better than the 48-49MPc measured with only one photodiode in loop.

Figure 2 shows the noise level decrease on the PD difference from the power reduction. At high frequency there should be a factor sqrt(3)=1.7 reduction due to lower shot noise at low power, at lower frequency the decrease is faster, or the order of proportional to the power, so that corresponds to an improvement in noise level projected on the sensitivity.

Leaving the interferometer with this lower DARM offset.

Images attached to this report

67915_1760167202_Screenshot from 2025-10-11 08-19-17.png

67915_1760178566_Screenshot from 2025-10-11 11-23-19.png

Comments to this report:

mwas - 16:13 Saturday 11 October 2025 (67917)

Figure 1 the BNS range has been fluctuating between 51Mpc and 54Mpc

Figure 2 this is due to the B1 PD2 noise that has been coming and going. With the smaller DARM offset the BNS range is much less sensitive to the extra noise on B1 PD2.

We should temporarily automate the setting of a smaller DARM offset, until B1 PD2 is replaced next week.

A simple solution could be to add a 5 minute long ramp on the DARM offset at the beginning of acquire LN3, the DARM UGF servo would take care of increasing the DARM gain to compensate, as it has done during the manual DARM offset reduction this morning.

LSC.DARM_SET.AcConstChSet(-0.1925, 300.0)

Images attached to this comment

67917_1760191764_Screenshot from 2025-10-11 15-09-17.png

67917_1760191805_Screenshot from 2025-10-11 15-08-34.png

mwas - 9:09 Monday 13 October 2025 (67924)

Figure 1 shows the trend of the measured B1 NULL rms (PD electronic noise + shot noise). On Oct 11 early in the morning it had increased sharply when the photodiode become noisy again, then around 10:00 UTC there is a decreas in steps due to the decreased in DARM offset (which also lower shot noise in terms of mW/rtHz), and finally around 13:30UTC there is a sharp drop which comes from the noise disappearing from the PD. It has not come back since then during the weekend.

Images attached to this comment

67924_1760339211_Screenshot from 2025-10-13 08-06-15.png

bersanetti - 18:22 Monday 13 October 2025 (67928)

The command for lowering the DARM offset has been added today at the beginning of the commissioning window, it should be necessary only up tomorrow afternoon when it will be removed.