Reports 1-1 of 1 Clear search Modify search
AdV-COM (AdV commissioning (1st part) )
swinkels, bersanetti, allocca - 1:23 Saturday 22 April 2017 (37334) Print this report
More alpha tuning
We continued yesterday's work of tuning the MICH noise subtraction. We further improved the filters for shaping the MICH and DARM noise, so that the measurements give good coherence down to 10 Hz, where it is needed. After that, we were able to do the actual noise injections via the injection script prepared by Franco, see /virgoDev/Automation/scripts/LSC/inject_lsc.py. Logfiles with the gps times of the injections are stored in /virgoData/NoiseInjections/LSC.

Using these injections, we could finally do a reasonably good measurement of alpha. Fig 1 shows a comparison between alpha filter on and off while injecting noise on MICH, while fig 2 shows the same in quiet conditions. Fig 3 shows a noise budget with the filter turned on (to be compared with the same plot from yesterday). This plot uses 120 seconds of clean data starting at gps = 1176846162. The noise subtraction is switched on by Metatron at the end of the state LOW_NOISE_1, and is switched off in the DOWN state.

Figure 4 shows that the obtained suppression is about a factor 10 over a broad frequency range. It is likely that the suppression is currently limited by variations in the optical gain. To solve this, we need to servo the gain as in the past.

Some other remarks:
-Especially during the beginning of the shift, the IMC shows a large oscillation after every unlock of the interferometer, which eventually unlocks the IMC, see fig 5. It typically takes about 2 minutes before the system calms down again.

-We thought that we could safely restart the CALnoise to change some filters. Switching the process off causes however some weird glitches which oscillate at 5kHz in some of the suspension signals, see fig 6. This might be a bug in how a missing channel is handled in Acl or in the DSPs.

Images attached to this report
Comments to this report:
mwas - 17:07 Monday 24 April 2017 (37354) Print this report
Made a noise budget (and calibration) for the quiet data with MICH (alpha) subtraction. For now the subtraction is modeled not measured so the noise projection doesn't fit well everywhere. The calibrated noise below 100Hz is well reduce compared to 10 days ago, about a factor 10, with the sensitivity limited by PD electronic (dark) noise from 40Hz, and the lowest point a ~1e-20/rtHz.

This correspond to a range of 145 kpc for BNS and 2 Mpc for 30 solar mass BBH. Still a long way to go, but it will soon start to make sense to look at the range!
Images attached to this comment
Search Help
×

Warning

×