KAGRA Logbook

MIF (Noise Budget)

kiwamu.izumi - 19:38 Friday 27 April 2018 (4868)

Improved sensitivity in 10-20 Hz band

Today we managed to reduce the Michelson noise in 10 -20 Hz band by a factor of 8 at most.

Noise in this band was identified to be an angle-to-length coupling of the BS even though it didn't show high coherence earlier. The BS has the optical lever controls fot both pitch and yaw with the control bandwidth of 3 Hz for both loops (4848). As they feed their sensing noises back to the BS they actually couple to the length via the combination of the beam spot position and the actuation nodes for the angular degrees of freedom. We reduced this noise coupling by minimizing the injected line in pitch and yaw at around 30 Hz and by changing the output matrix of the TM stage. The contributions from pitch and yaw were found to be equally bad.

On the other hand, because the coupling depends on the beam spot position on the BS, the noise in this band seems non stationary and we have no idea how long we can keep using the tuned output matrix. We will see.

Comments to this report:

kiwamu.izumi - 8:02 Saturday 28 April 2018 (4873)

Later, looking at the spectra of the BS optical lever signals, we see that the optical lever signals above 10 Hz are seemingly dominated by ADC noise.

Currently the sum for optical lever QPDs are about 1200 counts and 1700 counts for LEN and TILT, respectively i.e., the individual segments receive a few 100 counts only. We are hoping that the QPDs had been in the low gain mode so that we can increase the signal-to-noise ratio by simply switching them to the high gain mode.

Images attached to this comment

tomotada.akutsu - 15:57 Saturday 28 April 2018 (4878)

About the BS oplevs, I found that the count was around 2000-1000 cnts from the first; see here. If this is not because the light source itself, this is due to the reflectivity of the BS.

One option is to rotate the collimator so that the polarization of the light beam can be better (or not??)

Or check the coating of the BS and order another light source with different wavelength.