KAGRA Logbook

MIF (General)

satoru.takano - 20:33 Tuesday 16 July 2024 (30436)

OMC Commissioning Day1: Reduction of the floor noise

Komori (remote), Takano, Tanaka, Ushiba

Abstract

The white noise in OMC TRANS above 200 Hz seems to be introduced by detuning of the cavity. By adjusting the input offset of CARM servo, the white noise was well suppressed. Now we can see many structures below 600 Hz.

Detail

We started OMC commissioning. The first target is to improve the sensitivity reported here, which is limited by some unknown noise floor.

As reported in this log, the noise floor seems to come from some coupling of the control, such as OMC LSC, OMC ASC, CARM, etc. Inspired by this result, we tried changing the locking point of X arm by adding some offset at the input of the CARM servo, then we found that the floor level changes drastically, depending on the offset value, This implies that the locking point of X arm is somehow off resonance due to the residual RFAM.

Figure1 shows the spectra of OMC TRANS normalized by DC power of 27.6 mW with various offset in CARM servo. So far the best sensitivity is achieved with the offset of 14.5. Figure2 shows the comparison with ISS out of loop/ in loop signal and the input signal of CARM servo. Above 600 Hz, OMC TRANS signal is better than ISS signal, but we are not sure the reason, Perhaps we missed some filter in OMC TRANS or ISS OOL/IL. Above 2 kHz, we can see a tiny coherence ~ 0.15 between OMC TRANS and ISS signals.

Thanks to the reduction of the white noise, now it is obvious that below 600 Hz there are many structures in OMC TRANS. Comparing with the sensitivity at O4a, the noise level is still higher (for examle, at 100 Hz the current sensitivity is about 10 times larger). It could be that they come from the acoustic coupling. Another candidate is the control noise of the suspensions. We will investigate these possibility by acoustic injection and noise projection of the suspension controls.

Images attached to this report

Comments to this report:

kentaro.komori - 0:05 Wednesday 17 July 2024 (30438)

I checked coherence between the OMS trans DC and other interesting channels such as IMMT1 trans QPD PIT/YAW and OMC QPD PIT/YAW.
As shown in the attached figure, we can see strong coherence in OMC QPD PIT at a few Hz and around 10 Hz, but the coherence is almost zero in any channel above 100 Hz except for 130 Hz peak in IMMT1 trans QPD.
Therefore, it is unlikely that feedback of dark noise of OMC QPD signals causes noises above 100 Hz.

The next step is reduction of the noises above 100 Hz to perform better acoustic injection measurements.
If the mechanical structures originate from the input jitter, it should appear at IMMT1 QPD channel as well as the 130-Hz peak.
Thus, the input jitter should not be the noise sources above 100 Hz.

One possible explanation is that the OMC QPDs have poor sensitivities at high frequencies, so the actual jitter noise around OMC is completely hidden by the dark noise.
In that case, we might be able to reduce the jitter noise by better alignment of OMMT and OSTM to the OMC.
The beacon ASC might be helpful for the alignment.
In addition, tapping tests around the OMC will provide some hints to determine the origin of those noises.

Images attached to this comment

tomotada.akutsu - 8:57 Wednesday 17 July 2024 (30439)

> except for 130 Hz peak in IMMT1 trans QPD.

Is this the one discussed in klog24703? We did some treatment before closing the IFI chamber (29334, 29407) but unfortunately, seemingly this peak has not been well cropped... Hopefully the better alignment and proper common-mode rejection might help in the future. Dedicated narrow-footprint suspensions for STM1 and 2 would be wished, although I am well not sure what would be the main cause of this peak even found in IMMT1 QPDs.

Abstract

Detail

Next