KAGRA Logbook

MIF (General)

kentaro.komori - 23:05 Monday 14 October 2024 (31320)

First measurement of contrast defect in KAGRA

[Sugioka, YamamotoT, Komori]

We measured the contrast defect of KAGRA PRFPMI for the first time.
The measured value was 0.13 mW with an input power of 1.25 W, which is roughly consistent with Haoyu's simulation.

We created a plot of the OMC transmitted power versus the optical gain (W/m), following the LIGO method.
Further details will be reported later.

We succeeded in changing the DARM offset power from 20 mW to 3 mW.
The attached figure shows the sensitivity with different DARM offset power.
During this measurement, we measured K1:LSC-DARM1_IN1 with varying the input offset of K1:LSC-OMC_DC and compensating it by adjusting the gain at K1:LSC-OMC_DC by a factor of square root of relative power difference.

The sensitivity did not change significantly with different offset powers, although it slightly worsened at lower powers, likely due to an increased contribution from PD dark noise.
However, this issue should be resolved with higher input power to the IFO in the future.

In the most sensitive frequency range (60 Hz to 100 Hz), we observed breathing in the signal, which became more pronounced at 3 mW.
The interferometer (IFO) lock was lost within several minutes at 3 mW, but the lock was maintained for over an hour at 5 mW.

These measurements suggest the following:

It is likely that the current dominant noise source is not OMC acoustic noise but actual displacement noise, because OMC acoustic noise would improve with the square root of reduced DARM offset power.
The effect of alignment fluctuations (~mW) is much larger than the contrast defect (~0.1 mW), indicating that we could further reduce the DARM offset power by suppressing alignment fluctuation.
For stable operation at an IFO input power of 10 W with the same DARM offset displacement, we would need a DARM offset power of 40 mW, assuming no further suppression of alignment fluctuation. This power might be too high.

Images attached to this report

31320_1728914216_Screenshot from 2024-10-14 21-50-40.png

Comments to this report:

tomotada.akutsu - 23:24 Monday 14 October 2024 (31321)

> For stable operation at an IFO input power of 10 W with the same DARM offset displacement, we would need a DARM offset power of 40 mW...

Maybe, can we expect ASC improvement by increasing the input power as well?

kentaro.komori - 23:35 Monday 14 October 2024 (31322)

We left the IFO with the DARM offset power set to 5 mW and the gain of K1:LSC-OMC_DC set to 1.8, in orser to monitor the stability of the IFO.
While the offset power setting will be reset after a lock loss, the gain will not.
Therefore, when recovering the IFO, please reset the gain to unity.

kentaro.komori - 9:16 Tuesday 15 October 2024 (31325)

I reset the gain of K1:LSC-OMC_DC to unity.

Two lock loss occurred during this night with 5 mW DARM offset and the gain of K1:LSC-OMC_DC of 1.8.
The lock durations were five and three hours.
The IFO lock was recovered even with the higher DARM loop gain, but the sensitivity got worse probably because the DARM control was almost oscillating.

Tatsuya Sugioka - 0:02 Wednesday 16 October 2024 (31330)

Since the SNR in DC readout is proportional to the inverse of the square root of the DARM offset power (i.e., the OMC transmitted power), the smaller the offset power, the higher the SNR—provided the IFO lock is maintained. However, due to interferometer asymmetries, such as birefringence, the offset power has nonzero value even on perfect dark fringe; contrast defect. In principle, this defect limits how low the offset power can be. Near the dark fringe, OMC transmitted power is quadratically dependent on optical gain with contrast defect offset.

Now we have estimated the contrast defect and the value is 0.13 mW which is roughly consistent with Haoyu-san’s birefringence simulation.

We measured DARM LSC loop gain with various offset power. Offset changes are compensated by adjusting the gain at K1:LSC-OMC_DC to keep it locked, so the gain is roughly constant.

Fig1: K1:LSC-DARM1_IN1 for each DARM offset power

Optical gain is proportional to K1:LSC-DARM1_IN1/ K1:LSC-OMC_DC. We focused on 86.8 Hz and created a plot of the OMC transmitted power versus the optical gain (W/m).

Fig2: plot of the DARM offset(transmitted power) versus the optical gain(normalized to the value when offset power is 15 mW) @86.8 Hz, and its quadratic curve fitting
Fig3: enlargement of Fig2

In Fig2, the fitting curve shows contrast defect is ~ 0.13 mW.
Additionally, IFO lock is stable with 5 mW offset. This suggests that it is possible to keep the current SNR (with 15 mW offset) if the input laser power is tripled.

Images attached to this comment