I measured noise spectra of the three error signals (REFL 45, 56, and 17) directly reflected from PRM to estimate the sensing noise limit of these channels, particularly that of CARM.
Ideally, the sensing noise is completely limited by the shot noise.

The attached figure shows the result with some light.
In addition, it is compared to that without light to measure the electrical dark noise.

At high frequencies, the dark noise remains a dominant factor, which should be the shot noise.
We need to increase the input power to the PDs.
There are large excesses at low frequencies compared to the dark noise.
It might be because the optics inside the cahmber around IFI are still in air, so I will perform the same measurement after vacuum pumping.
Another potential reason is the amplitude modulation caused by detuning of the RF sidebands.
I will check the spectra after the fine tuning again, as described in klog:29385.

Furthermore, even-number harmonics of 60 Hz are prominent in the spectra with light.
Since there are no harmonics in the dark noise measurement, it should be generated by the IMC control or unknown electrical coupling peculiar to the configuration with light.

After vacuum pumping, I will calculate the CARM sensing noise limit in the unit of Hz/Hz and the projection on the DARM sensitivity.