[Tanaka, Ushiba]
Abstract:
We checked the effect of dewhitening filters for TM stage of Type-A suspensions.
Though noise floor is lower than the current sensitivity, it is better to increase the dewhitening filter stage from 1 to 2, espectillaty for ITMX, ITMY, and ETMY. because several peaks in DARM sensitivity is coming from the peak in DAC noise.
Detail:
Since DAC noise was projected by using the measured DAC noise with zero outputs but DAC noise with some outputs would be larger, we checked DARM spectrum with several conditions.
Figure 1 shows the spectrum with several dewhitening filter configurations of Type-A TM stage.
Blue, green, brown, and pink lines show the DARM spectra without dewhitening filters of ETMX, ITMX, ETMY, and ITMY, respectively (2-stage dewhitening filters are ON for the other suspensions).
Red line shows the DARM spectrum with 2-stage dewhitening filters for all suspensions.
According to the results, noise floor of DAC noise is about 1-2 times 10^-18 m/rtHz @ 100Hz without dewhitening filter, so to reach less than 10^-19 m/rtHz, we need to engage 2-stage dewhitening filters for all suspensions.
In addition, there are very large peak in DAC noise, especially for ETMY, which can be close to the current sensitivity even engaging 2-stage dewhitening filter.
However, ETMX TM coil is close to saturation when engaging 2-stage dewhitening filter (fig2), so to keep the stable lock, it is necessary to reduce the RMS of DARM LSC feedback.
According to the spectrrum of feedback signals (fig3), dominant frequency limiting the RMS of feedback signals is 2-2.5 Hz and 7.5 Hz, so reducing the noise around these frequencies are important.
Also, using both ETMX and ETMY is another solusion to reduce the RMS of LSC feedback to each suspension.
At this moment, ETMX DAC noise seems smaller than current sensitivity foretunately, so I changed the LSC_LOCK guardian so that 1-stage dewhitening filter for ETMX and 2-stage dewhitening filters for the other suspensions are engaged at LOW_NOISING_ETM state.
I also midified the LSC_LOCK guardian so that IMH and MNH additional dewhitening filters are turned on at LOW_NOISING_ETM state.
Note:
Thanks to dewhitening filter engagement, moving peak around 70 Hz was disppeared, which seems coming from ETMY and ITMX DAC noise.