[Ushiba, Tamaki]

Summary

We implemented the notch filter in CRY payload control to suppress the peaks in DARM sensityvity.
The peak at 22.5 Hz was suppressed to some extent, but we need to do the same thing one at a time.

Detail

At first, I compared the peaks shown in DARM sensitivity and the peaks in TFs (from suspension sumout to DARM) to identify to which suspension or DOFs each peak corresponds.
I'm not sure about 16.38, 34.75, 44.50 Hz, but for the other peaks we were able to identify which resonance they were, as summarized in Fig.1.
We used TFs which were measured in klog23334.

Then, we looked at how large the feedback signal was.
As a measure, we estimated how large it was relative to the DAC noise (approximately 1e-4 cnt/rtHz, klog24157 and klog24249) (Fig.2).
Based on the measurement results, we estimated how deep or Q-value notches should be placed and actually implemented them at the corresponding peak frequencies.

For example, the peak at 22.25 Hz was squashed by the notch in EY_L (and EY_{P,Y}) as shown in Fig.3.
In Fig3, red line is DARM sensitivity with new notch filter, and black line is one before install that filter (gray line shows DARM during O3GK).
The peak at 22.5 Hz was suppressed to some extent, but we need to do the same thing one at a time.