Today, I worked on the improvement of the performance of the blending strategy at EX following what I did at EY (see klog )

Below are some details of the work.

As the first step, I estimated the noise of the LVDT, seismometer, and geophone as sensors involved in the sensor fusion.

Pic1, Pic2, and Pic3 show the estimated noise of each sensor mentioned above.

I request the Align state to Guardian. Keeping this state, I measured the spectra of the IP, BF, and TM when the following control schemes were ON:

- IP LVDT control;

-IP LVDTSC control;

-IP blending at 80 mHz along L and T directions Blending at 150 mHz along the Y direction.

Pic4, Pic5, and Pic6 show the spectra of the IP compared to the CL noise and rms. Also here, It is possible to see that by changing from LVDT control to ID control the rms, and the re-injection of the seismic noise via LVDT are reduced, but there are still some rooms to improve the situation by reshaping the blending and control filters.

As done for EY, I reshaped the 80 mHz blending filters so to reduce a bit more the re-injection of the seismic noise in the region [0.12, 0.350] in L and T, and I moved the blending frequency from 159 mHz to 110 mHz in Y. I closed the loops with the same UGF and measured the spectra again. And as the last step, I reshaped the control filters along L and T and increased the UGF at 1 Hz.

Pic.7 and Pic 8 show the IP spectra in these new configurations. The rms along the 3 DOF are reduced.

In addition to this, I measured the spectra of H4 (out-of-loop sensor on the IP), BF (L, T), and TM (L, P, Y) in the same configurations.

Pic9 shows the spectra of IP acquired by the out-of-loop sensor H4. Here It is possible to see that changing from LVDT control to ID 80 mHz new and UGF 1 Hz, the suppression of the micro-seismic noise is about a factor of 13 in terms of rms.

Pic10 and Pic11 show the BF spectra along L and T directions. Also here it is possible to see the benefits of the new strategy: a factor of 1.8 of reduction in terms of rms in the frequency region [0.120-0.350]

Pic12, Pic13, and Pic14 show the TM spectra along L, P, and Y directions. No evidence of re-injection of geophone noise below 0.1 Hz has been detected.