I have implemented the sensor correction at HEPI-BSC8. Sensor correction improves the isolation performances of the HEPI in the X, Y and Z directions. In attachment, spectra of the L4C installed in the HEPI boots are presented in different configurations:
- Isolation without sensor correction (blue)
- Isolation with sensor correction (red)
There is no plots in the uncontrolled configuration (Robert S wants to keep the cavity locked).
Ugh -- no units on the vertical scales. Noise at 1 Hz goes down from ~100 to ~10; wish I knew what that means.
The spectra are measured before blending the L4Cs with the position sensors. The units are (nm/s)/sqrt(Hz) above 1Hz. To get the calibration in nm, you need to invert the idealized L4C (3 zeros at 0Hz and a pair of complex conjugate poles at 1Hz with a 45deg phase).
Even if there is no plots of the uncontrolled HEPI, the isolation provided by the controller without the sensor correction is visible in the blue curve (“V shape feature”). The isolation is defined by the blend filters (LP) of the position sensors (first segment of the “V” – Blend at 800mHz) and the super sensor suppression (second segment of the “V” – UGF at 10Hz).
When the STS-2 (ground) signal is added to the position sensors (sensor correction), it is like having a seismometer (STS-2) in the HEPI boot (TF from ground to top of the pier close to 1 below 10Hz). Now, the super sensor is an inertial sensor above 40mHz. With the sensor correction (red curves), the actual suppression follows the super sensor suppression. Consequently, the isolation improvement is visible in the 40mHz to 3 Hz frequency band.