Komori, Tanaka

To improve the 2 Hz gain peaking in klog36342, we increased the phase margin around 2 Hz by adding IM actuator in hierarical actuator, that is, we used MN, IM, and TM stages as an actuator of HARD_Y control. Previously, we avoided the 3 Hz resonance pole and zero in MN/TM actuator responce by implementing resG and notch filter. This resG and notch filter reduced the phase margin around 2 Hz. On the other hand, the 3 Hz resonance is IM mode. So we applied IM stage to control the 3 Hz resonance.

First, we measured the response from IM actuator to TM oplev. we compared the IM2TM gain around 3 Hz with the TM2TM gain around 3 Hz and decided the relative gain (-33 dB) between IM2TM and TM2TM so that IM stage works only around 3 Hz. After that, we measured the responce of DHARD_Y_EXC to TM_WIT_Y with the lead filter (zero:1.5 Hz, pole: 30 Hz) and the hierarical actuator. Fig.1 shows the result, solid lines are the responce by MN-IM-TM actuator and dash lines are the responce by MN-TM. Around 3 Hz, the peak height and the phase rotation got small. So the hierarical actuator including IM seems to work well.

Then, we closed {D,C}HARD_Y loops with new actuaors. Fig. 2 shows the OLTF of CHARD_Y, the cyan is the TF with the previous actuaor (MN-TM), the blue is the TF with the new actuator. The phase margin around 2 Hz is improved. Thanks to this, the gain peaking around 2 Hz got small.

Then, we tried to increase the gain to make the bandwidth more wider. We could enlarge the UGF to 4 Hz (red in fig.2). However, the DARM spectra become terrible at that time (fig.3) maybe due to contaminate the control noise... So we need to modify the filter design.