Lucia, Kiwamu,
As suggested by Ushiba-kun, we tried implementing an output matrix that accounts for the fact that one coil is not functioning on the test mass stage.
As a consequence, we now became able to engage Miyamoto-kun's TMP damping control on ETMX.
The first attachment shows the new hacked output matrix. Since the H2 coil is not functioning, we zeroed the corresponding element. We then put -0.5 on the H3 and H4 coils in order to eliminate undesired longitudinal excitation.
With this output matrix, we confirmed that the TM pitch to pitch transfer function now doesn't have an extra negative zero. See the attached below for the measured transfer function.
The red trace is the latest transfer function with the hacked output matrix put in. The black trace is the one from the last week where we were driving mainly the H1 coil only. Observe that a notch at around 0.5 Hz now disappeared presumably due to smaller cross-coupling to longitudinal. Finally the blue trace is the one Miyamoto-kun took in this August. You can see a qualitative agreement between his and our latest measurements. The DC gain was found to be 7 dB lower than his measurement which makes sense considering the fact that we now have one coil acting on the pitch degree of freedom.
The below shows a screenshot of the latest damping setup in the digital system. We didn't change the filters except for the gain filter banck which is now a factor of greater than before for the reason described above. So this is essentially the same damping loop as what Miyamoto-kun had when he achieved the best alignment (5906).
Finally the attached below is the performance of the day with the TM pitch damping engaged. It is clear that we are not limited by the 0.6 or 0.78 Hz peak any more.
