As often reported by the operators, the earthquakes use to unlock the ITF because they are able to trigger a DARM oscillation at 9.7 Hz (fig 1). At this frequency there is the well known problem of the roll resonance of the payloads, for which a digital damper is active. What is not well understood about that, is the fact that WE damper sometimes become unstable. It used to do like that at the beginning of O3, and for that reason it was prudentially disabled, until a fine tuning of the controller made the issue disappear. Then a similar issue appeared in coincidence with the earthquakes, but checking the WE damper, it looked fine. Finally, just the past week I was certain that WE roll was the culprit, because I had the opportunity to disable the damper during an earthquake, and the 9.7 oscillation was suddenly stopped.

At the same time, I realized that the oscillation was a side effect of the vertical slow oscillation of the suspensions (normally produced by the seismic wave in that condition), and in particular it was well synchronized with OB motion (fig 2). This is not too much surprising, because SDB1 motion produces a fake signal on the quadrant used to control COMMp, so the beam in the cavities starts to move a lot, and this fact could modify the roll to DARM coupling (surprising, but not too much).

Investigating more, I found that OB vertical control was quite bad because of a mistake in the blanding strategy for the Inertial Damping (it was like that also when the strange effect on the roll stability was not there, and the ITF was more resistant to the earthquakes).

I fixed the mistake, and I took the opportunity to improve the vertical control strategy, implementing a GIPC for SR/OB.

Yesterday there was the opportunity to check the effectiveness of the action, thanks to an earthquake. In fig 3 we can see that, for a similar vertical oscillation of SDB1 (plot 1,1) the SR/SDB1 relative motion is much smaller (plot 1,2), and as a consequence the COMMp_TX (plot 2,1) and SDB1 pitch control (plot 2,2) are much more stable.

Then, the ITF unlocked anyway, not for the roll oscillation. But this is a topic for the next entry.