HAM2-ISI features an unusual high Q resonance at 96.1Hz. This resonance is due to the top payload mass that still needs better boundary conditions with the ISI (see aLog comment #4025).
The seismic team uses generic controllers for the first level of the HAM-ISI controls. The watchdogs of HAM2-ISI would trip on the actuators' signal when turning on those controllers. Indeed, the generic controllers did not account for such feature and the unusual resonance would consume most of the gain margin (Simulation Plots - No Notches), causing the controllers to amplify motion at 96.1Hz (see p.4-9 of the attached pdf), and thus to only be stable on their own (see p.2 of the attached pdf).
In order to get HAM2-ISI under active-control, we decided to apply temporary notches on the level-1 controllers of HAM2-ISI. We started with the most unstable controllers: X and RY.
Simulation Plots – Notches on X and RY
After implementing the new X and RY isolation loops, we were able to have all HAM2-ISI controllers running at the same time. We took comparative spectra for the various configurations of the ISI: not damped, damped and controlled with the new controllers. They showed that the motion of the ISI was still amplified at 96.1Hz (Image #1).
In oder to correct that, we went through all the other controllers of HAM2-ISI and applied proper notches to them. The only one that did not need any was RZ.
Simulation Plots – Notches on all DOF
We then got rid of the excessive motion amplification at 96.1Hz (Image #2).
HAM2-ISI is now under control Level 1 with the generic 900mHz blend filters ON.
Note 1: HAM2-ISI is loaded with dummy masses only. This control work was made in anticipation of the future Control & Commissioning of the unit, and in preparation of the Control & Commissioning that will soon be performed on HAM3-ISI.
Note 2: The iquiery described above involved looking into te details of the HAM-ISI generic routines, and revealed the need for minor corrections. Routine #7 was corrected twice: SVN revisions 6028 and 6077
