Control algorithm and performance of the active optics system of the Giant Magellan Telescope

The GMT active optics (AcO) control problem is unique because the primary and secondary mirrors are both segmented. This paper describes the AcO control algorithms and assesses their performance for the Natural Seeing observing modes of the GMT. In this case, there are wavefront sensors in four off-axis probes, which are used to control the position and rotation of each of the seven primary and seven secondary mirror segments, as well as the bending modes of the primary mirror. The segmentation of both mirrors leads to a number of unsensed modes (e.g., segment piston and a rotation of the segment around the telescope optical axis) and poorly sensed modes (e.g., M1 segment translations compensated with the corresponding M2 segment translation). Uncertainties in the location of the probes and guide star coordinates also lead to blind modes. In this paper, we first introduce the GMT optical design in the context of the AcO system. The modes of the telescope controlled by the AcO system, and their counterparts in the wavefront sensing space, are presented next, including field-dependent and blind modes. The control architecture is then outlined along with a description of the singular modes of the system. Finally, performance results are provided with respect to various error terms.