Performance of MEMS-based visible-light adaptive optics at Lick Observatory: closed- and open-loop control

Abstract

At the University of California's Lick Observatory, we have implemented an on-sky testbed for next-generation adaptive optics (AO) technologies. The Visible-Light Laser Guidestar Experiments instrument (ViLLaGEs) includes visible-light AO, a micro-electro-mechanical-systems (MEMS) deformable mirror, and open-loop control of said MEMS on the 1-meter Nickel telescope at Mt. Hamilton. (Open-loop in this sense refers to the MEMS being separated optically from the wavefront sensing path; the MEMS is still included in the control loop.) Future upgrades include predictive control with wind estimation and pyramid wavefront sensing. Our unique optical layout allows the wavefronts along the open- and closed-loop paths to be measured simultaneously, facilitating comparison between the two control methods. In this paper we evaluate the performance of ViLLaGEs in openand closed-loop control, finding that both control methods give equivalent Strehl ratios of up to ~ 7% in I-band and similar rejection of temporal power. Therefore, we find that open-loop control of MEMS on-sky is as effective as closed-loop control. Furthermore, after operating the system for three years, we find MEMS technology to function well in the observatory environment. We construct an error budget for the system, accounting for 130 nm of wavefront error out of 190 nm error in the science-camera PSFs. We find that the dominant known term is internal static error, and that the known contributions to the error budget from open-loop control (MEMS model, position repeatability, hysteresis, and WFS linearity) are negligible.