Review of AO calibrations, or how to best educate your AO system

If the Real-Time Computer is the heart of an AO system, the Wavefront Sensor (WFS) its eyes, the Deformable Mirror (DM) its hands and the control strategy its nervous system, the sum of all those parts is made into a harmonious entity thanks to calibrations. This paper does not have the ambition to provide an overview of all the currently existing calibration strategies, but rather to focus on a few challenging problems and their recent evolution in the era of adaptive telescopes, mostly based on the experience of ESO's Adaptive Optics Instruments in general and the AO Facility in particular. Single most important calibration in post-focal AO system, the recording of the Interaction Matrix (IM) between WFS and DM has since long evolved to use fast modulation techniques, has shown to be feasible on-sky and is now almost free from measurements thanks to its pseudo-synthetic generation, quasi-mandatory solution in an adaptive telescope. Pseudo- because it requires an unprecedented knowledge of the components' characteristics, especially the WFS, DM and the optical registration between the two. Bigger telescopes and the use of Laser Guide Stars (LGS) also mean that the properties of the system will change in time and thus need to be constantly updated thanks to online diagnosis tools for spot size measurement, atmosphere monitoring, Wavefront Sensing and control optimization. New loops come into play like the one to minimize LGS Jitter and the one taking over the telescope active optics by means of offloading the DM low orders, and they all require calibration. More calibration means more time and one has to carefully balance the calibrations that require precious telescope night time, day time or for the best, no telescope time at all. Their importance sometimes underestimated, calibrations have repeatedly shown to be a vital part in the optimum functioning of present and future AO systems.