Figure 5: (a) The image of a single star with the focus pyramid introduced into the beam. Possible errors in the center-of-gravity determination of the four image centers do not hamper the focussing accuracy due to the high sensitivity of the subimages’ distances on the focus position. (b) The calibration curve used to convert the pyramid's subimage distances to the focus applied. (a) shows the measured FWHM of a star without the pyramid in the optical beam as a function of focus position (distance of the secondary mirror to M1). (b) shows the average distance of the four subimages in pixel with the pyramid inside the optical beam. This calibration has been done at a very early time in the commissioning of the telescope, when the telescope control system was still unstable in attaining certain focus positions. This is visible as the sometimes huge error bars in the focus position.