After an entrance window limiting the energy that enters the instrument to 5%, an image of the Sun (shown here in yellow) is obtained on the CCD detector using a Cassegrain-type telescope. Solar photons are selected spectrally using interference filters on a mobile mechanism. A shutter (not shown) defines the exposure time. To control the instrument’s stability, a system of four prisms is placed behind the entrance plate and delivers four secondary images to each corner of the detector. A point on the limb of this image and the corresponding point in the secondary image are separated on the CCD by a number of pixels that depends only on the prism angle and temperature. An interference filter prevents chromatic effects. Part of the corresponding beam is intercepted and placed on four diodes that centre the image of the Sun on the CCD by adjusting the position of the primary mirror.