Design issues for the active control system of the California Extremely Large Telescope (CELT)

We explore issues in the control and alignment of the primary mirror of the proposed 30 meter California Extremely Large Telescope and other very large telescopes with segmented primaries (consisting of 1000 or more segments). We show that as the number of segments increases, the noise in the telescope active control system (ACS) increases, roughly as root n. This likely means that, for a thousand segment telescope like CELT, Keck-style capacitive sensors will not be able to adequately monitor the lowest spatial frequency degrees of freedom of the primary mirror, and will therefore have to be supplemented by a Shack-Hartmann-type wavefront sensor. However, in the case of segment phasing, which is governed by a "control matrix" similar to that of the ACS, the corresponding noise is virtually independent of n. It follows that reasonably straightforward extensions of current techniques should be adequate to phase the extremely large telescopes of the future.