Luc Arnold
Formerly with the Laboratoire d 'Astrophysique Observationelle, Collège de France et Observatoire de la Côte d 'Azur, 2130 route de l'Observatoire, 06460 Caussols, France; now with the Laboratoire d'Astrophysique Observationelle, Collge de France, Observatoire de Haute Provence 04870, Saint-Michel-l'Observatoire, France.
e-mail may be addressed to arnold@obs-hp.fr

(Received 25 January 1995; revised manuscript received 31 July 1995.)

Explicit analytical expressions are derived for the elastic deformation of a thin or thick mirror of uniform thickness and with a central hole. Thin plate theory is used to derive the general influence function, caused by uniform and/or discrete loads, for a mirror supported by discrete points. No symmetry considerations of the locations of the points constrain the model. An estimate of the effect of the shear forces is added to the previous pure bending model to take into account the effect of the mirror thickness. Two particular cases of general influence are considered: the actuator influence function and the uniform-load (equivalent to gravity in the case of a thin mirror) influence function for a ring support of k discrete points with k fold symmetry. The influence of the size of the support pads is studied. A method for optimizing an active mirror cell is presented that couples the minimization of the gravity influence function with the optimization of the combined actuator influence functions to fit low-order aberrations. These low-spatial-frequency aberrations can be of elastic or optical origin. In the latter case they are due, for example, to great residual polishing errors corresponding to the soft polishing specifications relaxed for cost reductions. Results show that the correction range of the active cell can thus be noticeably enlarged, compared with an active cell designed as a passive cell, i.e., by minimizing only the deflection under gravitational loading. In the example treated here of the European Southern Observatory's New Technology Telescope I show that the active correction range can be enlarged by ~50% in the case of third-order astigmatic correction.