A Giant Step: from Milli- to Micro-arcsecond Astrometry Proceedings IAU Symposium No. 248, 2007 W.J. Jin, I. Platais & M.A.C. Perryman, eds. c  2008 International Astronomical Union DOI: 00.0000/X000000000000000X Design of a compact astrometric instrument for the GAME mission D. Loreggia, M. Gai, A. Vecchiato, D. Gardiol, S. Ligori and M.G. Lattanzi All Authors are from INAF - Astronomical Observatory of Turin, St. Osservatorio, 20, 10025 Pino Torinese, Torino, Italy. email: loreggia@oato.inaf.it Abstract. We present the design of a Fizeau interferometer to be implemented for the GAME mission. The aim is to measure the PPN γ parameter with the same technique used for the first time by Dyson, Eddington et al., but at a 10 -6 accuracy level. GAME will operate in a continuous scanning mode observing about 10 6 sufficiently bright stars at about 2 ◦ from the Sun. A dedicated space mission has the advantage of observing the light bending without waiting for an eclipse. Keywords. gravitational lensing, space vehicles, instrumentation: interferometers, techniques: high angular resolution 1. Introduction The Gamma Astrometric Mission Experiment - GAME (Vecchiato, et al.2008) will observe in a continuous scanning mode two symmetric fields, each about 2 ◦ from the Sun. The relative position of the targets is measured again 6 months later, without the Sun in between, giving a direct estimation of the light bending and the γ PPN parameter with an expected accuracy of 10 -6 . Hereafter we briefly discuss a possible telescope design, preserving some requirements for a small mission. 2. Optical design A 5-aperture Fizeau interferometer design is well suited for our propose of merging good performance with some geometrical constraints such as a distance between the main mirrors not larger that 1.5m, the size of the pupil compatible with diffraction limited (DL) images and with the size and the geometry of the interferometric apertures. These requirements must match the best sampling condition of the fringes (Loreggia, et al.), assuming a pixel size of 15μm in the high resolution direction (HRD) and asking for a baseline between each couple of apertures B=2L, where L the aperture side in HRD. The aperture area dimensioning sets directly all the instrument parameters and is driven by the photons collecting needs and the final encumber limitations. At λ ref =650nm, the best solution is obtained with apertures of 40x220mm over a pupil of 660mm, giving an EFL = 20600mm and a total baseline (between the outer apertures) B t =320mm, as shown in Fig.1. GAME will look at two lines of sight very close to the Sun, which combination and folding is implemented by mean of a beam combiner in front of the primary mirror. A baffling array is needed and is under study; we tested that looking at about 2 ◦ from the Sun, a typical baffling length may be ∼4.2m, which can be achieved with the folded configuration presented. 274