2 Recent developments relevant for the telescope design

The baseline telescope design for PLANCK is similar to that originally proposed for COBRAS. The primary mirror is illuminated to maximize the angular resolution. An enclosing radiation shield controls the high sidelobe levels far off-axis that are the unavoidable result of the strong edge illumination. The system is a Dragone-Mizuguchi compensated optic and only the centre of the focal surface is aberration-free. The region of the focal surface where aberrations are acceptable from the point of view of angular resolution and beam shape has to be large enough to contain the feeds of a single instrument. With the merger of COBRAS and SAMBA into a single mission (Bersanelli et al. 1996) the demands placed on the telescope increased, because the region of the focal surface with small aberrations had to accommodate the feeds and cryostat from the HFI as well as the feeds from the LFI. During the Phase A study of the merged COBRAS/SAMBA concept the wavefront error at various locations in the focal surface was calculated and found to satisfy standard criteria for optical performances over a region large enough to accomodate feeds from two instruments.

Since the Phase A study there have been several developments with important implications for the telescope design. Detailed designs of the feed arrays in several configurations have been made by both PLANCK instrument teams, taking into account optical, mechanical, and thermal considerations. In the "asymmetric configuration'' of the Focal Plane Unit (FPU) the LFI and HFI are accommodated side by side (Bersanelli et al. 1996). In the "symmetric configuration'', preferred by both teams for mechanical and thermal reasons and recently specified as the baseline (Mandolesi et al. 1998; Puget et al. 1998), the HFI is centered on the optical axis and the LFI feeds are distributed in a ring around it (see Fig. 1), well outside the central region of the focal surface, where the optical performances are particularly critical. The study provided by TICRA (1997) on directivities and beam shapes at all relevant frequencies as well as those based on new optical codes developed by the PLANCK LFI team (Valenziano et al. 1998; Villa et al. 1997, 1998a,b) allowed to assess the impact of beam distortions on the scientific objectives of the mission. New techniques have been developed to evaluate the beam distortions specifically in terms of their effect on measurements of CMB anisotropies rather than by more conventional but less relevant criteria based only on optical properties (Mandolesi et al. 1997; Burigana et al. 1998a,b). This analysis, described in Sects. 3 and 4, shows that the standard rules of thumb for evaluating the performance of optical systems are not stringent enough for the demands of accurate CMB observations. As a meaningful example of the impact of the recent increasing of the theoretical understanding of the importance of angular resolution in CMB anisotropy observations, we observe how the NASA Midex mission MAP has increased the size of its two telescopes from $1.3\times1.5$m to $1.4\times1.6$m in order to improve the goal of its angular resolution from 18' to 12' at 90GHz.

Figure 2: Contour plot of the focal surface of the PHASE A telescope with the positions of the LFI feeds overlaid according to the symmetric configuration. The coordinates are $u\equiv \sin{\Delta\hbox{Elevation}}\approx \Delta\hbox{Elevation}$ and $v\equiv \sin{\Delta\hbox{Azimuth}}\break \approx \Delta\hbox{Azimuth}$

Figure 3: The same as in Fig. 2 but for the BASELINE telescope. Note the difference in plate scale for the two telescopes

Figure 4: The same as in Fig. 2 but for the ENLARGED telescope. Note the difference in plate scale for the two telescopes