This paper is concerned with experiments which measure CMB anisotropies on
small angular scales. A certain coverage, a beam structure and a level of u
ncorrelated noise define each experiment. We focus our attention on the rev
ersion of the beam average. In each experiment, we look for the best pixeli
zation for reversion, namely, for the pixelization that - after reversion -
leads to good maps containing right spectra for the most wide range of ang
ular scales. Squared pixels having different sizes "smaller" than the beam
radius (theta(FWHM)) are considered. For a given size, the following questi
on arises: How well can we assign a temperature to each pixel? Various math
ematical methods are used to show that, in practice, this assignation - bea
m reversion or deconvolution - only leads to right spectra for pixel sizes
greater than a certain lower limit close to theta(FWHM)/2. This limit is es
timated for negligible and relevant levels of noise and also for sphericall
y symmetric and asymmetric beams. After this general study, we focus our at
tention on two feasible detectors (which have been proposed to be on board
of PLANCK satellite). For each of them, we estimate the size of the most ap
propriate pixelization compatible with beam reversion, diffraction, observa
tional strategy et cetera and, then, we answer the following question: Whic
h is the part of the angular power spectrum which can be extracted from app
ropriately pixelized maps after deconvolution? (C) 2000 Elsevier Science B.
V. All rights reserved.