The output from the correlator is integrated for some period, 
,
and then assigned the mean time for that integration period. For a source at a celestial
pole, this would amount to a rotation of a visibility through the angle
, where 
is the earth's angular velocity
(
rad s
) and 
is the timing error
in the assignment of u and v. This would cause the image to be rotated
through the same angle. For a pole source then, this error is the same
as averaging a series of images aligned at the phase centre, but with angular
offsets up to 
. The final image would thus
reflect an azimuthal distortion which is more severe with increasing
distance from the phase centre. This is in some sense complementary
to bandwidth smearing which produces a radial distortion. The general case
is of course more complicated than this. A useful result that is valid
for a variety of source locations and array types is very similar to that
of bandwidth smearing; the same figure can be used for the decrease in
amplitude of a point source with distance from the phase centre, with a
change of abcissa, as shown in the figure.
Last update : 27/11/93
1
Last update : 10/10/92