V. Sinceac et al., A NEW METHOD FOR CCD MEASUREMENTS OF THE SOLAR DIAMETER WITH AN ASTROLABE, Astronomy & Astrophysics. Supplement series, 128(3), 1998, pp. 605-615
Observing the Solar disk is a challenge and, as for the past visual ob
servations, we have many results depending on observers and/or instrum
ents. This was due to the differences in visual perceptions of the Sun
's limb, instrumental errors and atmospheric disturbances. After a lon
g series of visual observations at Calern Observatory, Francis Laclare
felt the need for more impersonal and automatic measurements of the S
olar diameter. After a series of analog CCD measurements (1989-1995),
a digital data acquisition and processing was tested by the Paris Obse
rvatory group (F. Chollet and V. Sinceac) during the 1996 spring at Ca
lern Observatory. Before starting a new continuous campaign of observa
tions, to confirm eventual variations of the diameter and solar flatne
ss, the aim is to find the best definition of the solar edge. The test
campaign was spent comparing different solutions that were tried on t
wo different astrolabes at Calern Observatory: The ''classical'' one,
outfitted with eleven zerodur ceramic prisms (S astrolabe), that has b
een used for twenty years in the Laclare series and on the other hand
an instrument equipped with a varying angle prism (V astrolabe) enabli
ng many measurements (385 in 1996) for perfecting the know how. This a
rticle focusses on acquisition techniques and their feasibility. Two p
rocedures were tried: The first one used alternately the direct and re
flected images (separated using a revolving shutter in front of the ob
jective) and the second one mathematically sorts out both components i
nside the computer (an image being a two-dimensional array of numbers)
. According to the principle of the astrolabe, the measured quantity i
s the exact time crossing the parallel of altitude (defined by the pri
sm angle) by the Sun's edge: i.e. the time of merging of the two image
s of the Sun in the focal plane of the telescope where the CCD matrix
stands. Here comes the definition of the Solar edge for one frame as t
he collection of the inflect points on the luminosity function along e
ach of the 256 useful lines (the matrix is 512 by 512 pixels). This me
ans that a numerical derivation is performed on every other line of th
e CCD video camera which has to stand as vertical as possible. Then, f
or every frame, and through the 2.56 points, a parabola is fitted, usi
ng the least squares method. The top of this parabola materializes the
prospective characteristic point. The sets of such points associated
with the corresponding times of acquisition, are collected for both im
ages and the exact time of contact of the two images may be obtained.
This time is also the time when the solar edge crosses the almucantar.
The results for the semi-diameter obtained during 1996 campaign are d
erived from sixty measurements with the revolving mask and sixty seven
without it, performed on the Solar Astrolabe. They give a mean value
of 959.'' 39 +/- 0.'' 03 with a scatter of 0.'' 29. It is interesting
to remark that the values of the error bar and the scatter obtained do
not depend on the definition of the Solar edge, whereas the mean valu
e does depend on it. It is noticed that going with the method is made
a systematic error which slightly shrinks the diameter, but this value
can be known statistically and the correction can easily be done. Cho
osing the best definition of the Solar edge will be the matter of a fo
llowing article. The main advantage of such a digital acquisition proc
edure has to be stressed, as it enables to store the full data for fur
ther reference and, if possible, better future processing.