Up: Speckle observations of double orbits
Data were recorded during three observing runs in December 1995,
January 1997 and June 1997 at the TBL. We had quite poor weather
conditions with very short coherence times
of a few milliseconds only and a FWHM seeing often larger than 2 arcsec.
A total of 33 double stars
and 15 multiple stars was observed; details are given in Table 1.
Table of measurements. denotes the
wavelength and the bandwidth.
The column labelled Mult. gives the number of stars in the system
The column labelled Obj. give the name of the measured components
of the multiple star.
Predicted values of and are computed from the
latest available orbits.
For orbits prior to 1982 (followed by a ), orbital elements
were found in the catalogue of
Worley & Heintz, 1983.
For Hya the angular separation ABxC was computed from
AB, AC and as the distance between C and the photocenter of AB. The columns
labelled and give the residuals in and
Table of measurements (continued). For observations
made with the PAPA detector,
a has been added in superscript to the exposure time
(usually 2 ms). (*) Cyg was in fact discovered as double
and then not used as reference star in the processing of BU 151
The focal instrumentation is the speckle camera of the Aperture Synthesis
group of Observatoire Midi-Pyrénées (OMP). For the major part of the nights it was coupled with the ICCD detector of Nice University. This instrumentation
is described in Aristidi et al. (1997b)
and Prieur et al. (1998).
During the runs in January 1995 and January 1997, the images were
recorded on video tape for further processing. In June they were
also sent to a PC hosting a Matrox Genesis digitizer board
equipped with a digital signal processor (C80) which enables near
real-time processing. As an example the power spectrum
for a 128128 image size is computed at a rate of
about 20 frames per second.
The use of a reference star for usual calibration of the telescope+atmosphere transfer function was sometimes avoided by computing
the cross-correlation between time-separated images and
subtracting it from the autocorrelation (Worden et al. 1977).
Though at a lower speed (9 frames/s) the system is also
programmed to compute the cross correlation between the images
and their square in order to find the absolute position angles
of binaries (Aristidi et al. 1997a). This system does not
include yet the classification of images according to the seeing that
is used in the data processing as described in Aristidi
et al. (1997b). Bright stars () were re-processed
from the video tape for magnitude difference determination
using probability imaging (Carbillet et al. 1998a).
This provides also the absolute position angle (PA) and
is useful to confirm the PA computed by the cross-correlation
technique (Aristidi et al. 1997a).
Problem of saturation is took into account the following way: before recording frames, the video
digitizer displays saturated pixels, allowing the observer to adjust the gain and the offset of the detector
to reduce the saturation. Moreover, at processing time (probability imaging), pixels at the maximum intensity level are rejected from
the calculation of the magnitude difference.
The PAPA camera was used during part of these observations but
because of technical testing and bad weather conditions it could
only lead to measurements
on the night of 23/06/97. It is actually a new version of the
original camera described in Papaliolios et al. (1982, 1985).
Modifications have been jointly implemented by
P. Nisenson (Harvard Center for Astrophysics), D. Gezari (NASA)
and L. Koechlin (OMP) in the last five years.
The current version has a new binary mask setup and a refurbished
A problem in the regulation of the high voltage power supply
caused a strong geometric distorsion and a variation of the overall scale
which imposed quasi permanent scale calibrations during the night.
A small "hole'' at the center of the autocorrelation function was
due to a lack of detectivity of the electronics after each
A whole reduction procedure had to be set up in order to correct
for these defects:
- Each star image was corrected with the distortion and scaling
coefficients interpolated from the two images of
the calibrating grid which were taken immediately before and after
the star observation.
- To remove the hole at the center of the autocorrelation, each
was correlated with the photo-events of a "gliding" time window located
at 1 msec after the arrival time of the photo-event.
The very poor atmospheric conditions did not allow us to use a
time window larger than 2 msec (as the typical coherence time was less than
Up: Speckle observations of double orbits
Copyright The European Southern Observatory (ESO)