|< Autoguider ]||[ Electronics >|
Fig. 1. Field of view indicators for imaging M64 with the TopBox autoguider.
The usable guiding regions for the TopBox autoguider are constrained to E and W offsets by the one-dimensional motion of the positioning stage. If the offset is too small, the pickoff mirror will intercept light that should reach the field of view of the main camera (vignetting). If the offset is too large, off-axis optical aberrations will distort star images so badly that their positions cannot be measured accurately enough to use as reference points for guiding. Optical design models for the field of view of the SBIG STL-1001E external guider camera, verified by real observations at the telescope, lead to the choice of two rectangular regions with dimensions of 2.74x7.51 arcmin on the sky with central offsets 27.16 arcmin E and 31.10 arcmin W from the main field center, shown as red rectanglar outlines in Fig. 1. Since a full frame image from the guider camera is 3.42x2.58 arcmin (blue outline in Fig. 1), guide stars within these search regions should not fall near the edges of the guider images.
Precise alignment of the autoguider optics with the main telescope optics is a tedious process requiring observing time for verification. Currently the center of the autoguider field is offset 1.97 arcmin W and 0.75 arcmin S from the center of the main telescope field of view, and this is the actual configuration shown in Fig. 1 and programmed into the TopBox driver software.
The pointing of the autoguider depends on the three Euler angles of the pickoff mirror (Fig. 2, left panel), which is mounted on a rotating stalk (vertical axis) with three adjustable points (a triangular array of screws threaded vertically through the stalk) that change the "tip" and "tilt" angles about two horizontal axes. The stalk must be removed from the TopBox to adjust these three screws, then tightened back into place in the correct rotational orientation with a 7/16" hex nut on its central threaded rod.
Fig. 2. Left: Detail of the autoguider pickoff mirror mounting. Right: Testing the alignment with a laser collimator.
A laser light source designed to fit into a standard 1.25" diameter eyepiece
holder ("laser collimator") is used to test the alignment. Beginning from an
approximate starting point set on the construction workbench (Fig. 2, right
panel), fine adjustments can be made with the TopBox mounted on the telescope,
using reflections in the telescope optics. With the camera mount removed from
the bottom of the TopBox it is possible to reach the pickoff mirror mounting
nut, make adjustments, and visually inspect the reflections of the laser beam
from the collimator installed in the autoguider focuser drawtube. The present
degree of alignment, measured by observations on the sky, was reached by
Filter Wheel 1 of the TopBox contains two Savart plates oriented with the directions of their beam displacements at 45 degrees to each other (Fig. 3). The converging telescope beam passing through each plate produces two images with orthogonal polarization states for each star in the field of view. Differential photometry of a pair of such double images taken through the two Savart plates yields independent measurements of both linear Stokes parameters, hence the degree and position angle of linear polarization for each star in the field.
Fig. 3. The two Savart plate polarization analyzers mounted in the TopBox filter wheel.
A common polarizing beamsplitter is a plane-parallel calcite plate with the optic axis at 45° to the surface in a plane containing the surface normal. This will split a converging beam to form two orthogonally polarized images, but the two images will have different focal properties because their optical path lengths are not equal. A Savart plate, however, consists of two such plates with the planes crossed at 90° (Fig. 4), so that the ordinary and extraordinary rays interchange at the boundary, which equalizes the optical path lengths to produce two identically focused orthogonally polarized images for incoming light at normal incidence. This reduces the error in differential photometry of star images for computing the linear Stokes parameters.
Fig. 4. Polarization beam splitting by a Savart plate.
For the strongly converging f/8 beam of the RRRT and the wide 17x17 arcmin
field of view with the SBIG STL-1001E CCD camera there are still significant
differences in the shapes of the double images: orthogonal polarization states
produce similarly elliptical images, but the ellipses are noticeably elongated
in the directions of the beam displacements, which differ by 90°. The
effect on the magnitude differences is negligible if the stars are well enough
isolated for simple aperture photometry with apertures large enough to include
most of the flux. However, the precision of standard PSF fitting photometry
methods for crowded fields is severely limited because two distinctly different
PSFs are required to deconvolve many of the overlapping star images.
The Savart plates in the TopBox were roughly positioned in their filter holders
by observing the splitting of a laser beam with a linear polarizer. A second
adjustment was made by physically rotating the Savart plates in their filter
holders by angles derived from observations of polarized standard stars and
laid out with an ordinary desk protractor. The final fine adjustment was made
in software by slight changes in the commanded filter wheel positions based on
further observations of polarized standard stars. The errors in the position
angles of the Savart plates are now on the order of a few tenths of a degree.
The Savart plates in the TopBox were roughly positioned in their filter holders by observing the splitting of a laser beam with a linear polarizer. A second adjustment was made by physically rotating the Savart plates in their filter holders by angles derived from observations of polarized standard stars and laid out with an ordinary desk protractor. The final fine adjustment was made in software by slight changes in the commanded filter wheel positions based on further observations of polarized standard stars. The errors in the position angles of the Savart plates are now on the order of a few tenths of a degree.
|Last modified: Nov 15, 2012 David McDavid||