The MegaPrime baffle
MegaPrime is equipped with a light baffle (Figure 1) meant to stop stray
light from outside the optical beam. Without the baffle, MegaCam "sees"
the entire caisson central including its bright reflective parts (hand rails),
as well as sides of the equatorial fork (painted bright yellow). The light baffle
was installed one year after first light to reduce the stray light
contamination. It is made of aluminum and its rigidity has been tested
prior to its installation on MegaPrime (it attaches at the bottom of the
wide-field corrector). It however needed to be tested when
mounted on the telescope to ensure that it does not cause some vignetting at
extreme telescope angles due to mechanical flexure. The conclusion of this
study making use of a pinhole mask, is that while the extreme corners do
suffer from a bit of vignetting (5% maximum) there are no measurable *variable*
vignetting caused by the MegaPrime light baffle (or/and the wide-field corrector
internal light baffling) depending on the position of the telescope on
the sky. This guarantees that MegaCam's photometry is consistent across the
whole sky from a photon gathering point of view.
Figure 1. The MegaPrime baffle is the lowest extending structure (in red, 5 feet tall).
The pinhole mask was designed to obtain two non-overlapping images of an image hole
per CCD. It was obtained with the CFHT's laser-cutting machine (LAMA) giving
highly accurate holes diameter and spacing. The mask was then mounted in a
standard MegaCam filter holder. When installed in the beam, the mask is only
a couple of inches away from the focal plane. The Figure 2 presents two versions of
the same image: the high contrast version on the left shows that the great majority
of the light comes from the mirror but there are inevitable contaminations from the sides of
the wide-field corrector lenses ring holder in particular (the dim rings). The low
contrast version of the image (right) shows however that no light source compares to what
reflects off the primary mirror.
The right image of Figure 3 shows a low contrast version of a pinhole image with the
telescope at zenith. The spiders of MegaPrime (4) can be seen as well as the large
central light obstruction caused by MegaPrime itself. The horizontal bar in the upper
half of the mirror image is a reflected structure from the dome (rail). The "dents" around
the mirror are the mirror cover inner sides which have been painted black in 2003 but still
reflect a bit of light. Because MegaCam's field of view is so large, the baffle
is designed in order not to vignet the extreme corners of the mosaic, in
consequence, the pinhole image at the center of the field shows a larger
viewing diameter than just the primary mirror.
Figure 2. MegaCam image of the pinhole mask with the mirror covers open and lights
on in the closed dome (left: high contrast, right: low contrast).
Figure 3. Left: the 9 positions of the telescope. Right: individual pinhole image.
Vignetting across the field of view
The pinhole mask creates soft images and it is difficult to measure precisely
the position of the primary mirror image's edge. A simple approach is to
run the images through an edge algorithm detection, followed by some unsharp
masking. The result is much easier to grasp visually. Because of various
features on the dome reflecting on the mirror, like a crane or rails, it is
not trivial to run an automatic detection algorithm, hence the following study
is based essentially on visual control.
Figure 4 shows the pinhole image on 5 different positions of the mosaic
(CCDs: 00, 08, 13, 27, 35) for the image taken at zenith.
It is obvious here that there is indeed some
vignetting in the corners versus the center. The red circle
which encircles the "light" border (the limit of the primary mirror)
defined by the black circle, matches the central position, and is the
same for the four other positions.
If the dark circle is smaller than the red circle, this means vignetting
is present. This vignetting effect on the photometry is however easily
corrected by the combination of the flat-field and the photometric flat-field
component. Note that the vignetting does not exceed 5% (precisely measured
with surface difference in this case) in the most extreme bottom left corner
Figure 4. Vignetting in the 4 corners of the mosaic versus central position.
Vignetting versus telescope position
The most important point of it all is if the vignetting is changing with the
telescope position: this would be disastrous for the photometry!
Figure 3 (left) shows the 9 positions in which pinhole images were taken.
Note that 60 degrees from zenith is almost 10 degrees more than the most
extreme angle CFHT goes to for science observations. The following
figures (5, 6, 7, and 8) demonstrate that for a given location on the
mosaic, the vignetting does not depend on the telescope position
(all this within the limitations set by the use of the pinhole mask).
Nothing jumps out at a level higher than a fraction of a percent for
the corners at the extreme angles (where in such case the photometry
gets limited by atmospheric effects).
Figure 5. Central (CCD13) vignetting.
Figure 6. Top left (CCD00) vignetting.
Figure 7. Top right (CCD08) vignetting.
Figure 8. Bottom left (CCD27) vignetting.
Figure 9. Bottom right (CCD30) vignetting.