Figure 1, MODS exploded

1
Detector Systems for the MODS Spectrographs
Bruce Atwood, Daniel Pappalardo, Mark A. Derwent,
Thomas P. OBrien The Ohio State University
Abstract The detector plan for the four cameras
on the work-horse spectrograph of the worlds
largest optical telescope, the Large Binocular
Telescope, is presented. The two cameras of MODS
1 will be outfitted for initial deployment in
November, 2006 with 4K x 4K detectors processed
by the University of Arizona Imaging Technology
Laboratory. A custom run of 8192 x 2880
detectors is in the planning stages. These
larger detectors will be used for MODS 2 and
retrofits to MODS 1. The Dewar and electronics
design are outlined. Projections of the limiting
magnitude for the completed spectrograph at S/N5
and 10 are given.
Figure 1, MODS exploded
MODS will be the work-horse optical spectrograph
on the LBT over the wavelength range of 300 to
1000 nm with resolutions, as we purchase a full
compliment of gratings, of 2,000 to 16,000. The
11.8 m equivalent diameter LBT, and MODS, are
fully described in http//medusa.as.arizona.edu/lb
to/, Osmer, P.S. et. al., 2000, Proc. SPIE 4008,
40-49, and Byard, Paul L., OBrien, Thomas P.,
2000, Proc. SPIE 4008, 934-941. A clever man
once said that in optical design anyone can add
surfaces, the trick is to take them out while
still getting the job done. Minimizing the
number of surfaces reduces cost and increases
throughput. We have adopted a very simple
traditional design for astronomical grating
spectrographs a decentered paraboloid collimator
and a Schmidt camera. A removable dichroic beam
splitter located just behind the slit directs the
incoming light to separate red (Lgt5000 Angstroms)
and a blue (Llt5000 Angstroms) channels. The red
and blue optimized collimators produce a 230 mm
diameter beams for the red and blue four-position
grating turrets. Dispersed light goes from the
gratings to red and blue cameras that, while
similar in design, use different materials,
coatings and prescriptions. MODS includes full
multi-slit capability with a 25 position mask
interchange mechanism, a flat field and
wavelength calibration system and a guide and
wavefront-sensing camera which provides the
signal to the LBT active optics system. The
opto-mechanical modules are supported on a welded
steel space frame. While the frame is designed
to have low hysteresis, displacement of the
spectra on the detectors due to the variable
gravity vector and thermal gradients would be
unacceptable. To compensate for flexure an IR
reference beam is launched from the telescope
focal plane and passes through all the optics of
the spectrograph with the exception that is
strikes a small by-pass grating mounted in a
hole in the main gratings. A Ge quad cell in the
detector plane generates an error signal which
steers the collimators to maintain the spectra in
a fixed location on the detectors. A total of 30
stepper motors are used to position the optics.
All motors are controlled with Micro-Lynx
controllers. Mechanism positions are sensed with
proximity switches. Linear motions are
accomplished with ball screw slides which include
spring-applied/electrically-released breaks. A
computer model of the structure and the 10
sub-assemblies is shown in Figure 1. All
mechanisms are fabricated and in test. The
collimator and camera mirrors and small optics
are complete. The camera correctors are in
fabrication. On site commissioning will begin in
the 3rd Q 2006. Historically astronomical
spectrograph cameras have needed faster f-ratios
as telescope size has increased. This simply
reflects the fact that the desired slit width in
arc-sec and physical pixel size was remaining
relatively constant. LBT and MODS represent a
significant departure from this trend in two
ways. LBT (and many of the current generation of
telescopes) will produce seeing limited images
that are significantly smaller that has
previously been the case. The basic seeing
limited slit width for MODS is 0.6 arc-sec. The
MODS cameras have a large focal plane, 120 x 40
mm and will be outfitted with detectors with high
pixel count, 4096 x 4096 initially and 8192 x
2880 when our custom detectors are ready. This
allows us to bin pixels for low resolution work
in average conditions and still operate at higher
resolution and correctly match to smaller images
when available due to unusually good seeing or
when adaptive optics is available. The
combination allows us to use a relatively slow
f/3 camera. The decentered Schmidt design has the
advantages of no vignetting and that ghosts from
the detector are not returned to the grating The
computer model of one of the MODS blue cameras is
shown in Figure 2. Figure 3 is a picture of the
one of the Cameras on its handling cart. The
support structure of four spectrograph cameras,
two red and two blue, is described in Atwood, B.,
OBrien, Thomas P, 2003, Proc. SPIE 4841,
403-410. The Dewar for the MODS camera, now
out for bid, is shown in Figure 4. An 8 liter
capacity LN2 tank will give a hold time gt 30
hours. A cold link connects the LN2 reservoir to
the detector box which will house either the 4k x
4k or 8k x 3k CCD. Vacuum is maintained by cold
charcoal, to pump N2 and O2, and warm zeolite to
keep the interior of the Dewar dry even when
warm. The camera field flattener lens is the
vacuum window. MODS will be commissioned
initially with 4k x 4k STA0500A CCDs thinned,
coated and packaged by the University of
Arizonas Imaging Technology Laboratory. The red
detector and the QE of both the red and blue
detectors are shown in Figures 5 and 6. A
schematic of the 8k x 3k detector, now under
development with the Imaging Technology
Laboratory and Scientific Detector Associates, is
shown in Figure 7. The control electronics is
based on the Ohio State ICIMACS. The 32 channel
clock bias board, shown in Figure 8 with 8
channels installed, will be essentially unchanged
while the post-amplifier daughter boards will now
include on board ADCs. The Sequencer is being
ported from the ISA bus to PCI. The expected
limiting magnitude for S/N5 and 10 from a 1 hour
exposures as a function of read noise in a single
pixel is presented in Figure 9. Included are the
effects of sky noise, readout noise, combining
the signal from the two MODS, the size of a
resolution element, and system throughput.
Figure 2, Blue Camera Design
Figure 3, Blue Camera Structure
Figure 4, MODS CCD Dewar
Figure 5, MODS I 4k x 4k CCD
Figure 6 , QE of MODS 1 CCDs
Figure 9, MODS limiting magnitudes
Figure 7 , 32 output, 8192 x 2880 MODS CCD
Figure 8, 32 Channel Clock Bias Board