The Astronomical Journal, 146:32 (40pp), 2013 August doi:10.1088/0004-6256/146/2/32
C
2013. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
THE MULTI-OBJECT, FIBER-FED SPECTROGRAPHS FOR THE SLOAN DIGITAL SKY SURVEY
AND THE BARYON OSCILLATION SPECTROSCOPIC SURVEY
Stephen A. Smee
1
, James E. Gunn
2
, Alan Uomoto
3
, Natalie Roe
4
, David Schlegel
4
, Constance M. Rockosi
5
,
Michael A. Carr
2
, French Leger
6
, Kyle S. Dawson
7
, Matthew D. Olmstead
7
, Jon Brinkmann
8
, Russell Owen
6
,
Robert H. Barkhouser
1
, Klaus Honscheid
9
, Paul Harding
10
, Dan Long
8
, Robert H. Lupton
2
, Craig Loomis
2
,
Lauren Anderson
6
, James Annis
11
, Mariangela Bernardi
12
, Vaishali Bhardwaj
6
, Dmitry Bizyaev
8
, Adam S. Bolton
7
,
Howard Brewington
8
, John W. Briggs
13
, Scott Burles
14
, James G. Burns
9
, Francisco Javier Castander
15,16
,
Andrew Connolly
6
, James R. A. Davenport
6
, Garrett Ebelke
8
, Harland Epps
5
, Paul D. Feldman
1
,
Scott D. Friedman
16
, Joshua Frieman
11
, Timothy Heckman
1
, Charles L. Hull
3
, Gillian R. Knapp
2
,
David M. Lawrence
7
, Jon Loveday
17
, Edward J. Mannery
6
, Elena Malanushenko
8
, Viktor Malanushenko
8
,
Aronne James Merrelli
18
, Demitri Muna
19
, Peter R. Newman
8
, Robert C. Nichol
20
, Daniel Oravetz
8
,
Kaike Pan
8
, Adrian C. Pope
21
, Paul G. Ricketts
7
, Alaina Shelden
8
, Dale Sandford
5
, Walter Siegmund
6
,
Audrey Simmons
8
, D. Shane Smith
9
, Stephanie Snedden
8
, Donald P. Schneider
22,23
, Mark SubbaRao
12
,
Christy Tremonti
24
, Patrick Waddell
25
, and Donald G. York
26
1
Department of Physics and Astronomy, Johns Hopkins University, Baltimore, MD 21218, USA; smee@pha.jhu.edu
2
Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544, USA
3
Observatories of the Carnegie Institution of Washington, 813 Santa Barbara Street, Pasadena, CA 91101, USA
4
Physics Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
5
UC Observatories and Department of Astronomy and Astrophysics, University of California, Santa Cruz,
375 Interdisciplinary Sciences Building (ISB) Santa Cruz, CA 95064, USA
6
Department of Astronomy, University of Washington, Box 351580, Seattle, WA 09195, USA
7
Department of Physics and Astronomy, University of Utah, Salt Lake City, UT 84112, USA
8
Apache Point Observatory, Sunspot, NM 88349, USA
9
Department of Physics and Center for Cosmology and Astro-Particle Physics, Ohio State University, Columbus, OH 43210, USA
10
Department of Astronomy, Case Western Reserve University, Cleveland, OH 44106, USA
11
Fermi National Accelerator Laboratory, P.O. Box 500, Batavia, IL 60510, USA
12
Department of Physics and Astronomy, The University of Pennsylvania, 209 South 33rd Street, Philadelphia, PA 19104, USA
13
HUT Observatory, Mittelman Family Foundation, P.O. Box 5320, Eagle, CO 81631, USA
14
Physics Department, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA 02139, USA
15
Institut de Ci encies de l’Espai (IEEC-CSIC), E-08193 Ballaterra, Barcelona, Spain
16
Space Telescope Science Institute, Baltimore, MD 21218, USA
17
Astronomy Centre, University of Sussex, Falmer, Brighton BN1 9QJ, UK
18
Department of Astronomy, California Institute of Technology, Pasadena, CA 91125, USA
19
Center for Cosmology and Particle Physics, New York University, 4 Washington Place, New York, NY 10003, USA
20
Institute of Cosmology and Gravitation (ICG), Dennis Sciama Building, University of Portsmouth, Portsmouth PO1 3FX, UK
21
High Energy Physics Division, Argonne National Laboratory, 9700 South Cass Avenue, Lemont, IL 60439, USA
22
Department of Astronomy and Astrophysics, The Pennsylvania State University, University Park, PA 16802, USA
23
Institute for Gravitation and the Cosmos, The Pennsylvania State University, PA 16802, USA
24
Department of Astronomy, University of Wisconsin-Madison, Madison, WI 53703, USA
25
NASA Ames Research Center, Moffett Field, CA 94035, USA
26
Department of Astronomy and Astrophysics and the Fermi Institute, The University of Chicago, Chicago, IL 60637, USA
Received 2012 August 10; accepted 2013 May 15; published 2013 July 12
ABSTRACT
We present the design and performance of the multi-object fiber spectrographs for the Sloan Digital Sky Survey
(SDSS) and their upgrade for the Baryon Oscillation Spectroscopic Survey (BOSS). Originally commissioned in
Fall 1999 on the 2.5 m aperture Sloan Telescope at Apache Point Observatory, the spectrographs produced more than
1.5 million spectra for the SDSS and SDSS-II surveys, enabling a wide variety of Galactic and extra-galactic science
including the first observation of baryon acoustic oscillations in 2005. The spectrographs were upgraded in 2009
and are currently in use for BOSS, the flagship survey of the third-generation SDSS-III project. BOSS will measure
redshifts of 1.35 million massive galaxies to redshift 0.7 and Lyα absorption of 160,000 high redshift quasars over
10,000 deg
2
of sky, making percent level measurements of the absolute cosmic distance scale of the universe and
placing tight constraints on the equation of state of dark energy. The twin multi-object fiber spectrographs utilize a
simple optical layout with reflective collimators, gratings, all-refractive cameras, and state-of-the-art CCD detectors
to produce hundreds of spectra simultaneously in two channels over a bandpass covering the near-ultraviolet to
the near-infrared, with a resolving power R = λ/FWHM ∼ 2000. Building on proven heritage, the spectrographs
were upgraded for BOSS with volume-phase holographic gratings and modern CCD detectors, improving the peak
throughput by nearly a factor of two, extending the bandpass to cover 360 nm <λ<1000 nm, and increasing the
number of fibers from 640 to 1000 per exposure. In this paper we describe the original SDSS spectrograph design
and the upgrades implemented for BOSS, and document the predicted and measured performances.
Key words: cosmology: observations – instrumentation: spectrographs – surveys
Online-only material: color figures
1
The Astronomical Journal, 146:32 (40pp), 2013 August Smee et al.
1. INTRODUCTION
The Sloan Digital Sky Survey (SDSS; York et al. 2000)
project was conceived in the mid-1980s as an ambitious en-
deavor to understand the large-scale structure of the universe.
SDSS and its extension, SDSS-II, conducted a coordinated
imaging and spectroscopic survey from 2000–2008 over ap-
proximately 10,000 deg
2
of high Galactic latitude sky. Now in
its third phase of operation, SDSS is one of the most successful
projects in the history of astronomy. The survey has produced
an enormous catalog consisting of five-band digital images that
include nearly one billion unique objects, and spectra of 930,000
galaxies, 120,000 quasars, and 460,000 stars, all publicly
available (Abazajian et al. 2009, and references therein).
To obtain these imaging and spectroscopic data, a dedicated
2.5 m telescope (Gunn et al. 2006), wide-field mosaic CCD cam-
era (Gunn et al. 1998), and twin multi-object fiber spectrographs
were constructed and installed at the Apache Point Observatory
(APO) in Sunspot, New Mexico. The telescope, built to accom-
modate the requirements for both imaging and spectroscopy,
is shared by the camera and spectrographs, which mount at
the Cassegrain focus. The imaging survey was carried out on
clear, dark nights with good seeing using the 120 mega-pixel
camera, which operated in drift-scanning mode using a 5 × 6
array of 2048 × 2048 pixel detectors to obtain ugr iz (Fukugita
et al. 1996), photometry. The imaging data, once reduced and
calibrated (Smith et al. 2002; Pier et al. 2003; Ivezi
´
cetal.
2004; Tucker et al. 2006; Padmanabhan et al. 2008), were used
for spectroscopic target selection. Spectroscopy was performed
using the two multi-object fiber spectrographs, collecting 640
spectra over the 3
◦
diameter field in one exposure.
In this paper, we describe the design and performance of the
SDSS spectrographs, and their recent upgrade for the Baryon
Oscillation Spectroscopic Survey (BOSS; Schlegel et al. 2009;
Dawson et al. 2013). BOSS is the flagship survey in the third-
generation SDSS-III program currently underway at the 2.5 m
SDSS telescope (Eisenstein et al. 2011). BOSS will measure
the cosmic expansion history of the universe to percent-level
precision by mapping an immense volume of sky to obtain the
spatial distributions of galaxies and quasars, and from it, the
characteristic scale imprinted by baryon acoustic oscillations
(BAO) in the early universe (for a review of BAO with a respect
to other cosmological probes, see Weinberg et al. 2013). A
measure of the scale at low redshifts, out to z ∼ 0.7, will
be obtained by carrying out a redshift survey of 1.35 million
massive galaxies from 10,000 deg
2
of SDSS data. BOSS will
also observe Lyα absorption in the spectra of 160,000 high-
redshift quasars to measure large-scale structure at redshifts of
z ∼ 2.5.
Each SDSS spectrograph utilizes a dual-channel design with
a common reflecting collimator and a dichroic to split the beam
into a blue channel and a red channel. In each channel, just
downstream of the dichroic, a transmitting grism disperses
the light, which is imaged by an all-refractive camera onto a
CCD. For BOSS, the basic optical design has been retained,
with several improvements. The ruled gratings have been
replaced by volume-phase holographic (VPH) grisms (gratings
sandwiched between two prisms) and the CCDs have been
replaced with more modern devices. These changes produce a
significant improvement in throughput and a modest extension
of the wavelength range in both the blue and red channels.
Additionally, smaller diameter fibers that are better matched to
the angular scale of BOSS targets have been installed, allowing
the total number of simultaneous spectra obtained from the two
spectrographs to be increased from 640 in the original design to
1000 in the BOSS configuration.
The remainder of this paper is organized as follows. In
Section 2 we begin by describing the design and construction of
the original SDSS spectrographs in some detail, published here
for the first time. This is followed in Section 3 by a discussion
of the spectrograph upgrades completed in 2009 for BOSS. The
performance of both the original SDSS spectrographs and the
upgraded BOSS design is presented in Section 4. Finally, some
highlights of the scientific research enabled by these instruments
is provided in Section 5.
2. SDSS SPECTROGRAPH DESIGN
2.1. Design Requirements
The requirements for the SDSS spectrographs were set by its
primary scientific goal: the creation of athree-dimensional wide-
area map of the universe to reveal its large-scale structure. The
SDSS imaging survey provides the two-dimensional locations
of nearly one billion celestial objects, and spectroscopy of a
selected subset of targets is then used to determine redshifts and
thus distances. The project set as a requirement spectroscopy
of one million galaxies and 100,000 quasars distributed over
approximately 10,000 deg
2
.
Acquisition of a large number of spectra simultaneously over
a large field of view, with moderate resolution sufficient for
accurate redshift measurements, naturally led to the choice of a
fiber-fed multi-object spectrograph. The spectrograph design
was dictated in large part by the design of the telescope, which
was itself optimized for both wide-field, multi-band, imaging
and multi-object spectroscopy. Requirements were specified
when possible; however, the instrument design was largely
driven by technology available at the time.
In what follows throughout Section 2.1, we summarize
the requirements that dictated the design of the SDSS
spectrographs.
2.1.1. Telescope Design
The 2.5 m SDSS telescope is a modified distortion-free
Ritchey–Chr
´
etien design with a 3
◦
diameter field of view, and
f/5 final focal ratio, which provides a good match to fibers
for spectroscopy (180 μm diameter, 3
) and to the imaging
CCDs (pixel size 24 μm, 0.
4). The optical design incorporates
two aspheric corrector lenses, a Gascoigne-type design located
near the vertex of the primary mirror, and two interchangeable
secondary correctors, one used for imaging and the other for
spectroscopy. The imaging corrector is a thick fused silica lens
located close to the focal plane and is incorporated into the
SDSS camera, where it serves a mechanical function in addition
to providing optical correction. The spectroscopic corrector
is a thinner lens located further from the focal plane and
optimized for chromatic focus. The plate scale for spectroscopy
is 3.627 mm arcmin
−1
. The spectroscopic focal surface is
slightly curved, with a maximum deviation from a plane of
2.6 mm. One important detail of the spectroscopic optics is that
the central ray for each field point is not perpendicular to the
focal plane, necessitating a clever correction scheme for fiber
placement that will be described below.
2.1.2. Number of Fibers
Spectroscopy of approximately one million objects over
10,000 deg
2
, plus 10%–20% additional fibers for calibration
2
The Astronomical Journal, 146:32 (40pp), 2013 August Smee et al.
sources and sky background measurements, implies a density
of 120 deg
−2
. The 2.5 m telescope has a field of view of 7 deg
2
,
but each plate will view a unique area on the sky of about
5deg
2
. The higher density of plates is due to the need for overlap
between fields to ensure complete sky coverage without gaps,
and to allow multiple observations to cross-calibrate the entire
survey. The required number of fibers is therefore approximately
600 per plate.
A practical limit on the number of fibers was imposed by the
detector format, camera design, and fiber mounting scheme. For
proper spectral sampling, the fiber images on the detector should
be about 3 pixels in diameter, with an equal space between
spectra to reduce crosstalk and allow for a measurement of
the scattered light floor. Thus, each spectrum used six detector
columns, and the 2048×2048 pixel detector could accommodate
a maximum of 341 spectra. The actual number was reduced to
320 spectra to avoid camera optical distortions near the detector
edges and to allow for extra gaps between groups of 20 fibers,
which was necessary for the fiber mounting scheme described
in Section 2.2. These larger gaps turned out to be quite useful for
measurements of scattered light in the wings. The final choice
of 640 fibers per plate, or 320 per spectrograph, provided some
contingency over the required 600 fibers, allowing for broken
fibers, additional calibration fibers, and/or ancillary programs
utilizing the extra fibers.
2.1.3. Fiber Diameter
The fiber diameter is set by the desire to maximize the
signal-to-noise ratio (S/N) for an extended source given the sky
background. For the galaxies of interest around redshift z = 0.1
and the sky conditions at Apache Point, this corresponds to a
fiber size of around 3
, or a fiber diameter of 180 μm. Fibers of
good optical quality were also readily obtainable in this size.
2.1.4. Wavelength Range
Redshifts are determined either from absorption lines or
emission lines—in both cases only a few lines contribute most
of the signal. In absorption, three features are dominant: the Mg
b triplet at λ = 5180 Å, Ca at λ = 5270 Å, and the Na i doublet
(D lines) at λ = 5890 Å. At shorter wavelengths, the Ca ii K
and H lines at λ = 3933, 3969 Å and the G band 4300 Å may
also be detected in absorption. In emission, Hα = 6353 Å is the
strongest (and often the only) line, although the [O ii] doublet
may also be visible at λ = 3727 Å.
Given the availability of these spectral features, and consider-
ing practical limitations on UV throughput, the short wavelength
cutoff was set at 3900 Å to ensure that the H and K lines of Ca ii
are observable even at zero redshift, while the [O ii] doublet is
observable at z>0.05. Redshift determination for most nearby
galaxies could have been accomplished with a single blue arm
extending up to 6000 Å; however, the SDSS imaging camera
was designed to measure to the detector red limit cutoff, so it
was decided to take advantage of the detector sensitivity in the
spectrographs and add the red channel. This would enable ob-
servation of Hα to a redshift of z = 0.2 or more, as well as the
observation of quasars out to redshifts beyond z = 5.
Extension of the upper wavelength cutoff to 9100 Å opened
up a rich new vein of scientific discovery that was not anticipated
at the time of the instrument design. In particular, pushing the
long wavelength cutoff as high as possible extended the limit for
redshift determination of luminous red galaxies (LRGs) using
the 4000 Å break. The LRG sample (Eisenstein et al. 2001)was
used to make the first observation of the BAO feature, which in
turn motivated the future upgrade of the spectrographs to even
longer wavelengths for BOSS, as discussed later in this paper.
2.1.5. Resolving Power
The spectroscopic resolution is defined as the full width at
half-maximum (FWHM) of the one-dimensional point-spread
function (PSF), in wavelength units (a resolution element). The
resolving power is the wavelength divided by this quantity,
and we will often use the phrase “higher resolution” to mean
higher resolving power, as is the normal usage. Given a fixed
number of pixels in the dispersion direction and requiring proper
sampling, increasing resolving power reduces the wavelength
range. Higher resolving power also reduces the number of
source photons per pixel, increasing the exposure time required
to exceed the CCD read noise. On the other hand, if the resolving
power is too low, absorption lines cannot be resolved and this
will ultimately degrade both the accuracy and success rate of
redshift measurements.
The resolution was therefore set by the requirement to obtain
spectroscopic redshifts of galaxies to an accuracy limited only
by the broadening due to typical velocity dispersions of 100
to 200 km s
−1
. This corresponds to a resolving power of
1500–3000.
The actual resolving power as a function of wavelength was
allowed to vary within these limits to optimize the red–blue
channel wavelength split location and the total wavelength cov-
erage, while maintaining well-sampled spectra with 3 pixels per
resolution element on the CCD over the full wavelength range.
These choices of spectrograph parameters were chosen to opti-
mize the overall redshift success rate for a given exposure time.
2.1.6. Throughput and Signal-to-noise Ratio
The requirement on throughput was set by the desire to obtain
one million spectra over 10,000 deg
2
to a limiting Petrosian
magnitude of r = 18.15 in five years, corresponding to roughly
100 deg
−2
galaxies. Given the number of fibers and average
weather at APO, this implied an average exposure time of one
hour.
Provided that the spectral resolution is sufficient to resolve the
absorption lines, the minimum S/N needed to derive a redshift
depends mainly on the strength of the absorption lines. For
convenience, the S/N per Å of the spectral continuum will be
quoted. For an elliptical galaxy with strong absorption features,
spectra obtained in the Center for Astrophysics redshift surveys
(Huchra et al. 1983; Falco et al. 1999) demonstrated that one
can measure a reliable redshift with S/N per Å > 8, i.e., one
needs to collect 64 object photons per Å, assuming that the
noise is dominated by photon statistics from the source. This
number must be increased, however, if sky background and/or
readout noise is significant. A significant problem for some
galaxies is that they have weak absorption lines (presumably
because they have a significant amount of light from early-
type stars) and yet lack strong Hα emission. In these cases
one may need two or three times as many photons to derive
an absorption-line redshift. We adopt as a guide the goal of
obtaining spectra with S/N of 15 per Å. Simulated galaxy
and quasar spectra indicated that we could in fact reach
this goal with exposures of somewhat less than one hour in
typical conditions for seeing and atmospheric extinction. The
corresponding throughput requirement, including atmospheric
extinction and the telescope throughput, varies as a function of
wavelength; the maximum requirement is about 17% at 7000 Å,
3
The Astronomical Journal, 146:32 (40pp), 2013 August Smee et al.
Figure 1. Rendering of a fiber cartridge. The fiber cartridge consists of a cast
aluminum body that supports the fiber harness, the two slitheads, and the plug-
plate, which has a diameter of 800 mm. The slitheads are attached to the cartridge
body with a spring-loaded seating system that provides alignment for insertion
into the spectrograph bodies, but then allows the slithead to float free from
the cartridge body and engage the slithead-to-spectrograph kinematic mounting
system. Kinematic mounts around the periphery of the cartridge casting ensure
accurate and repeatable placement of the cartridge with respect to the telescope.
with requirements of roughly 10%, 15%, and 10% at 4000 Å,
6000 Å, and 8000 Å, respectively.
2.2. Fiber System Design
2.2.1. Overview of Fiber System
Light is transmitted from the telescope focal plane to two
identical spectrographs by fiber optic strands 180 μmindiame-
ter (3
on the sky). Light enters the fibers at the telescope focal
plane in a cone of numerical aperture 0.1 (f/5 beam), and the
spectrograph collects light emitted from the other end of the
fibers in a slightly larger cone with numerical aperture 0.125
(f/4), due to focal ratio degradation (FRD) that occurs as the
light travels down the fiber. Any light emitted outside this cone
will be lost, so a primary requirement on the fiber system is
to limit FRD so as to maximize throughput. To this end, the
spectrographs are mounted on the telescope to avoid any rela-
tive motion between the two ends of the fibers and the potential
stress that can result in increased FRD (an issue that was not
well understood at the time). References available in those days
(early 1990s) suggested that the macrobending of fibers is be-
nign (Angel etal. 1977; Heacox 1986; Clayton1989); however, a
recent study shows that the repeated bending of fibers, as would
be the case for a bench-mounted spectrograph, can increase
FRD over several years of operation (Murphy et al. 2012). This
scheme also maximizes throughput by keeping the fibers short,
minimizes fiber throughput variations due to physical motion
and stress, and avoids the problems of routing and protecting
long fiber runs. The sky ends of the fibers are plugged into
drilled 800 mm diameter aluminum plates called plug-plates
that position the fibers on the spectrograph focal plane, and the
other ends of the fibers are terminated in one of two slitplates.
Each thin slitplate is mounted to a rigid frame with precision
locating features for accurate placement in the spectrograph.
The assembly of plug-plate, fibers, and slitheads is mechani-
cally supported by a portable aluminum cartridge that can be
Figure 2. Photograph of a BOSS fiber cartridge. Fibers plugged into the back
of the plug-plate are routed in bundles to the slitheads (the two boxes standing
upright at the left and right side of the cartridge). The design shown is identical
to that used for SDSS except for the number and size of the fibers. For SDSS,
320 fibers are routed to each slithead, while for BOSS each slithead carries
500 fibers.
Figure 3. Photograph showing a fiber cartridge being installed on the telescope.
The twin spectrographs are the green instruments on either side of the focal
plane. The cartridge is raised by a hydraulic lift in the floor below the
primary cell. When raised, the cartridge engages kinematic mounts for precise
location. At the same time, the two slitheads engage the spectrographs, each
of which is located by its own kinematic mounting features integral to the
slithead and spectrograph optical bench. Installation takes approximately three
to five minutes.
installed on the telescope by a single operator in a few minutes.
New plug-plates are mounted on the cartridges during the day
and plugged with fibers, then sequentially mounted on the tele-
scope during the night. A rendering and photograph of a fiber
cartridge are shown in Figures 1 and 2, respectively.
For each new sky field, a cartridge is wheeled under the
telescope using the Linde cart (named for its designer Carl
Lindenmeyer) and attached to the telescope rotator using pneu-
matic clamps. At the same time the attached slitheads enter the
spectrographs through the open slithead doors and are clamped
in place. A kinematic mounting interface ensures accurate, re-
peatable placement. The photograph in Figure 3 illustrates the
operation. Eight cartridges were fabricated for SDSS to provide
sufficient pre-plugged plates for an entire night of observing.
Each cartridge also has a set of coherent fibers that are placed
on pre-selected guide stars and viewed by the guider camera.
These stars are used for field rotation and translation to align
the plug-plate to the field.
4
The Astronomical Journal, 146:32 (40pp), 2013 August Smee et al.
2.2.2. Cartridges
The fiber cartridge consists of a machined aluminum cast
body that supports the optical fiber harnesses, spectrograph
slitheads, and plug-plate. Assembling these components into
a single robust unit protects the fragile fibers during the
manipulations necessary for plugging, transport to and from
the telescope, and mounting onto the instrument rotator. The
cartridges are plugged during the day, and are designed so they
can be quickly installed on the telescope at night under often
difficult conditions of low light and cold temperatures.
The plug-plate holder consists of two large bending rings that
warp the plug-plate to match the telescope best-focus surface.
An adjustment rod centered on the back side of the plug-plate
is used to fine-tune the plate curvature. The bending rings
are mounted to the cartridge body with a set of kinematic
pin mounts. The alignment of the cartridge to the telescope
is provided by another kinematic mount employing v-groove
blocks on the cartridge that engage with v-blocks on the
telescope. This system ensures proper and repeatable alignment
between the plug-plate and telescope focal plane.
The two slitplates, each supporting 320 fibers, are mounted in
their respective slitheads. These slitheads are aluminum assem-
blies mounted outboard of the cartridge body that support and
protect the slitplates. The slitheads are attached to the cartridge
body with a spring-loaded seating system that provides align-
ment for insertion into the spectrograph bodies, but then allows
the slithead to float free from the cartridge body and engage the
slithead-to-spectrograph kinematic mounting system. When not
mounted on the telescope, these slitheads are protected by slid-
ing covers to prevent contamination and/or mechanical contact
with the delicate slitplates.
All cartridge operations occur at the same elevation, on the
telescope platform and the adjacent support building. In the
plugging lab, the exposed plug-plates from the previous night’s
observing are unplugged and removed from their cartridges and
new plates are installed. Once plugging and fiber mapping is
completed (a process that takes 30–45 minutes), the 145 kg
cartridge is stowed on a lift table installed in a bay that provides
both interior and exterior bay door access. At night, the outside
door is opened to allow the cartridges to equilibrate to the
temperature of the ambient air.
To install a new cartridge on the telescope, an outside
manipulator arm is employed to move the cartridge from the
storage bay to one of two receiving plates on the Linde cart.
The cartridge is then wheeled from the storage bay to the
telescope. With the telescope parked at zenith and locked
into position, the Linde cart is rolled under the mounted
cartridge to align the empty receiver plate with it. Aided by a
hydraulic lift, the observer removes the exposed cartridge from
the telescope. Then, maneuvering the Linde cart to align the
unexposed cartridge onto the hydraulic lift, the observer mounts
the new cartridge onto the telescope instrument rotator. Once the
new cartridge is latched and the cart receiver plate is lowered
back onto the Linde cart, the cart is rolled out from under the
telescope. The telescope is now ready to move to the next field
and to begin another exposure. Only three to five minutes is
required to perform this cartridge change.
As the cartridge is lifted into place and clamped to the
telescope, the slitheads are simultaneously inserted into sockets
in the spectrographs. The slitheads are attached to the cartridge
frame by stiff springs so that they can move slightly with respect
to the rest of the cartridge. Once the cartridge has been correctly
positioned and clamped to the telescope, each slithead is loaded
Figure 4. Two schematic views of the cartridge mounted on the telescope. Top:
a cutaway side view showing the slitheads inserted into the spectrographs (only
nine fiber harnesses are shown). Bottom: bottom view showing the cartridge
located between the two spectrographs, which are mounted to the instrument
rotator (depicted as the large outer circle).
(A color version of this figure is available in the online journal.)
against a three-point kinematic mount on the spectrograph by
a single pneumatic clamp. A flexible rubber seal between the
slitheads and the spectrograph bodies prevents extraneous light
from entering during exposures. Each slithead is coded and
its identification relayed to the observer’s workstation when
it is inserted. This information allows adjustments for each
slithead, e.g., image placement on the CCD and focus, to be
made automatically. Figure 4 shows two schematic views of the
cartridge mounted on the telescope with the slitheads inserted
into the spectrographs.
2.2.3. Optical Fiber
The selected optical fiber material is a silica UV-enhanced
step-index fiber with a core diameter of 180 μm, a thin cladding
and a polyimide protective layer. The actual fiber was Polymicro
Technologies, Inc.
27
FHP 180-198-218, where the numbers
refer to the diameter of the bare fiber, plus cladding and plus
polyimide buffer.
2.2.4. Fiber Harnesses
A fiberharness consists of 20fibers of length 1.865 ±0.025 m;
each fiber cartridge contains 32 fiber harnesses. The fiber
27
Polymicro Technologies, Inc., http://www.polymicro.com
5
