LONG-RANGE RECONNAISSANCE IMAGER ON NEW HORIZONS

A. F. Cheng (1), H. A. Weaver (1), S. J. Conard (1), M. F. Morgan (1),
O. Barnouin-Jha (1), J. D. Boldt (1),  K. A. Cooper (1),
E. H. Darlington (1), M. P. Grey (1),  J. R.  Hayes (1),
K. E. Kosakowski (2), T. Magee (1), E. Rossano (1), D. Sampath (2),
C. Schlemm (1), H. W. Taylor (1)

1. The Johns Hopkins University Applied Physics Laboratory, 
11100 Johns Hopkins Road, Laurel MD 20723
Tel. 240-228-5415
andrew.cheng@jhuapl.edu

2. SSG Precision Optronics, Wilmington MA

ABSTRACT

The LOng-Range Reconnaissance Imager (LORRI) is the high resolution
imaging instrument for the New Horizons mission to Pluto, its giant satellite
Charon, its small moons Nix and Hydra, and the Kuiper Belt, which is
the vast region of icy bodies extending roughly from Neptune's orbit
out to 50 astronomical units (AU). New Horizons launched on January
19, 2006 as the inaugural mission in NASA's New Frontiers
program. LORRI is a narrow angle (field of view=0.29 degrees), high
resolution (4.95 microrad pixels), Ritchey-Chretien telescope with a 20.8
cm diameter primary mirror, a focal length of 263 cm, and a three lens
field-flattening assembly. A 1024x1024 pixel (optically active
region), thinned, backside-illuminated charge-coupled device (CCD)
detector is used in the focal plane unit and is operated in frame
transfer mode. LORRI provides panchromatic imaging over a bandpass
that extends approximately from 350 nm to 850 nm.  LORRI operates in
an extreme thermal environment, situated inside the warm spacecraft
with a large, open aperture viewing cold space. LORRI has a silicon
carbide optical system, designed to maintain focus over the operating
temperature range without a focus adjustment mechanism. Moreover, the
spacecraft is thruster-stabilized without reaction wheels, placing
stringent limits on the available exposure time and the optical
throughput needed to satisfy the measurement requirements.

1. Introduction

The New Horizons mission launched on January 19, 2006 on its way to
perform the first reconnaissance of the Pluto-Charon system and the
Kuiper Belt. First, however, New Horizons made a Jupiter swingby with
closest approach on Feb 28, 2007. Extensive observations of the Jovian
atmosphere, rings, and satellites were acquired. Pluto closest
approach will occur on July 14, 2015. The prioritized measurement
objectives of the New Horizons mission, and the contributions from
LOng-Range Reconnaissance Imager (LORRI) observations, are summarized
in Table 1.

Table 1. Pluto and Charon Prioritized Measurement Objectives 

Pluto is an icy dwarf planet with a significant atmosphere consisting
mainly of nitrogen, CO and methane. High resolution images from LORRI
will yield important information on Pluto's geology and surface
morphology, collisional history, and atmosphere-surface
interactions. Will Pluto have a young surface, with evidence of
endogenic activity like plumes or cryovolcanism? Will there be
evidence for tectonism in the form of faulting or ridge and groove
formations? Will there be layered terrains? Will there be evidence for
atmospheric hazes or for surface winds forming dunes (dunes on Pluto
may be mostly grains of nitrogen ice)? Pluto is known to have an
active surface, with changes in surface colors and reflectances
observed by Earth-based telescopic monitoring. LORRI's high resolution
images will reveal features as small as 100 m on Pluto (260 m on
Charon).

Charon is Pluto's giant satellite: at about half the size of Pluto, it
is larger than any other planetary satellite relative to its
primary. Unlike Pluto, Charon has no detectable atmosphere, and it
probably has an old surface that may preserve a cratering record from
collisional evolution within the Kuiper Belt. LORRI data will play a
critical role in determining the crater size distribution and
morphologies on Charon. Equally important, LORRI images will provide
precise measurements of the shapes and sizes of both bodies. In
addition, LORRI will obtain high resolution images of the newly
discovered moons Nix and Hydra (Weaver et al. 2005), comparable in
terms of resolution elements across the illuminated disk to the
Galileo images of Gaspra.

After the Pluto-Charon encounter, New Horizons will make the first
visit to one or more Kuiper Belt objects. Owing to the likely small
size of these targets, LORRI's high resolution is especially important
to capture as much surface detail as possible. Will these Kuiper Belt
objects look like the asteroid Eros, or will there be bizarre surface
features like the flat-floored, steep-walled depressions (craters?)
found on the nucleus of comet 81P/Wild 2?

The New Horizons mission has a long focal length, narrow angle imager
for several reasons. Pluto is a dwarf planet, and New Horizons flies
by quickly, so the encounter science observations occur within one
Earth day - but with LORRI, New Horizons will be able to image the
Pluto system at higher resolution than any Earth-based telescope can
(even the Hubble Space Telescope, or its successor in 2015) for 90
days prior to encounter. These images will provide an extended time
base of observations, for studies of the shapes, rotations, and mutual
orbits of both Pluto and Charon, and for characterizing surface
changes.

Moreover, Pluto and Charon both rotate at the same rate as for their
6.38 day mutual orbit, always keeping the same faces towards each
other. Hence, during the near encounter, which lasts less than an
Earth day, only one hemisphere of each body, that which faces New
Horizons, can be studied at the highest resolution. The opposite faces
of both Pluto and Charon are last seen some 3 days earlier, when the
spacecraft is still ~4 million km away. Despite this distance, LORRI
will obtain images with 40 km resolution. These will be the best
images of the portions of Pluto and Charon which are not visible
during the near encounter period.

Finally, we have not yet discovered the Kuiper Belt object(s) to which
New Horizons will be targeted after the Pluto-Charon encounter, and
extensive Earth-based observing campaigns are searching for potential
targets. However, even after discovery, the heliocentric orbits of the
targets cannot be measured with sufficient accuracy from Earth to
enable the New Horizons spacecraft to fly to them, unless the targets
are also observed directly from the spacecraft. The direction in which
the target is seen from the spacecraft is then used to steer the
spacecraft to the target by optical navigation. LORRI is expected to
play a key role, by making the first and highest resolution detections
of the Kuiper Belt target object from New Horizons, more than 40 days
before closest approach. 

The New Horizons instrument payload includes three imaging
investigations: LORRI, the focus of this manuscript; Alice, a
ultraviolet imaging spectrometer; and Ralph, a visible imager and
infrared imaging spectrometer. These optical instruments are all
approximately co-aligned to view a common boresight direction, and the
spacecraft will maneuver as required to provide pointing during the
various planetary encounters. The Ralph instrument includes a
multispectral imaging channel (MVIC), which is a four-color, medium
angle imager with a field-of-view (FOV) 5.7 degrees wide, scanned by
time-delay integration. LORRI provides complementary imaging data, as
a panchromatic, narrow angle (field-of-view 0.29 degrees) framing camera.

2. LORRI Requirements

LORRI is a long focal length imager (see Figure 1) designed for high
resolution and responsivity at visible wavelengths. LORRI will perform
its primary measurements while New Horizons approaches Pluto and its
satellite Charon in July, 2015, obtaining images from a closest
approach distance of 11100 kilometers to Pluto, at a fly-by speed of
13.77 kilometers per second. New Horizons performed a Jupiter flyby,
with closest approach on Feb. 28, 2007. During the Jupiter system
flyby, LORRI imaged the atmosphere of Jupiter, its ring system, and
several of its satellites. Finally, after the Pluto and Charon
encounter, the New Horizons spacecraft will be targeted to encounter a
Kuiper Belt object (KBO), at which LORRI will again obtain high
resolution images. 

Figure 1  (left) LORRI telescope assembly, showing SiC mirrors and
metering structure; (right) LORRI composite baffle and flexure mount
on test stand. 

LORRI is required to obtain high resolution, monochromatic images
under low light conditions. Other instruments on New Horizons have
multispectral and hyperspectral capabilities. At Pluto encounter, 33
AU from the Sun, the illumination level is ~1/1000 that at Earth, but
Pluto is an unusually bright object with a visible albedo of ~0.55. At
the Kuiper Belt object, likely to be encountered outside 40 AU from
the Sun, the illumination will be still lower, and the object will be
darker, with an albedo of typically ~0.1.

The main objectives of LORRI are: (1) obtaining high resolution images
of Pluto and Charon during the approach phase, including the
hemisphere that will not be observed during closest approach, (2)
taking images at closest approach with instantaneous field of view
(IFOV) of approximately 50 meters per detector element and (3)
obtaining optical navigation images required to support trajectory
corrections. LORRI's reflective telescope is a Ritchey-Chretien
design, with a field-flattening lens group near the focal plane. The
digital image is captured by a frame transfer CCD detector. The FOV is
0.29 degrees x 0.29 degrees; the 1024 x 1024 square detector elements
have a pixel resolution (or IFOV) of 4.94 microrad. A summary of the 
characteristics of the LORRI instrument appears in Table 2.

Table 2. Summary of LORRI Characteristics 

The LORRI instrument is mounted inside the New Horizons spacecraft,
the interior of which is designed to remain near room temperature. As
the telescope views cold space, and the CCD is designed to operate
below -70 degrees C, the thermal implementation of the system was a
challenge. In order to maintain optical performance, a material with
high thermal conductivity and low coefficient of thermal expansion
(CTE) was required. As a result, the LORRI optical telescope assembly
(OTA) has both primary and secondary mirrors and a metering structure
fabricated from silicon-impregnated silicon carbide (see Figure 1).

During the approach to Pluto, which occurs under high sun conditions
(small phase angle or Sun-Pluto-spacecraft angle), LORRI is required
to image the surface of Pluto at signal-to-noise ratio (SNR) > 100 in
single frames. The encounter geometry is such that near closest
approach to Pluto, where the highest resolution images would be
obtained, LORRI views regions near the terminator under low sun
conditions and still less illumination; here LORRI has the goal of
imaging at SNR > 20 in single frames. Likewise, at the Kuiper Belt
object LORRI has an SNR goal > 20.

The resolution requirement near Pluto closest approach is for LORRI to
resolve 100 meters per line pair at a distance of 10,000 km from the
surface. An IFOV required to be <5 microrad (Table 3) was derived, with a
modulation transfer function (MTF) goal of 0.05 for spatial modulation
at 1 cycle per 10 microrad. LORRI is designed to meet imaging requirements
not only at nominal operating temperature (as low as approximately
-100 degrees C for the telescope), but also at room temperature.

Table 3. LORRI Measurement Requirements at Pluto

For optical navigation, LORRI is required to be able to image a star
of visual magnitude V=11.5 at SNR>7 in a single 100 ms exposure, with
full width at half maximum (FWHM) >1 pixel. It is not desirable for
too great a fraction of the energy from a point source to be imaged
onto a single pixel, because stellar images become too
undersampled. LORRI has a 4x4 pixel binning mode, for which its
limiting magnitude requirement is V>17 in a single exposure of 9.9
s. This 4x4 pixel-binning mode will be used to search for the target
KBO and to perform optical navigation on approach.  A special
spacecraft guidance mode is available for the KBO search in which the
spacecraft will hold the target within the 4x4 pixel pointing
tolerance for 10 second exposures. At 40 AU from the Sun, LORRI is
predicted to be able to detect a 50 km diameter object, of albedo 0.04
and at phase angle 25 degrees, from a distance of 0.35 AU, more than 40 days
before the object would be encountered. This is ample time for
targeting of the spacecraft.

Observations of the Pluto system begin at least 90 days prior to the
July, 2015 encounter, with both Pluto and Charon already resolved (see
Table 4). Initial observations are planned to refine the orbits of
Pluto, Charon and the two newly discovered moons Nix and Hydra. Pluto
and Charon can be imaged in single frames or in 2x1 mosaics through at
least ten full orbits (orbit period 6.38 days) ending about 14 days
before Pluto closest approach (c/a), to refine orbital ephemeredes and
especially the eccentricity. In the last week prior to c/a, searches
for librations of Pluto and Charon are performed, where Pluto subtends
123 pixels one half rotation before c/a. Nix and Hydra are expected to
be detectable in 4x4 binned images about 90 days before c/a, and in
unbinned single frames within the last 14 days.

Table 4. LORRI Planned Observations of the Pluto System

The last full frame image of Pluto is acquired about 10 hours before
c/a, and two 3x3 global mosaics of the full illuminated disk are
acquired during the near encounter. The full illuminated disk
observations will be useful to construct global base maps, to
determine the global shape of Pluto, and to search for oblateness and
tidal bulges. Additional images near c/a are obtained at successively
higher resolutions but covering smaller portions of the illuminated
disk. The Pluto near encounter image dataset will be used to study
surface morphology, geologic processes, and atmosphere-surface
interactions. Likewise, the full illuminated disk of Charon will be
imaged with 3x3 LORRI mosaics at resolution of better than 0.5 km per
pixel, meeting the Group I panchromatic imaging requirement for Charon
(the corresponding 0.5 km per pixel imaging requirement for Pluto will
be met by MVIC).

The LORRI data sets acquired during the Jupiter encounter in January
through March of 2007 are summarized in Table 5. Images of the Jovian
atmosphere and its clouds and storms include full disk rotation
sequences, acquired up to two months before Jupiter c/a on February
28, as well as high resolution 2x2 mosaics of specific features
acquired near c/a. Imaging sequences were executed for each of the
Galilean satellites, including observations of the night sides of Io,
Europa, and Ganymede acquired while the respective satellites were in
eclipse (i.e., in Jupiter's shadow). The Jupiter ring system was
imaged at both low and high solar phase angles, as well as during the
ring plane crossing. Images to study the shapes and photometric
properties of Elara and Himalia, two of Jupiter's irregular
satellites, were obtained.

Table 5. LORRI Observations of the Jovian System 

3. LORRI Instrument Description

3.1 LORRI Overview

The LORRI instrument was designed and fabricated by a combined effort
of The Johns Hopkins University Applied Physics Laboratory (JHU/APL)
and SSG Precision Optronics Incorporated (SSG), of Wilmington,
Massachusetts, USA. LORRI has four subassemblies in close proximity
connected by electrical harnesses. These are the OTA, the aperture
cover door, the associated support electronics (ASE), and the focal
plane unit (FPU). Except for the door, all are mounted inside the
spacecraft on its central deck; the door is mounted to an external
spacecraft panel. LORRI is electronically shuttered, with no moving
parts aside from the cover door. The ASE implements all electrical
interfaces between LORRI and the spacecraft except for the door
control, several spacecraft thermistors, and two decontamination
heaters. Figure 3 is a block diagram of the instrument. Conard et
al. (2005) give a detailed description of LORRI design, manufacture
and test.

Figure 3. LORRI block diagram , showing subassemblies: optics (OTA)
and aperture door; focal plane unit (FPU); and associated support
electronics (ASE) with three slices, which are the low voltage power
supply, the event processing unit, and the imager input/output board.
 
A summary of LORRI instrument specifications is given in Table 6. The
LORRI boresight is required to be aligned within 0.1 degrees of the boresight
of the Ralph imager (MVIC). For a combined summary of the
fields-of-view for the three imaging instruments on New Horizons, see
the payload overview companion paper (Weaver et al. 2007).

Table 6. LORRI Instrument Specifications

The OTA, aside from the focal plane unit and thermal blanketing, was
designed and built by SSG Precision Optronics, Inc. The primary and
secondary reflecting optical elements are constructed of SiC. The
telescope is a 2630 mm focal length, f/12.6 Ritchey-Chretien
design. Three field-flattening fused silica lenses, located in front
of the focal plane unit, are the only refractive elements in the
system.

The SiC metering structure of the telescope holds the mirrors and
field flattening lens cell. It is a monolithic structure consisting of
a primary mirror (M1) bulkhead, short cylindrical section, and
three-blade spider with secondary mirror (M2) mounting. The field
flattener assembly mounts to the M1 mounting plate and protrudes
through the M1 mirror. The metering structure is mounted to the
graphite composite baffle using three titanium, vibration-isolating
feet. The baffle assembly is mounted to the spacecraft using six
glass-epoxy legs, which provide thermal isolation. The entire OTA is
covered with multi-layer insulation (vented away from the OTA), except
for the entrance aperture.

LORRI is protected from contamination and solar illumination using a
one-time-open door mechanism. The door is mounted to the exterior of
the spacecraft, and the LORRI baffle tube extends into the door to
form a contamination seal. The door is aluminum, with thermal blankets
for temperature control prior to deployment. The mechanism uses
redundant springs and redundant paraffin actuators for deployment. A
port allowing for installation of a witness mirror or small window is
also part of the door. The door was opened successfully in flight on
August 29, 2006.

The ASE provides the data and control interfaces to the spacecraft and
to the focal plane unit. It consists of three 10 cm by 10 cm printed
circuit cards electrically interfaced to one another via stackable
connectors. They are in a magnesium housing, mounted directly to the
spacecraft deck, a short distance from the OTA.

The LORRI FPU has a back-illuminated, thinned, high quantum efficiency
CCD (an E2V Technologies Model 47-20). The FPU consists of a magnesium
box mounted to the spacecraft deck, housing a 15 cm by 10 cm circuit
card that controls the frame transfer CCD and provides interfaces to
the imager board of the ASE. This circuit card is connected by a flex
circuit to a small electronics board mounted at the focal plane. The
small focal plane board holds the CCD itself and is mounted on thermal
stand-offs, allowing the CCD and the small board to operate at < -70 degrees
C while the magnesium box operates near room temperature. The CCD is
in a window-less mount to avoid scatter and multiple reflections, with
a black, anodized aluminum plate installed over the CCD storage
readout area.

3.1.1 Design requirements and trades

The stringent optical, thermal and structural requirements for the
LORRI OTA presented many design challenges. The primary design driver
for LORRI instrument was the resolution requirement as described in
section 2.2. The resolution is limited by a number of factors,
including the stability of the spacecraft while an image is being
exposed. The pointing stability of the spacecraft is characterized as
a typical drift rate of 25 microrad per second. The minimum exposure time
is limited by the frame transfer time of the CCD. In order to remove
completely the image smear which occurs during transfer, the exposure
time should preferably exceed the frame transfer time (approximately
13 milliseconds), although acceptable image quality has been achieved
in flight at exposures as short as 1 millisecond. LORRI was designed
for an exposure time range between 50 and 200 milliseconds, with 100
milliseconds the nominal design value. Over this exposure range, the
spacecraft orientation would drift ~2 to ~7 microrad.

After the range of exposure times was determined, IFOV had to be
traded. A smaller IFOV yielded higher resolution, though it would
ultimately be limited by spacecraft stability. Diffraction limited
system resolution when the entrance pupil became appreciably smaller
than 200 mm diameter. Strict mass limitations combined with cost
limitations prevented increasing the aperture much beyond 200 mm
diameter. An aperture of 208 mm was selected.

With reasonable assumptions about the type of telescope and detector
and their associated efficiencies, and with the range of nominal
exposure times, a nominal IFOV of 5 microrad was chosen, which defined an
effective focal length of 2630 mm. The FOV of the final design was
0.29 degrees square.

The aperture requirement drove the telescope to reflecting
optics. Mass and cost limitations, combined with the FOV and imaging
requirements, drove the design to a Ritchey-Chretien
design. Refractive elements were used as field flattening lenses, as
the Ritchey-Chretien focal plane curvature over the flat CCD would
have limited imaging performance without them. The field flattening
lenses allowed LORRI to meet the requirement of <0.1% distortion at
all points over the full field of view. There was no requirement for
color imaging. An optical ray trace layout for LORRI is shown in
Figure 4.

Figure 4. Optical Ray Trace Layout, showing primary and secondary
mirror surfaces, plus fused silica lenses.

During the New Horizons mission, LORRI is exposed to the following
radiation environments: a total dose of ionizing radiation 15 krad and
a neutron fluence (worst case) of 1.4 x 1010 n/cm2. The neutron
fluence is from the radioisotope thermal generator over the course of
the mission. In the OTA, the refractive optics, antireflective
coatings, and reflective coatings were required to be tolerant to the
specified radiation environments. The LORRI CCD is operated below
-70 degrees C to mitigate degradation of charge transfer efficiency from
exposure to the neutron fluence.

3.1.2 Optical design

The LORRI OTA is a Ritchey-Chretien design, with high system
throughput required because of the short allowed exposure time and low
light level at Pluto. The complete LORRI OTA design was evaluated with
a computer-aided design model including stray light analysis. Specular
reflections and bidirectional reflectance distribution functions of
the Aeroglaze Z-306 black paint, primary and secondary mirrors, field
group optics and focal plane were included in the model. The primary
and secondary baffle tubes were sized to minimize obscuration and
suppress direct paths to the FPA's active area. The telescope
magnification and obscuration were balanced, affecting the optical
sensitivity and MTF, respectively. Out-of-field stray light was
evaluated by generating point source transmittance curves with angular
scans across the boresight in two orthogonal directions (-70 degrees
to +70 degrees for each scan) to search for any obscured paths with 
unacceptable amplitude.

The OTA design required multiple baffle vanes, fabricated from
graphite composite material, surrounding the metering structure to
suppress stray light. In addition, an inner baffle is used extending
from the hole in the primary mirror. This inner baffle has both inner
and outer vanes plus threading in its interior. All baffle design
features were optimized through TracePro ray tracing analysis. This
analysis shows that out-of-field stray light is adequately suppressed
and that ghosting is acceptable. The derived OTA system
root-mean-squared (rms) wavefront error requirement based on the MTF
requirement was <0.10 waves @ 632.8 nm, over an operating
environmental temperature range of -125 degrees C to 40 degrees C and after
g-release. The system throughput requirement drove a surface roughness
requirement on all optics to a goal of <20 Angstroms rms.

3.1.3 Thermal requirements and design

The extreme thermal environment posed a design challenge. LORRI is
mounted within the interior of a spacecraft with deck temperature in
the range 0 to +40 degrees C temperature, while it has an open, 208 mm
aperture radiating into cold space. The challenge was to minimize
defocus due to gradients developed in the OTA metering structure. The
derived requirement was to limit the heat loss from the spacecraft to
the OTA to less than 12 W, requiring conductive isolation at the
interface and near zero radiative coupling to the spacecraft. The
gradients in the OTA metering structure were limited to 2.5 degrees C axial
and 1.0 degrees C lateral, and 0.5 W heater power was available for mirror
gradient control without any in-flight focus adjustment mechanism.

The thermal requirements were met by an athermal, single material
solution, which is self-compensating under soak conditions. SSG's SiC
55A formulation was chosen due to its inherent high conductivity,
which acts to minimize gradients, and its low coefficient of thermal
expansion, which minimizes the thermal strain impact of such
gradients. Invar 36, a good match to SiC 55A over the temperature
range of interest, was chosen for the metallic inserts that allow
bolting together of the OTA assembly. All invar inserts, as well as
the secondary mirror foot, were epoxy bonded to the SiC. In addition,
the OTA is mounted inside of the telescope baffle tube, made of highly
conductive K13C graphite composite. The baffle tube provides a
relatively uniform cold sink along the length of the telescope which
helps to reduce longitudinal thermal gradients.

Additional thermal control features on the LORRI telescope help to
reduce system gradients. The telescope is mounted to the spacecraft
via G-10 isolators which isolate the OTA conductively from the
spacecraft deck. Covering the appropriate OTA and spacecraft surfaces
with multi-layer insulation minimizes radiative coupling between the
OTA and the spacecraft deck. The LORRI multi-layer insulation, which
represents 15% of the instrument mass, consists of 23 separate pieces.

Areas where material mismatches occur are flexured or otherwise
configured to minimize induced strain and resulting optical
degradation. The secondary mirror is mounted to a flexured Invar 36
mount plate, which in turn mounts to the end of the SiC metering
structure. The secondary and primary mirror magnesium baffles are
flexure mounted to the OTA structure (SiC) and the lens cell (Invar
36), respectively. The system aperture stop is made of aluminum
6061-T6 and is flexure-mounted to the middle ring of the OTA
structure. The CCD mount plate is attached to the OTA structure by
titanium flexures which have the dual purpose of mitigating thermal
strain and thermally isolating the CCD from the structure. The G-10
mounts have titanium post flexures on either end. Additionally, the
OTA itself is mounted to the K13C baffle tube via titanium isolators,
which act to mitigate thermal strain and vibrations and to provide
conductive isolation.

The baseline mission requires decontamination of the CCD at
temperatures >-18 degrees C before door opening, with decontamination
heaters. Decontamination can be achieved at lower temperatures, but
long time periods may be necessary to desorb water.

An additional thermal requirement is that the CCD must be maintained
at a temperature of < -70  degrees C while acquiring science data. This
requirement is met by mounting the CCD to a bracket which is mounted
to the OTA via conductively isolating titanium flexures; the CCD
bracket is in turn attached to a gold-coated beryllium S200F
conduction bar that is bolted to a gold-coated beryllium S200F thermal
radiator whose outside surface is painted white with Aeroglaze
A276. Since the radiator is mounted to a separate spacecraft panel
from the OTA itself, there must be some compliance to allow motion
between the two, provided by a highly conductive 1100 series aluminum
alloy S-link.

The LORRI in-flight temperatures were predicted via the creation of a
finite difference thermal model that included all conductive and
radiative heat transfer. A thermal balance test of the instrument was
performed that validated the thermal model in five separate test
cases. The results of thermal balance testing predict that the 0.5
Watt gradient control heater will not be required in flight.  Flight
data shows a gradient of 1.5  degrees C, well within the 2.5 degrees C requirement
for gradient.

3.1.4 Telescope mechanical requirements and design

The structural requirements of the OTA were geared towards achieving
minimum mass, while maintaining performance over operational
temperatures and allowing for the stiffest design that will survive
the launch environment. The requirements were a maximum mass of 5.64
kg, a minimum resonant frequency of 60 Hz, and survival under
launch-induced vibration and stresses.

The inherently high stiffness-to-weight ratio of SiC, (~4.5 times that
of aluminum) allowed the fabrication of a low-mass structure with a
light-weighted primary mirror of open back, hub-mounted design to
minimize weight. The main baffle tube was fabricated from graphite
composite (K13C2U, M55J, and T300), another very high
stiffness-to-weight material (~2.5 to 4 times that of aluminum). The
smaller internal baffles were fabricated from a light-weight magnesium
alloy (ZK60A).

Vibration isolation was required to survive the launch
environment. Titanium isolators were incorporated to mount the
structure to the main baffle tube at three points approximately at its
center of gravity location. Another important structural design
consideration was to minimize any potential for mount-induced
distortion of the optics. Intimately connected to this design
consideration is the requirement for the OTA to mount to a surface
with only moderate mounting coplanarity, namely, a spacecraft aluminum
honeycomb panel. To avoid degradation of optical quality, a 3-point
mount was adopted, with the bases of each of the three mount locations
on the OTA outfitted with a ball joint that can be loosened and
re-tightened if necessary.

At the interface between the main baffle tube and the OTA inner
assembly, the vibration isolators also act to mitigate any
mount-induced strains. The flexurized mount plate at the secondary
mirror serves the dual purpose of controlling thermally induced
distortions, as well as mount-induced distortions; the CCD flexure
mounts also serve this dual purpose.

Due to its mass, the primary mirror was not flexure-mounted, as the
low resultant frequency and dynamic responses would have increased the
risk to the OTA under vibration. Instead the primary mirror is
hub-mounted, with a post and a foot bolted via three invar inserts to
a mount plate. The mount plate is in turn bolted to the
structure. Because the lens cell is made of Invar and mounts to the
structure, close to the primary mirror, the mount plate helps to
separate any induced thermal strain in the structure from being
transferred to the primary mirror.

3.1.5 Instrument integration, focus and alignment

When the OTA was assembled at SSG, a convex spherical reflector was
centered at the focus using interferometry. Metrology, combined with
knowledge of the shim size used to connect the reflector to the
carrier plate, determined the location of the focus relative to the
interface location on the carrier plate. SSG also provided a reference
mirror on the back of the secondary mirror mount, such that the
telescope line-of-sight was parallel to the reference mirror's
normal. Two optical reference flats at right angles to the
line-of-sight were mounted to the LORRI metering structure for use in
alignment monitoring after mounting to the spacecraft.

The depth of focus for LORRI at the detector, with mechanical
tolerances on the CCD, allowed initial shim sizes to be selected for
system focusing at the SSG-provided focus location. A 300 mm aperture,
f/5 off-axis parabolic collimator was used to project a point-like
image into LORRI. This image was produced by a laser unequal path
interferometer (LUPI) at the collimator focus. A series of exposures
were made at nine points in the LORRI field by moving within the field
using a fold mirror. At each location, the spot was centered on a
detector element by viewing the live image from the CCD through ground
support electronics, and balancing the wings of the image
symmetrically about the center detector element. This was repeated for
seven focus adjustments of the collimator, with slight measured
changes from nominal to allow for deterministic shimming of
LORRI. Data from these exposures were examined to find the best LORRI
focus versus the collimator adjustment. New shims were then installed
to move the plane of the CCD onto the plane of best focus.

Once at nominal focus, the CCD was centered on the optical axis by use
of a theodolite viewing both the reference mirror on the back of the
secondary mount and the primary mirror. The theodolite was
autocollimated on the reference mirror, and the azimuth and elevation
recorded. The theodolite then viewed the four corners of the CCD off
the primary mirror, and these azimuth and elevation values were
recorded and averaged. The CCD location was shifted in the plane of
best focus such that the average of the four CCD corners was within
tolerance of the normal to the reference mirror.

After the best focus was found in ambient conditions, LORRI was
installed in the NASA Goddard Space Flight Center's Diffraction
Grating Evaluation Facility for a focus check at flight-like thermal
and vacuum conditions. This was done by viewing a collimated beam,
which projected a small point-like image into LORRI, and stepping the
spot over the field by tilting LORRI on a gimbal platform.

Prior to delivery to the spacecraft, LORRI's line-of-sight was
measured relative to the optical reference flat mounted to the back of
the secondary mirror support. Additionally, the roll angle was
measured by viewing the CCD corners, and referencing to the orthogonal
alignment mirrors. These data, combined with measurements performed
referencing the flats to the spacecraft coordinate system, showed that
LORRI's line-of-sight was within mission requirements. Tracking of the
LORRI alignment references through the spacecraft environmental test
program did not show any significant movement relative to the
spacecraft coordinate system.

3.1.7 Contamination control

The LORRI telescope assembly remained under a nitrogen purge during
all phases of integration and test until launch, except for limited
times when images were taken or when put under vacuum. The internal
cleanliness requirements for the LORRI OTA per the LORRI performance
specification are a beginning-of-life specification of 250 A/2 and an
end-of-life specification of 300 A per MIL STD 1246C. The resulting
predicted reduction in optical efficiency (Conard et al. 2005) is <
4%.

3.2 Electronics

As shown in Figure 3, LORRI electronics consist of the ASE and
FPU. The ASE contains three printed circuit cards. These are the low
voltage power supply (LVPS), the event processor unit (EPU), and the
imager input/output (IM I/O). The ASE is the primary interface between
the spacecraft and the FPU, which mounts and controls the
CCD. Additional information can be found in Conard et al. (2005).

3.2.1 Focal plane unit

The LORRI FPU is required to read out a complete image in 1 second,
with the charge level in each pixel represented by a 12 bit binary
word. A highly sensitive CCD with antiblooming was required, leading
to the choice of the E2V Technologies CCD47-20. This is a 1024x1024
pixel frame transfer CCD with 13 micron square pixels. This device is
back-illuminated for high quantum efficiency and has a frame transfer
time of 13 milliseconds. The FPU noise is required to be <40 electrons
per pixel, well above the CCD read noise which is calculated to be
about 10 electrons at the readout time of about 0.7 microsecond per
pixel. In general, exposures of 50 to 200 milliseconds are typical for
LORRI, although the maximum exposure time is 29.9 seconds. The FPU
design includes switchable 4x4 pixel on-chip binning. The FPU also
includes two small incandescent bulbs that can illuminate the CCD
through multiply scattered light, so that testing can be performed
even when imaging is not possible through the optics (e.g., when the
cover door is closed).

All CCDs have a number of clocks that must be driven to specific
levels for satisfactory operation. The E2V CCD uses three phase clocks
for image zone, memory zone and line transfer. These clocks are highly
capacitive, particularly for the image and memory areas of the chip,
and they have capacitive coupling between different phases. LORRI uses
Micrel MIC4427 drivers which are designed to drive high capacitance
loads at the required voltage levels from logic level inputs. They are
switching, not linear, devices so that low and high voltage levels are
obtained by suitable choice of supply voltages, and transition rates
must be adjusted at the output. This is done with series resistance
which adds to the internal switch resistance of the drivers, forming a
simple time constant with the capacitance of the CCD phase. The CCD
requires 29 volts bias for the output field effect transistor. A
charge pump with pre- and post-regulation was used to generate this
voltage.

The LORRI FPU uses an Analog Devices AD9807 integrated circuit that
performs correlated double sampling, signal amplification, and analog
to digital conversion to 12 bits at maximum rates of 6 MHz,
comfortably above the pixel readout rates which are close to 1.5
MHz. The AD9807 is susceptible to latch-up from ionizing radiation in
space, and the LORRI FPU incorporates latch-up protection
circuitry. The CCD output is low enough for the AD9807 amplifier to
contribute significant noise, so a low noise, wide band operational
amplifier is added between the CCD and the correlated double sampler,
and the AD9807 is run at low gain.

3.2.2 Associated support electronics

The LORRI EPU controls the instrument via interfaces to the LVPS and
IM I/O boards. The EPU communicates to the spacecraft using an RS-422
link, which receives commands and transmits engineering data. The EPU
uses a RTX2010RH processor and runs FORTH code.

The main function of the IM I/O board is to receive serial image data
from the FPU and transmit that data to either of two Integrated
Electronics Modules (IEMs) in the required format. There are dual,
redundant IEMs on the New Horizons spacecraft, which provide command,
data handling, and telemetry functions. One IEM is active and one is a
back-up; LORRI provides both interfaces (LVDS and RS-422) to both
IEMs. Secondary functions of the LORRI IM I/O board include the
ability to:  store and transmit the image header, receive commands
from the RTX processor, calculate a 32-bin histogram, generate test
patterns without an FPU present, command the FPU mode and exposure
times based on input from the RTX. 

The Imager I/O board contains two field programmable gate array (FPGA)
designs. The first is called the imager-interface FPGA, and the second
is called the RTX-bus FPGA. The main function of the imager-interface
FPGA is to read images from the FPU and send them to the IEM. As an
added feature, the IM I/O board can generate test pattern images
across the IEM interface without an FPU present. The first pattern
consists of a horizontal ramp and the second pattern consists of a
vertical ramp. The imager interface FPGA can also receive data from
the RTX. These data are sent across the ASE backplane and through the
RTX-bus FPGA. The data are used to set the FPU mode and exposure time,
to set the active IEM low voltage differential signaling port and to
write the 408 bit header. The FPU mode data is transmitted to the FPU
across the pixel data signal at the beginning of each second. The
header data replaces the first 34 pixels when data is sent to the IEM,
to provide redundancy in associating instrument engineering data with
each image. The instrument engineering data are transmitted by the
spacecraft within a separate data stream from the science data, and
the engineering data must be associated with individual images in
ground processing. For LORRI, the critical header information is also
encoded into the images themselves, at the cost of 34 pixels in the
first row of each image. 

The RTX-bus FPGA also calculates a 32-bin histogram of the FPU image
data currently being transmitted. This histogram is then made
available to the RTX for future exposure time calculations. The RTX
FPGA also collects the FPU status and temperature data, making it
available to the RTX.

The LVPS provides 2.5 V, 6 V, and 15 V power as required by the other
boards within the ASE and by the FPU. The input voltage from the
spacecraft is 30+/-1 V. The LVPS also provides for current, voltage, and
temperature monitoring via an I2C serial interface to the EPU
board. The LVPS board provides switching to control power on/off to
the FPU and the telescope trim heaters. 

3.2.3 Flight software

The LORRI RTX-2010 processor shares a common design with that of the
PEPSSI instrument on New Horizons (see accompanying paper). This
common design extends into the software. The common flight software
provides packet telemetry and command handling services. Besides
handling LORRI-specific packets, the common software automatically
generates a variety of standard packets, including
housekeeping/status, command echo, memory dump, and alarm
packets. Similarly, besides handling LORRI-specific commands, the
common software also handles standard commands for memory loads and
memory dump requests. The command handling software also provides
storage and execution of command sequences. The common software has
timekeeping, voltage and current monitoring, and memory management
services and a standard boot program.

The LORRI-specific flight software controls heaters, collects
voltages, currents, and temperatures from the LVPS board, and manages
the FPU. In the FPU, the software controls the exposure time, either
by manual command or automatically based on the hardware-provided
image histogram, generates an image header, and enables routing of the
image to the spacecraft. The software also controls the FPU's test
patterns and calibration lamps. Whenever the software has nothing to
do, it reduces the processor's clock rate to save power.

3.3 Laboratory Test and Calibration

The LORRI instrument was subjected to an environmental qualification
program. Performance and environmental testing was performed at both
the subassembly and instrument levels. Typically, performance tests or
calibrations were done before and after environmental testing.

3.3.1 Subassembly Test and Calibration

The OTA was tested by SSG. Wavefront testing was performed using a
LUPI operated in double-pass mode. Wavefronts were measured before and
after vibration test and during thermal vacuum test. No change was
detected due to the vibration test. Some change was noted at cold
temperature during thermal vacuum testing, but it was determined that
the level of change was acceptable within the LORRI performance
requirements. As a secondary verification of performance, modulation
transfer function testing was also performed by projecting small
features into the OTA, and recording highly magnified images on a
non-flight detector. This testing was performed only in ambient
conditions and showed excellent correlation to wavefront data.

The FPU and CCD were also subjected at JHU/APL to testing at the
subassembly level. They were calibrated at predicted operating
temperatures and then subjected to environmental testing. The
calibration of the FPU and CCD was not repeated after the
environmental series due to time limitations.

3.3.2 Instrument Level Calibration 

Results of the LORRI instrument-level laboratory calibrations are
discussed in detail by Morgan et al. (2005). The instrument-level
optical calibrations under ambient conditions were performed at
JHU/APL. Due to schedule conflicts with another instrument development
program at JHU/APL, LORRI instrument-level optical calibrations in
thermal vacuum were performed at the NASA Goddard Space Flight Center,
in the Diffraction Grating Evaluation Facility (DGEF).

The objective of radiometric calibration is to determine the
conversion from the observed signal in instrumental units to the scene
radiance in physical units. For LORRI, the radiometric calibration
equation relates the digital output S expressed as DN (12-bit 'data
numbers') to the scene radiance (or surface brightness) I, see
Equation 1, where subscripts x; y denote dependence on detector
position (i.e., line and pixel numbers), lambda on wavelength, T on
temperature, t on exposure time, phi on optical power at other points
within the FOV, and Phi on total optical power within the telescope. The
quantities and their units are: I the scene radiance in W m-2 sr-1
nm-1; S the observed signal in DN; Bias the electronic offset of the
CCD signal in DN; Dark the detector dark current in DN; Smear the
signal in DN acquired by a pixel as it is shifted through the scene
during frame transfer; Stray the signal in DN due to stray light; FF
the (dimensionless) flat field response which normalizes responses of
individual pixels using a uniform diffuse source; R the absolute
responsivity in [DN s^-1 pixel^-1]/[W m^-2 sr^-1 nm^-1]; and t the exposure
time in seconds.

Of the terms in the radiometric calibration equation, bias and smear
were characterized over the relevant parameters; dark is insignificant
at operational temperatures; stray light was tested at discrete
off-axis source angles; flat field response was measured, but there
were difficulties obtaining the desired 0.5% accuracy at all pixels
from ground-based data, and in-flight measurements are planned to
replace or correct the current flat field; and absolute response was
measured with sufficient accuracy for LORRI science goals. In addition
to radiometric calibration, instrument characteristics such as
detector read noise, the imaging point spread function (PSF),
effective focal length and field-of-view were determined during
calibration. The PSF was measured at several locations across the FOV,
and the field-of-view and read noise were determined.

3.3.3. Calibration setup

To confirm stability of LORRI's focus and response, key measurements
were conducted at the nominal operating temperature as well as
temperatures at the upper and lower ends of the operational range. The
telescope assembly temperature ranged from -97 degrees C to -60
degrees C during calibration, while the CCD temperature ranged from
-96 degrees C to -78 degrees C. At these low temperatures it was
necessary to operate LORRI in vacuum to
avoid condensation. The calibration was conducted at the NASA/GSFC
DGEF which operates as a large vacuum chamber for optical
calibrations. Calibrations at DGEF were conducted in July, 2004 prior
to instrument-level environmental (vibration and thermal vacuum)
testing of LORRI, and again in September, 2004 after environmental
testing, to establish stability of the calibration.

Figure 5 is a block diagram of the DGEF calibration setup for LORRI. A
light source outside the vacuum chamber was fiber optically coupled
into the chamber, feeding an integrating sphere. The integrating
sphere backlit a rotating target wheel that contained various
apertures such as pinholes and resolution targets. The target wheel
was at the focal plane of a 38 cm diameter collimator, the output of
which illuminated LORRI, which was mounted on a two-axis gimbal
approximately two meters from the collimator aperture. Between LORRI
and the collimator, a 30.5 cm diameter off-axis paraboloid (OAP)
reflector could be rotated into the beam to focus a pinhole image onto
a calibrated photodiode, providing a measurement of the integrating
sphere port radiance for radiometric calibration.

Figure 5. Calibration Facility at the NASA Goddard Space Flight Center

A 150 Watt, high pressure, ozone-free xenon arc lamp was used for the
LORRI calibration in order to provide a high radiance, solar-like
spectrum. Since there are significant spectral variations in response
over the LORRI bandpass, it is important to approximate the expected
spectral radiance distribution in order to infer broadband radiometric
response from calibration measurements with reasonable
accuracy. Pluto, Charon, Jupiter, and Jovian satellites have spectra
ranging from generally neutral to moderately red relative to a solar
spectrum.

The arc lamp output was collimated using a fused silica condenser and
transmitted through two six-position filter wheels, one with bandpass
(BP) filters and the other with neutral density (ND) filters. The
output then passed through an airmass zero (AM0) filter and was
refocused into a fiber optic providing input to an integrating sphere
inside the vacuum chamber. The AM0 filter attenuates the near-infrared
portion of the spectrum to make the overall spectral energy
distribution more solar-like. The BP filters were centered at 400 nm,
500 nm, 600 nm, 700 nm, and 850 nm. Bandpasses at full width half
maximum (FWHM) for the filters were 65 nm for the 400 nm filter, 119
nm for the 850 nm filter, and approximately 90 nm for the others. 

The radiance at the integrating sphere output port, spectrally
integrated over the LORRI bandpass, was 46 microW cm^-2 sr^-1. Figure 6
shows the source spectrum. Although the lines appear to dominate the
spectrum in Figure 6, they contribute a modest fraction of the total
energy in the spectrum because their widths are relatively narrow. The
calibration source spectrum is compared to the spectra of Pluto and
Jupiter, as well as LORRI's calculated spectral response in Figure 6
(Morgan et al. 2005). 

Figure 6. (left) Arc lamp calibration source spectrum and (right)
Spectra of Pluto, Jupiter normalized to unity at 607.6 nm and LORRI
response normalized to unity at maximum (Morgan et al. 2005)

The target wheel (Figure 5) contained image masks, including pinholes
ranging in size from 5 microns to 1000 microns, an NBS 1963A resolution target,
and an open position. The target wheel was located at the focal plane
of a Cassegrain collimator with a 38 cm aperture, 460 cm focal length,
and 0.5 degrees unvignetted field of view. The collimator aperture was large
enough to overfill the LORRI aperture without vignetting over the
LORRI FOV and gimbal motion range, with tolerance for easy
co-alignment. The two-axis gimbal was stepper motor-driven with a
Unidex controller at 63 steps per arcsecond to scan LORRI across
targets. A LORRI pixel subtends roughly one arcsecond. The reference
photodiode was a Hamamatsu S1336-8BQ, whose responsivity was
calibrated by the vendor at 10 nm intervals from 200 to 400 nm, and at
20 nm intervals from 400 to 1180 nm.

For thermal control during calibration, a shroud cooled by liquid
nitrogen was constructed surrounding the telescope. A small annular
shroud encircling the telescope inside the cold shroud was heated to
control the telescope temperature to the desired set point. The
thermal conductivity of the telescope structure was sufficient to
guarantee a uniform temperature distribution despite the uneven
heating. The shrouds were fixed, with LORRI gimbaled inside
them. LORRI viewed the collimator through a ten inch aperture in the
main cold shroud, which kept the shroud edge well out of the LORRI FOV
over the maximum gimbal motion range used. LORRI calibration
temperatures are listed in Table 7.

Table 7. LORRI calibration temperatures 

3.3.4. Laboratory calibration results

Results of pre-flight calibrations are summarized here, and
preliminary analyses of in-flight calibrations from instrument
commissioning are summarized in the next section. Pre-flight and
in-flight calibrations are fully consistent, considering limitations
of ground test equipment.

The detector read noise was estimated from zero exposure time
exposures taken in the DGEF at nominal temperature with no
illumination. One hundred of these dark images were acquired, and the
standard deviation of the signal of each pixel was evaluated. The
distribution of the standard deviations is plotted in Figure 7. The
most probable standard deviation is 1.2 DN, which is adopted as a read
noise estimate.

Figure 7. Read noise distribution, from 100 dark images measured
pre-flight. The estimated read noise is 1.2 DN.

The detector bias (signal at zero illumination due to electronic
offset) depends on the focal plane electronics board temperature. From
subsystem level tests, the bias in unbinned 1x1 mode is Equation 2 in
units of DN as a function of board temperature T in degrees C. The observed
in-flight board temperature range is +34 (+/-0.7) degrees C. The bias ranges
from 546 to 547 DN over that range. The bias in 4x4 binned mode is 2
DN higher. Detector dark current is negligible over the flight CCD
temperature range, for exposure times up to at least one
second. Planned exposure times are on the order of 100 ms. Also from
subsystem level testing, the FPU gain is 22 (+0,-0.5) electron/DN for
CCD temperatures of -50 degrees C and -70 degrees C and for board 
temperatures of 0 degrees C and 40 degrees C. This was determined from 
analyses of photon counting noise in flat field exposures using
constant illumination with a green light emitting diode. A gain
estimate of 22+/-0.4 electron/DN was also obtained with analyses of
noise in flat field exposures versus net signal with eight different
exposure times. The linearity of the FPU was found to be within +/-1%
over DN values from 900 to 3900, where the saturation level is 4095.

During calibration at the DGEF, point source imaging performance was
evaluated by observing a 3x3 grid of 5 micron pinholes. The grid target
was arranged with 4 pinholes near the corners of the FOV, 4 along the
edges, and one near FOV center. LORRI was gimbaled through a 6x6
subpixel grid of positions spanning approximately 1.5 pixels,
acquiring 10 images at each step. A fitting process was applied to the
images to estimate the sub-pixel irradiance distribution. Results
indicated a PSF with FWHM on the order of 2 pixels, with no
discernible dependence on position within the FOV. No significant
variation of the FWHM with temperature was observed, verifying focus
stability from ambient temperature to -100 degrees C. These tests were
adversely affected by diffraction by the collimator secondary
obscuration, which was larger than the LORRI secondary.

More accurate PSF measurements were obtained in bench tests at ambient
temperature at APL. A source was projected through a Laser Unequal
Pathlength Interferometer (LUPI) and then an off-axis parabola
collimator with an unobstructed aperture to generate a collimated beam
that completely filled LORRI's aperture. For LORRI tests, the beam was
reflected from an optical flat mounted with tip/tilt controls, which
allowed focus and PSF measurements at five positions in the LORRI FOV
(near the corners and near the center). Image analysis indicated a PSF
with FWHM of 1.5 pixels with little variation across the FOV, when fit
with a two dimensional Gaussian function. With the spot near pixel
center, ensquared energy in a pixel exceeded 0.3 at all four corners
and the center of the CCD.

Although LORRI flat field images were obtained during the
post-environmental calibration at the DGEF, the CCD was subsequently
cleaned by blowing ionized nitrogen across it. New flat field
measurements conducted at APL with a quartz halogen lamp (see Figure
8) showed that the number of pixels affected by particle shadowing was
reduced to 62 from over 150. These new flat field measurements were
conducted in air at room temperature. There is vignetting of
approximately 4% in the corners, consistent with expectations based on
the design of the baffles and field stop. Low amplitude, dark
"doughnut" rings (up to 1% amplitude) are seen from particles probably
on the field-flattening lenses, and dark spots (typical 20% signal
loss in single pixels) are seen from particles on the CCD. Comparison
with in-flight images of Jupiter and with calibration lamp images
shows that the particles have not moved since pre-launch calibration,
and no new particles have been introduced since launch.

Figure 8.  Provisional flat-field measured at APL: logarithmic stretch
from 0.95 (black) to 1.05 (white). Dust particles form 'donuts' and
dark spots. There is 4% vignetting in the corners, but in the upper
left corner and in the center of the field, there is excess brightness
from image ghosting which is not removed in preliminary processing.  

A central region of approximately 200 pixel radius is brightened by a
maximum of 1.5% by ghost images that result from multiple reflections
between surfaces in the field flattening lens group. There is also
excess diffuse brightness from image ghosting in the upper left corner
of Figure 8 such that the known vignetting is masked, and arc-like
ghost features have been removed for preliminary processing. The
ghosts are dominated by out-of-field illumination at the red extreme
of the LORRI passband, depending on the radiance distribution over
field angles just outside the FOV up to approximately 0.37 degrees
off-axis. The ghosts are strongly dependent on source spectrum and
will be characterized extensively with Jupiter observations. Also seen
in Figure 8 is fringing of 0.5% amplitude from constructive and
destructive interference in the CCD.

The DGEF flat fields showed no significant temperature dependence. The
maximum differences between the flat field at nominal temperature and
those at the hot and cold extremes were on the order of 1%, with fewer
than 250 pixels exceeding 0.5% variation in either
case. Bandpass-filtered flat fields showed somewhat more variation
with temperature, but for most of the filters the variations are not
statistically significant. Only 10 images were used for the filtered
flats, compared to 100 for the panchromatic flats. There appeared to
be significant temperature dependence in the 850 nm flat, possibly due
to interference in the CCD with the emission lines in the xenon arc
lamp spectrum. These strong lines do not occur in natural targets, and
the panchromatic flat fields are practically independent of
temperature for flight conditions.

However, laboratory flat field calibrations apply only imperfectly to
the flight system, because of limitations such as use of non-solar
source spectra. The provisional flat field of Figure 8 will be
replaced with in-flight flat field observations, using Jupiter as an
extended source during the Jupiter flyby, either to use directly or to
be combined with the laboratory flat field data. Additional in-flight
flat field data are planned using solar stray light.

The absolute radiometric response of LORRI was also determined at the
DGEF. When system responsivity varies significantly over the spectral
bandpass, as is the case for LORRI, the measured signal depends on the
scene spectrum, which will depend on the target. As can be seen in
Table 3, LORRI has requirements for relative, not absolute,
radiometry. However, the source spectra that LORRI will view in flight
will be measured by other instruments on New Horizons (namely Ralph,
see companion papers). With the aid of independent spectral
observations, LORRI can provide useful radiometric data.

LORRI's absolute responsivity as a function of wavelength was
determined (Morgan et al. 2005) using panchromatic absolute
measurements together with a calculated relative response
curve. First, the LORRI responsivity spectrum was calculated from
geometrical throughput of the optical design and from spectral
characteristics of optical components and the CCD. It was assumed that
an unknown constant factor would scale this response spectrum to the
correct absolute level. The calculated spectral responsivity accounted
for the measured mirror reflectance curves and lens transmissions, and
it used E2V's typical quantum efficiency curve for the model 47-20 CCD
which was not individually measured for the flight CCD. 

Second, the absolute LORRI response was estimated by scaling the
calculated spectral responsivity to match the DGEF flat field
observations. The flat field source spectrum was corrected to an
absolute scale using reference photodiode measurements taken
immediately before each set of flat field images. The 1 mm pinhole was
imaged onto the calibrated photodiode with the OAP mirror to obtain
the spectral radiance of the integrating sphere port (see Figure
6). Correction was made for signal reduction from the partial blockage
of LORRI's aperture by the collimator secondary, which is larger than
LORRI's secondary. The LORRI obscuration by spider and secondary is
11%. The LORRI signal was taken from the median flat field signal,
which was 15908 DN/s/pixel viewing this diffuse source spectrum. From
this signal, the absolute LORRI response curve shown in Figure 9 was
derived (correcting the figure shown in Morgan et al. 2005). The flat
field source was also observed through the BP filters of Section 3.3.3, and a
similar procedure was used to obtain the responsivity within each
bandpass, which was plotted at the signal-weighted mean wavelength for
each filter. The BP filter data suggest that LORRI's response peaks at
a shorter wavelength than would be inferred from the calculated
spectral responsivity. The absolute response will be refined in future
work using in-flight calibration data.

Figure 9. Absolute monochromatic response in
(DN/s/pixel)/(W/cm2/sr). Asterisks indicate bandpass filter
measurements, which agree well with the absolute response curve at the
three shorter wavelengths but lie below the curve at longer
wavelengths (Morgan et al. 2005). 

Lastly, the response to realistic scene spectra was calculated
numerically by integrating the product of the assumed scene spectra
and the LORRI response spectrum derived from calibration. From the
calibrated response curve, LORRI's responsivity to specified scene
spectra can be calculated by integrating the product of the LORRI
absolute response curve and the normalized scene spectrum. For a
source with the average Pluto spectrum, the absolute response of LORRI
is estimated from the DGEF calibrations as Equation 3 giving the
conversion from spectral radiance at the LORRI pivot wavelength
(Horne, 2004), which is 607.6 nm, to signal expressed in units of
DN/pixel/s. The spectral radiance in the denominator is understood to
correspond to the pivot wavelength. The absolute response depends on
the shape of the adopted scene spectrum; for additional details see
Morgan et al. (2005) and discussion of in-flight observations below.

3.3.5 In-Flight Calibration 

LORRI has successfully completed its in-flight commissioning
tests. LORRI's cover door opened on August 29, 2006, and successful
observations have been acquired of the open star cluster Messier 7 and
the prime mission target Pluto as well as the planets Uranus, Neptune,
and Jupiter. Instrument commissioning tests through September, 2006
are summarized in Table 8.

Table 8. LORRI In-Flight Commissioning and Calibration Test Summary

The commissioning test sequences LORRI-005, LORRI-006 and LORRI-025,
with the door still closed, were used to verify the FPU noise and bias
characteristics in-flight versus the ground calibrations. Results of
the in-flight tests have been consistent with the ground calibrations
of the FPU. For instance, the difference of two consecutive
zero-exposure time images, obtained from LORRI-005 on February 24,
2006, yielded a read noise estimate of 1.11+/-0.12 DN, where DN ~ 22
e-. The cosmic ray "hit rate" was also measured from LORRI-005, with
the rate of hits (mean +/- standard deviation) determined as 15.7+/-13.3
per frame for nominal exposures <0.5 s, where a "hit" was defined to
be a pixel with DN at least 5 standard deviations above the mean.

The result of preliminary in-flight measured bias correction, from
analysis of 25 frames acquired during the LORRI-006 test performed on
30 July, 2006, is shown in Figure 10. A total of 100 frames were
acquired in that test and are under analysis. The preliminary analysis
used 25 of the available dark frames, all of which are exposures of
zero seconds. For each of these dark frames, the median DN is found of
all the pixels in the four dark columns (which are the masked columns
at the right side of the image), and the resulting scalar is
subtracted from each pixel value in the image zone to obtain a
corrected dark frame. After removal of outlier pixel values (more than
3 standard deviations from the mean of the 25 values at the same row
and column numbers), the 25 corrected dark frames are then averaged to
obtain the so-called delta_bias frame which is shown in Figure
10. This frame shows the pixel-level variations of the bias and
demonstrates the excellent uniformity of the LORRI flight CCD. The
low-level vertical banding in Figure 10 is the result of amplifier
oscillation in the FPU; the maximum peak-to-trough amplitude of this
wave-like feature is ~0.8 DN. 

Figure 10. Delta_bias frame: in-flight measurement of FPU bias
correction, based on analysis of 25 frames acquired 30 July 2006. This
image is linearly stretched between -1 DN and +1 DN. 

The delta_bias frame is used in the LORRI pipeline data processing to
remove the FPU bias from a LORRI image as follows. First, the median
of the pixels in the four dark columns of the raw image is determined
and subtracted from each pixel in the image zone of the raw
image. Then, the delta_bias frame is subtracted to obtain the final
bias-corrected result. The statistics of delta_bias frame are: the
mean value in DN is -0.0049; the standard deviation is 0.26; the
maximum value is 1.64; the minimum value is -1.56. Since the pipeline
processing uses the absolute bias level estimated from the dark
columns of every image, the bias estimate from Equation (2) is not
used, although the latter provides an independent check.

The next step in LORRI pipeline processing is to apply the flat field
correction. Acquisition of in-flight flat field observations is
planned during and after Jupiter encounter in 2007. The in-flight
observations to date have been reduced with flat fields determined in
pre-flight laboratory testing, see Morgan et al. (2005).

Also required to complete the data reduction is removal of an effect
called readout smear. This arises from the operation of the frame
transfer CCD used in LORRI, where first the image zone is flushed,
then an exposure is taken, and finally the image is transferred into
the storage zone. Hence a pixel of the raw image is exposed to the
scene radiance from the corresponding geometrical element of the
scene, but it is also exposed to the radiances of all the scene
elements in the same image column during the image transfers. Thus the
raw image is the superposition of the scene radiance and the signal
acquired during frame transfers, which is called readout smear. 

The readout smear is removed as follows. Let P^meas_i,j = measured
image array in DN where i, j are the
column and row indices, respectively. Let the exposure time be written
T_exp, with the transfer times for the frame scrub T_f1 and the frame
storage T_f2 and with N the number of rows (which is 1024 for 1x1
images). Let T_favg be the average of T_f1 and T_f2 to define the
constant A in Equation 3.25.  The desmeared image is then Equation
3.5.  In-flight tests have verified desmear by this technique using
observations of Jupiter obtained at exposure times as short as 1
ms. For the image in Figure 11, T_favg = 10.5 ms and T_f1/ T_favg =
1.044, while  T_f2/ T_favg = 0.956.

Figure 11. Jupiter observed 9/4/2006. Panel a. Raw image MET 19683344
on logarithmic scale, showing readout smear. Panel b. Same image after
calibration on same logarithmic scale, showing bias and smear
removal. Panel c. Average of indicated columns showing different
readout smear on either side of Jupiter. Panel d. Same as c. for
calibrated image, showing removal of readout smear. Panel e. Image of
panel a. after calibration, on linear scale. Panel f. Five LORRI
images including that of panel a., all with 6 ms exposures, after
calibration and co-addition. Io and Europa (both barely resolved) and
their shadows are indicated.

In-flight photometric calibration was obtained using observations of
the open star cluster Messier 7 obtained from LORRI-007 on 29 Aug
2006. The visual magnitude is given by V = -2.5 log S + 18.94 + CC
with S in [DN/second]. We find the color correction CC = -0.06 for OB
stars, CC = 0 for FG stars, and CC = 0.4 for K stars.

The Messier 7 observations (see Figure 12) also confirm from stellar
images that the point source function of the LORRI system, including
the effects of spacecraft pointing jitter, is 1.8 pixels FWHM. Stars
to at least 12th magnitude are detected. The plate scale, based on
comparisons of pixel separations of bright stars in the field versus
cataloged positions, is 4.955 microrad/pixel. Any geometric distortion in
this image is less than 0.5 pixel.

Figure 12. LORRI image of open cluster Messier 7 obtained in
flight. This image has been logarithmically stretched. North is up,
east is to the left.

4. Summary

The LORRI instrument was successfully developed, qualified, and
calibrated, and it was delivered to the New Horizons spacecraft on
time and within budget. It has operated successfully in flight, and
its performance is completely nominal. It will obtain the first high
resolution imaging observations in the Pluto system and at one or more
additional Kuiper Belt objects.

Acknowledgments

We thank NASA and the New Horizons Project teams for support. We also
thank many more members of the LORRI team whose contributions have
helped to make LORRI a success.

References

S. Conard, F. Azad, J. Boldt, A. Cheng, K. Cooper, E. Darlington, ,
M. Grey, J. Hayes, P. Hogue,
K. Kosakowski, T. Magee, M. Morgan, E. Rossano, D. Sampath,
C. Schlemm, and H. Weaver, "Design and fabrication of the New Horizons
Long-Range Reconnaissance Imager," in Astrobiology and Planetary
Missions, G. R. Gladstone, ed., Proc. SPIE 5906, 2005.

F. Morgan, S.J. Conard, H.A. Weaver, O. Barnouin-Jha, A.F. Cheng, H.W. Taylor, K.A.
Cooper, R.H. Barkhouser, R. Boucarut, E.H. Darlington, M.P. Grey, I. Kuznetsov, T.J.
Madison, M.A. Quijada, D.J. Sahnow, and J.M. Stock. Calibration of the
New Horizons Long-Range Reconnaissance Imager, in Astrobiology and
Planetary Missions, G. R. Gladstone, ed., Proc. SPIE 5906, 2005.

Cheng, A.F. and Young, L.A., Mission Seeks to Explore Pluto-Charon,
EOS Trans. AGU, 83, pp. 384, 389, 2002.
K. Horne, The ways of our errors: optimal data analysis for beginners
and experts, star-www.stand.ac.uk/ kdh1/pub0/ada/woe/woe.ps,
University of St. Andrews, 2004.