(Note: the majority of this text was derived from that in
  [MALINETAL1991])
 
  Instrument Overview
  ===================
    The Mars Observer Camera (MOC) was initially developed as part of
    the Mars Observer instrument complement.  After the loss of MO,
    the MOC flight spare was completed and flown on the Mars Global
    Surveyor spacecraft.  To avoid a confusing change of acronym, the
    instrument on MGS is called the Mars Orbiter Camera, but is often
    referred to internally to the MOC project as MOC2.  Regardless,
    the two instruments are essentially identical.
 
    MOC is a three-component imaging system (one narrow-angle and two
    wide-angle cameras) designed to take high spatial resolution
    pictures of the surface and to obtain lower spatial resolution,
    synoptic coverage of the surface and atmosphere [MALINETAL1992,
    MALINETAL1998].  The cameras are based on the 'push broom'
    technique, acquiring one line of data at a time as the spacecraft
    orbits the planet.  Using the narrow-angle camera during the
    Mapping Phase of the mission, areas ranging from 2.8 x 2.8 km to
    2.8 x 25.2 km (depending on available internal digital buffer
    memory) can be imaged at about 1.4 m/pixel.  Additionally,
    lower-resolution pictures (to a lowest resolution of about 11
    m/pixel) can be acquired by pixel averaging; these images can be
    much longer, ranging up to 2.8 x 500 km at 11 m/pixel.
 
    The following table summarizes MOC characteristics.
 
 Camera         Min wave-  Max wave-  Focal   Aperture  F number  Reso-
                 length     length    length                     lution
                  (nm)       (nm)      (cm)                     (380 km,
                                                                m/pixel)
---------------------------------------------------------------- ------
 Narrow angle     500        900      350.0    0.35 m   10        1.5
 Wide angle red   600        630        1.1    1.7 mm    6.4      230
 Wide angle blue  420        450        1.14   1.8 mm    6.3      230
 ----------------------------------------------------------------------
 
 
  Scientific Objectives
  =====================
    High-resolution data will be used to study sediments and
    sedimentary processes, polar processes and deposits, volcanism,
    and other geologic/geomorphic processes.  The MOC wide-angle
    cameras are capable of viewing Mars from horizon to horizon and
    are designed for low-resolution global and intermediate
    resolution regional studies.  Low-resolution observations can be
    made every orbit during the Mapping Phase, so that in a single
    24-hour period a complete global picture of the planet can be
    assembled at a resolution of at least 7.5 km/pixel.  Regional
    areas (covering hundreds of km on a side) may be imaged at a
    resolution of better than 250 m/pixel at the nadir.  These images
    will be particularly useful in studying time-variable features
    such as lee clouds, the polar cap edge, and wind streaks, as well
    as acquiring stereoscopic coverage of areas of geological
    interest.  The limb can be imaged at vertical and along-track
    resolutions of better than 1.5 km.  Color filters within the two
    wide-angle cameras permit color images of the surface and
    atmosphere to be made to distinguish between clouds and the
    ground and between clouds of different composition.
 
 
  Calibration
  ===========
    MOC is fixed to the nadir panel of the spacecraft, and therefore
    cannot view any spacecraft-mounted calibration targets.  Through
    a combination of pre-launch calibration over temperature ranges
    anticipated at Mars, and occasional coordinated observations with
    TES, especially during regional or global dust storms when the
    atmosphere obscures the surface and presents a relatively
    featureless target, it is anticipated that MOC data can be
    calibrated photometrically to about <=10% in an absolute sense,
    although the relative photometric precision of the data (pixel to
    pixel) should be better than 3%.
 
 
  Operational Considerations
  ==========================
    The pre-mapping MOC observations were taken while MGS was in its
    elliptical capture and aerobraking orbits.  Imaging was only done
    for a few minutes immediately after periapsis, when the
    spacecraft was either tracking the nadir or performing the
    so-called 'rollout', a reorientation maneuver that returned it
    from the aerobraking drag pass to an Earth-pointed attitude.  (A
    handful of images were taken when the spacecraft was explicitly
    slewed to point at targets of interest, including the Viking and
    Pathfinder landing sites, Phobos, and the 'Face on Mars'.)
    Altitudes were at times lower than those expected in mapping, but
    the high velocity of the spacecraft usually precluded clocking
    the line array fast enough to produce images with a 1:1 pixel
    aspect ratio.  In this period, a wide variety of summing modes
    and line times were used to address imaging goals in the best way
    possible.  Global coverage with the wide angle cameras was not
    possible in this period; typically a small fraction of the MOC
    buffer was devoted to reduced-resolution WA images for regional
    monitoring, while the lion's share of the buffer was used to
    store targeted NA images.
 
    The remainder of this discussion describes observations planned
    from the mapping orbit.  Low resolution observations can be made
    every orbit, so that in a single 24-hour period a complete global
    picture of the planet can be assembled at a resolution of at
    least 7.5 km/pixel.  Of course, because the Mars Global Surveyor
    orbit is sun synchronous, this global picture shows how each part
    of Mars appears at approximately 2:00 PM local time.  From an
    imaging perspective, this global map also illustrates two
    principal limitations in the Mars Global Surveyor orbit: the
    absence of diurnal coverage (it samples Mars only at two times of
    day, one of which is unsuitable for visible imaging) and the poor
    illumination angle (chosen prior to the selection of a camera for
    the mission).  Relief will be less apparent in images taken at
    low incidence angles near the equator, but at higher latitudes
    (i.e., nearer the terminator), relief shading will become more
    prominent.  Shadows will be visible near the poles.
 
    The fields of view of both the NA and WA cameras, the nature of
    the Mars Global Surveyor spacecraft and mission, and specific
    aspects of the MOC design place certain constraints on viewing
    Mars.  The width of MOC NA frames is limited by the camera FOV to
    be about 2.9 km.  Even if MGS orbits were uniformly spaced, it
    would take over 600 days for the same location near the equator
    to be viewed twice.  Given the 687 nominal mission duration and
    the vagaries of the Mars Global Surveyor orbit that result from
    gravitational perturbations and atmospheric drag, MOC will be
    fortunate to pass over each equatorial area once.  At higher
    latitudes, of course, the number of opportunities to image a
    given location increases.  Unfortunately, the along- and
    cross-track orbit prediction uncertainties are larger than the NA
    FOV owing to mission operational constraints on Earth.  Thus,
    targeting a given feature will be a probabilistic activity.
    Contiguous 'mosaics' of NA images are not possible.  The WA FOV
    covers approximately 1300 km on the surface from nadir to each
    limb.  At the equator, MGS orbits are spaced about 1500 km, so
    the limb on one orbit is at the nadir on the subsequent orbit,
    and again at the limb on the orbit after that.  There is
    therefore good overlap at the equator, and excellent overlap
    closer to the poles.  Contiguous WA mosaics are possible.
 
    MOC image sizes are determined by the camera FOV (cross-track
    dimension, as noted above) and by the size and utilization of the
    12 MB digital buffer (along-track).  Using realtime compression,
    it is possible to acquire an image longer than the orbit
    determination along-track uncertainty.  However, other contents
    of the buffer may prevent its use in this manner.  Buffer space
    is thus an important resource that requires close management.
    For this reason, MOC images can vary in size and compression
    factor.  For the NA, areas ranging from 2.9 km X 2.9 km to 2.9 km
    X 25.2 km (depending on available buffer memory) can be imaged at
    about 1.4 m/pixel.  Additionally, lower resolution pictures (to a
    lowest resolution of about 11 m/pixel) can be acquired by pixel
    averaging.  Contingent as well upon available power, these images
    can be much longer, ranging up to 2.9 X 500 km at 11 m/pixel.
 
    Since payload selection and the Mars Observer project 'new start'
    in 1986, the data return from the MOC has grown significantly
    with better understanding of the spacecraft and ground
    telecommunication system.  Mission data rates, and the MOC data
    rate allocation, vary substantially over the course of the MGS
    mission, owing to Deep Space Network link performance.  They
    include recording at 3.5, 7, and 14 Kbps, and transmitting in
    realtime at 34.2 Kbps.  MOC data rates vary from 0.7 Kbps during
    the lowest record rate to 9 Kbps during the highest record rate,
    and 29 Kbps during realtime transmissions.  These rates permit
    the daily transmission of roughly 2, 4, and 8 uncompressed 2048 X
    2048 images during 'record only' days and, once every three days,
    14 such images during an eight hour realtime pass.  These numbers
    can be larger or smaller, depending on the compression factors
    used and the resolution of the global map that is being acquired
    at the same time.  MOC is capable of simultaneously sending data
    to both the record and realtime streams, and is further capable
    of matching any data rate available on-board the spacecraft.
 
 
  Detectors
  =========
    To minimize mass and complexity, and to increase reliability, all
    three MOC camera heads are line-scan imagers.  This allows the
    detectors to be electronically shuttered, eliminating the need
    for a mechanical shutter.  The substantial advantages of
    line-scan systems comes at the cost of fixed exposure times.
    This is not a problem for the WA (f/6, 13 lines per second), but
    is a substantial problem for the NA (f/10, 2200 lines per
    second).  In order to meet the signal to noise requirement of
    20:1 over an appreciable part of the planet, the overall signal
    chain noise performance must be better than 100 electrons.  This
    performance must be met at the raw acquisition pixel processing
    rate of 5 million pixels per second.
 
    In order to meet the noise and other performance requirements,
    two custom CCDs were fabricated for the MOC by Ford Aeronutronics
    (now Loral Aeronutronics): one for the MOC NA (2048 13 um pixels,
    two outputs) and one for the MOC WA (3456 7 um pixels, one
    output).  Development work also resulted in devices with thinner
    polysilicon layers than used in Ford's standard process,
    increasing their quantum efficiency in the blue (400 to 450 nm).
 
    The custom CCDs are individually packaged and mounted on custom
    hybrid ceramic substrates.  Each hybrid is attached to a small
    printed circuit board (PCB) that contains a small amount of
    pre-amplification circuitry inside the radiation shielding to
    minimize low level signal path lengths.  The hybrid and PCB are
    mounted together in a focal plane assembly that provides
    mechanical alignment of the detector to the optics, radiation
    shielding of the detector (down to <1 kilorad for the NA, <2
    kilorad for the WA), and thermal control.  Each FPA has a heater
    to keep it warm during the cold cruise to Mars, and each
    incorporates a thermal path to conduct away heat produced during
    operation.  In the case of the NA, the thermal path runs from the
    FPA to an 'internal radiator' behind the primary mirror.  This
    radiator loses its heat by radiation to the back of the primary
    mirror which, in turn, loses heat by radiation to Mars during the
    nominal mission, or to space during interplanetary cruise.  The
    focal plane assemblies are attached to the main analog circuitry
    through flex cables with multiple isolated return paths to
    minimize cross-talk.
 
    Based on breadboard measurements, the NA signal chain has a
    readout noise is less than 70 electrons, and a dark current noise
    of less than 3 electrons rms at its operating rate.  The WA
    signal chain has a readout noise of less than 15 electrons, and a
    dark current noise of less than 25 electrons at WA operating
    rates.  In-band quantum efficiency for the NA detector (i.e.,
    between 500-900 nm) is better than 35%, for the WA red detectors
    (575-625 nm) better than 35%, and for the WA blue detectors
    (400-450 nm) better than 10%.  Both detector designs have charge
    transfer efficiencies better than 0.999995.
 
    Combined with the optics described previously, the 2048 X
    1-element line array with 13 um pixels provides permits a NA
    resolution of 1.41 m/pixel from 380 km altitude and better than
    1.5 m/pixel over the entire range of operational altitudes, while
    the WA cameras, using the 3456 X 1-element line array with 7 um
    pixels, achieve nadir resolutions of 233 m/pixel (blue) and 242
    m/pixel (red) from 380 km altitude.
 
 
  Electronics
  ===========
 
    Analog Electronics
    ------------------
      Separate but similar processing chains are used for the NA and
      WA cameras.  Both chains provide control of gain to 3 db ('root
      2') resolution.  Offset control in 256 steps covering the range
      from 0 to 5 times the full scale of the analog-to-digital
      converter (A/D) is provided to subtract the effects of, for
      example, atmospheric scattering prior to A/D conversion.  All
      camera outputs are digitized to eight bit resolution.
 
      Correlated double sampling (CDS) is implemented in both cameras
      to attenuate reset and low-frequency noise.  In the WA this is
      performed by the fairly conventional means of using two
      separate sample and hold circuits, one holding the reset level
      and the other the video level, both driving the inputs of a
      difference amplifier.  Owing to the high speed of the NA (each
      of the two NA CCD outputs produces pixels at 400 ns intervals),
      a 200 ns delay line comb filter is used in each output chain to
      form the difference between the reset and video period, and
      this difference is further filtered and sampled before the
      offset and A/D operation.
 
      All variable gain is taken before the CDS, while the signal is
      effectively 'chopped,' providing immunity from offset drift in
      these amplifiers.  Monolithic transimpedance amplifiers are
      extensively used in the NA for their compactness and
      high-frequency performance.  A slow feedback loop maintains the
      average reset level of the signal at a fixed voltage, keeping
      the reset and video levels in a stable, linear range of the
      amplifiers while the clock feedthrough transients are removed
      by clipping stages to keep the amplifiers out of saturation.
 
      For each of the NA and WA, dual-tracking linear field effect
      transistor-series regulators, operating with about 300 mV of
      headroom, provide overall post-regulation and ripple rejection
      for the analog 10 V power rails.  L-C-R decoupling is used to
      isolate individual amplifiers.  These regulators provide
      overcurrent power cycling and current monitoring to the
      housekeeping telemetry.  They operate as switches to shut down
      the analog load by microprocessor command, to conserve power.
 
      Both NA and WA analog systems show long-term gain stability of
      better than 1% and gain stability over temperature excursions
      of better than 0.5% per 10 degrees C.  The NA system uses about
      700 parts and draws about 3 W; each of the WA systems uses
      about 450 parts and draws 1 W.  Noise through the WA analog
      system is 25 electrons, while noise through the NA analog
      system is less than 80 electrons.
 
 
    'Slow Digital' Electronics
    --------------------------
      The MOC digital system is divided into two parts: that
      constrained to operate at the clock speed of the microprocessor
      (the 'fast' digital electronics) and that which is not so
      constrained (the 'slow digital' electronics).  In order to
      isolate potential noise interactions between the 'slow' and
      'fast' electronics, and to minimize the amount of circuitry
      that had to be placed on one face of a 35 cm diameter PCB, they
      have been placed on separate faces of two-sided PCB assemblies.
 
      The MOC 'slow digital' electronics include the power supply,
      bus interface unit (BIU), housekeeping/engineering measurement
      circuitry, and power switching elements (PSE).  The power
      supply consists of two flyback-type switching converters, each
      capable of operating asynchronously at approximately 45 KHz but
      normally synchronized to a common 50 KHz clock.  Both
      converters share a common input filter.  One converter (the
      'digital' converter) operates whenever the spacecraft-provided
      28 VDC bias is present.  It provides +5 V to the digital logic
      and +20, +10 and -10.5 V to various sensitive analog loads (for
      which separate return lines are maintained).  The second
      converter (the 'analog' converter) also draws from the
      spacecraft 28 VDC power, but may be turned off using secondary
      control logic.  It supplies only analog referenced voltages:
      +22.5, +10.5, -10.5, +12, -5, and -20 V.
 
      Voltages from each of two 'slow digital' boards are 'diode
      or'd' to provide power redundancy to the analog and digital
      systems.
 
      The bus interface unit is Mars Global Surveyor Project-provided
      hardware that acts as the link between the MOC and the
      spacecraft Payload Data Sub-System (PDS).  Three functions are
      accomplished by the BIU: 1) it receives MOC uplink commands
      from the PDS (in the form of a serial bit-stream that is
      converted by the BIU to a 16-bit parallel format at the BIU/MOC
      interface), 2) it requests and receives from the MOC science
      and engineering data in 16-bit words on specific polling
      schedules, and 3) it provides a spacecraft timing reference
      signal (the Real-Time Interrupt, RTI) eight times a second.
      The PDS to BIU link is Manchester II encoded, and the BIU
      incorporates a custom encoder/decoder chip to decode commands.
      All three signals are carried over shielded differential
      twisted wire pairs.
 
      MOC housekeeping/engineering monitoring occurs for 48 points
      within each redundant system--34 voltages and currents and 14
      temperatures.  Test points are selected using analog switches;
      the voltage from a selected test point is fed to a voltage
      controlled oscillator (VCO) which is allowed to settle for two
      output cycles.  VCO cycles are counted by an A/D converter,
      with sampling initiated on an integer RTI and completed one RTI
      later.  The final count may be read immediately thereafter.
      Housekeeping telemetry is sampled four times each second.
 
      Owing to power constraints, the MOC implements power management
      circuity.  Power switching elements permit various sections of
      the MOC digital and analog electronics to be power-cycled to
      reduce orbit-average power consumption.  Some of these elements
      also incorporate circuitry to sense overcurrents that result
      from cosmic-ray induced single-event latchups (SEL).
      Power-cycling these sections enables the MOC electronics to
      avoid the deleterious effects of such latch-ups.
 
 
    'Fast Digital' Electronics
    --------------------------
      The MOC digital electronics incorporate many advanced
      technologies.  Three aspects of the all-CMOS digital
      electronics are highlighted here: the microprocessor, the three
      custom ASIC designs (on two chips), and the DRAM buffer memory.
      Figure 5 shows the layout of the 'fast digital' board,
      illustrating the relationships between the main elements of the
      digital design.
 
      The MOC control system is based on a radiation-hard version of
      the National Semiconductor 32C016 microprocessor developed by
      Sandia National Laboratories (the Sandia SA3300).  The SA3300
      is a 32-bit microprocessor with a 16-bit bus and a 24-bit
      address space.  The MOC operates its SA3300 at 10 MHz, giving a
      computational performance of just under 1 MIPS.  Among the
      advantages of this processor is that all flight software could
      be written in a high level programming language ('C').  All
      memory and control registers are located in the logical address
      space of the microprocessor.  All of its address space is also
      directly accessible by ground command through the PDS by direct
      memory load.
 
      All of the MOC digital design except the microprocessor,
      memory, and BIU is implemented in gate arrays.  Three basic
      designs are used, with two occurring on the same chip.  The
      three designs are called the Control Gate Array (CGA), the
      Buffer Gate Array (BGA), and the Narrow Angle Gate Array
      (NAGA).  Four parts (1 CGA, 2 BGA, and 1 NAGA) are used in each
      of the redundant half-systems.
 
      The CGA provides the link between the SA3300 microprocessor and
      the rest of the system by performing all address decoding and
      memory mapping.  It incorporates interfaces to the program
      memory and operational memory, three direct memory access (DMA)
      channels to the BIU, and the housekeeping/engineering VCO.  It
      also includes the reset/bootstrap controller and 'deadman'
      timer.  The CGA permits direct memory/register access without
      CPU intervention, allowing the MOC to continue to function
      should its microprocessors fail during the mission, albeit in a
      degraded fashion.
 
      The BGAs provide DRAM interfaces to each of two separate
      'banks' consisting of 6 MByte each.  Each BGA also provides WA
      CCD clocking, analog setting control, and downtrack summing for
      one of the two WA cameras, as well as WA data DMA.  Each BGA
      allows data DMA for both the Mars Balloon Relay (MBR) link and
      the NA camera, and applies ECC to the MBR data.
 
      The NAGA provides acquisition control and image processing
      capabilities for the NA camera.  It generates the NA clocking
      and acquisition control signals.  Image processing capabilities
      include cross-track summing, and one- and two-dimensional
      predictive compression, resampling, and Huffman encoding.  All
      data acquired through the NAGA may be optionally impressed with
      a 16-bit polynomial ECC syndrome capable of correcting 2 bit
      errors in 256.
 
      Each of the MOC's two half-systems incorporates more digital
      memory than has cumulatively flown on planetary missions:
      ninety- six 1 Mbit DRAMs arranged in two banks of 3 million
      16-bit words each.  The primary use of this memory is image
      storage and downlink data rate buffering, although
      approximately 1 MByte is allocated for flight software use
      (e.g., command sequences, scratch space, etc.).  Access
      bandwidth to the RAM buffer is sufficient to accommodate NA,
      WA, and MBR acquisitions, and memory refresh and PDS DMA
      transfers, all simultaneously and at their maximum rates.  CPU
      accesses are given lower priority and are delayed as necessary
      to allow other system activities the access they require.
 
      The DRAM devices are susceptible to both cosmic-ray induced
      single event upset and latchup (SEU and SEL).  Based on tests
      performed by the MOC design team, SEUs will probably occur less
      than 100 times each day under normal solar conditions.  The ECC
      applied to information stored in the DRAM was designed to
      repair such data corruption; the software refresh rate is
      selectable to match the actual upset rate.  Such procedures
      will not work during energetic solar flares, but these are
      expected to occur rarely during the Mars Observer mission.
      Additionally, every set of three devices (i.e., one bit in the
      16-bit word) has a separate power sensing element capable of
      detecting SEL-induced overcurrent, and power-cycling those
      parts.  As these devices use capacitive storage, the short-
      duration power-cycles are non-destructive (i.e., the memories
      retain their contents).
 
      The MOC includes two other memory banks--one each for program
      memory and operational memory.  The program memory consists of
      128 KBytes of ultraviolet-erasable programmable read-only
      memory (UVEPROM), while the operational memory consists of 192
      KBytes of radiation-hard static random access memory (SRAM).
      The flight software is stored in the UVEPROM, with ECC.  It is
      copied into the SRAM for execution.  A small amount of SRAM is
      also used as a high- speed line buffer for the NAGA.
 
 
    Flight Software
    ---------------
      The MOC flight software is very sophisticated.  It is capable
      of establishing the acquisition parameters of many different
      data types, while simultaneously conducting routine
      housekeeping and data compression tasks.  Among its
      housekeeping tasks is the continual refreshing of the DRAM
      buffer ECC.
 
      WA commanding constitutes the most complicated task of the
      flight software.  Owing to the desire to simultaneously acquire
      global monitoring, NA context, and local and regional WA
      coverage, a schema of 'virtual cameras' was implemented.  The
      BGA incorporates a 16-line 'ring buffer' into which WA
      observations are continuously acquired on each sun-lit pass;
      the flight software routes lines from this buffer to separate
      addresses, each representing a different acquisition or
      'camera.' Downtrack and cross-track summing can be performed
      separately from full- resolution acquisitions.  Different pixel
      positions along the detectors may also be selected so that, for
      example, both nadir 'context' frames and limb observations can
      be acquired at the same time without returning to Earth the
      full 3456 pixel FOV.
 
 
    Data Compression
    ----------------
      MOC has considerable capabilities to compress images to
      optimize use of its buffer and downlink bandwidth allocation.
      Three different forms of compression are available.  During
      acquisition of images (i.e., in realtime), a predictive
      compression technique can be employed, in either lossless or
      'lossy' form, to acquire more data than would otherwise fit in
      the MOC buffer.  As noted above, predictive compression is
      implemented in hardware for the NA camera; the same algorithm
      can be applied in software to WA data.  Depending on scene
      content, lossless compression factors from 1.5 to 2.5 are
      possible; with loss, the factors are larger but image quality
      degrades rapidly, primarily by contouring that results from
      insufficient bits to represent the actual image gray levels.
      Once otherwise uncompressed data are in the buffer, they can be
      compressed using two transform compression techniques: Walsh-
      Hadamard (WH) and Discrete Cosine (DC).  WH transform
      compression, which uses square-wave approximations of spatial
      frequency content, can be used to compress images by factors of
      3 to 5 with acceptable degradation, that appears mostly in the
      form of 'blockiness' associated with the loss of high frequency
      information.  DC transform compression, using cosine
      representation of spatial frequencies, provides the highest
      compression factors with the least degradation, but takes about
      30 minutes to compress a typical image (2048 X 2048).
      Compression factors of 10:1 show little visual evidence of
      degradation, although they may have increased radiometric noise
      (in the 1-3 DN range) within each 16 X 16 pixel transform
      block.  Visual degradation of DC transformed images typically
      focuses on the boundaries between the transform blocks, which
      become more obvious when these blocks contain different
      information owing to requantization of the higher frequency
      components.
 
 
  Optics and Filters
  ==================
    The NA camera is a 3.5 m focal length (f/10) Ritchey-Chretien
    telescope, filtered to operate in the 500-900 nm bandpass.  The
    WA system consists of two cameras with the same FOV, one
    optimized in the 400-450 nm ('blue') bandpass and the other in
    the 575-625 nm ('red') bandpass.  The blue WA camera (f/6.3) has
    an on-axis focal length of 11.4 mm, while the red WA camera
    (f/6.4) has an on- axis focal length of 11.0 mm.
 
    The MOC structure is approximately 88 cm in length and 40 cm in
    diameter, including four major components.  The largest component
    of the camera is the NA tube (called the Secondary Mirror Support
    Assembly or SMSA), which holds the secondary mirror spider and
    acts as both an optical baffle and the primary/secondary distance
    metering bench.  The SMSA is fabricated of pseudo-isotropic
    P75S/ERL1962 graphite/epoxy in a (0, 45, 90, 135) layup.
    P75S/ERL1962 graphite/epoxy was selected as the main structural
    material within the MOC primarily for its high strength and
    stiffness per unit mass and low density (leading to a low mass)
    and low coefficient of thermal expansion over a large temperature
    range (allowing it to maintain focus without active mechanisms or
    complicated matched-material thermal metering).  Attached to the
    side of the SMSA is the WA assembly, consisting of a mechanical
    support tube (the Wide Angle Support Assembly, or WASA) also made
    of P75S/ERL1962 graphite/epoxy with Invar fittings, the two WA
    optics and focal planes, and a WA baffle.  Beneath the SMSA is
    the P75S/ERL1962 graphite/epoxy Main Body Structural Assembly
    (MBSA), which supports the primary mirror, NA focal plane and
    associated electronics, and connects these to the SMSA.  Below
    the MBSA is the electronics assembly, housing three 2-sided
    electronics boards within an aluminum chassis/radiator assembly,
    connected to the MBSA by a T300/934 graphite/epoxy conical
    'flexure skirt.'
 
    Owing to the high magnification on the secondary mirror (7.75X)
    required to fold a 3.5 m focal length into 45 cm (17.8 inches),
    the NA optics are extremely sensitive to focus: a 1 um shift
    between the primary and secondary mirrors results in a 60 um
    movement of the plane of best focus, and the relatively fast
    optics provides only a 115 um depth of field.  Thus, the NA
    design is extremely sensitive to gravitational, vibrational, and
    thermal loads, and to moisture absorption, which arises from the
    hygroscopic nature of the epoxy.  Great care has been taken to
    minimize the effects of temperature variations on the camera.
    For example, the secondary spider vanes are arranged such that
    expansion or contraction of their lengths will rotate rather than
    despace, decenter or tilt the secondary mirror.  Adoption of a
    passive, graphite/epoxy-based structural metering system, and the
    uncertainties in its performance prior to actual testing of the
    Flight system, led to the inclusion of a bakeout heater to
    dehydrate the structure during cruise to Mars.
 
    Owing to their short focal lengths and compact design, the WA
    optics (each is a 9-element all refractive design) and their
    titanium housings are relatively insensitive to gravitational,
    vibrational, and thermal loading.
 
    The MOC optics/structure weighs 9.5 Kg, the primary mirror and
    its mechanical assembly accounting for 3.4 Kg (36%).  The primary
    mirror has been lightweighted and sags slightly under its own
    weight in a one-g field.  The WA lenses and housings, and the NA
    secondary mirror and its housing, together weigh 0.6 Kg.  The
    graphite/epoxy structures and associated fittings weigh 3.5 Kg,
    and the aluminum electronics housing and radiator and
    miscellaneous hardware, weigh 2.0 Kg.
 
 
  Operational Modes
  =================
    The Narrow Angle system has:
 
          16 gain settings
        2 56 offset settings
           8 summing settings (1 to 8)
         128 crosstrack size settings (16 to 2048 by 16)
      3 2768 downtrack size settings
         256 data compression settings
 
    The Wide Angle systems have:
 
          20 gain settings
        2 56 offset settings
          32 summing settings (1 to 32)
         128 crosstrack size settings (16 to 2048 by 16)
      3 2768 downtrack size settings
         256 data compression settings
           4 crosstrack summing modes
 
    The modes in effect for each image are described in keyword-value
    pairs in the header of each image.