ABSTRACT
  ========
    The dual technique magnetometer system onboard the Cassini orbiter is
    described. This instrument consists of vector helium and fluxgate
    magnetometers with the capability to operate the helium device in a scalar
    mode. This special mode is used near the planet in order to determine with
    very high accuracy the interior field of the planet. The four-year orbiting
    Cassini mission will lead to a detailed understanding of the Saturn/Titan
    system. In addition to the prime scientific measurement of the planetary
    field, the instrument will also make measurements of the planetary
    magnetosphere, and the interactions of Saturn with the solar wind, of Titan
    with its environments, and of the icy satellites within the magnetosphere.
 
  INSTRUMENT DESCRIPTION
  ======================
    The MAG instrument comprises a fluxgate magnetometer (FGM) and a
    vector helium magnetometer capable of operating in both vector and
    scalar mode (V/SHM). The instrument is intended to measure small
    changes in fields spanning four orders of magnitude with extremely
    high sensitivity. This goal is achieved in part by mounting the
    sensors on an 11-metre spacecraft boom; the V/SHM at the end of the
    boom, the FGM halfway along. The magnetometer boom distances the
    sensors from the magnetic field associated with the spacecraft and
    its subsystems, and especially from spacecraft-generated temporal
    field variations. Spacing the sensors at different distances along
    the boom allows the spacecraft fields to be better characterised and
    removed from the observations. However, mounting the sensors on a
    boom could result in their orientation with respect to the
    spacecraft axes changing from time to time, for example after
    spacecraft manoeuvres. A means of sensor-alignment determination has
    been provided by the Cassini project - the Science CAlibration
    Subsystem, SCAS. This system consists of two, perpendicular, coils
    rigidly mounted on the spacecraft body with a known alignment to the
    spacecraft axes. These coils produce well-defined magnetic fields on
    command which can be detected by the sensors and used to correct for
    any changes in sensor orientation.
 
    Both magnetometers are capable of measuring the magnetic-field
    vector at rates from 0 Hz up to 10 Hz (VHM) or at least 30 Hz (FGM).
    The VHM optimises low-frequency vector measurements in weak fields.
    The FGM is best suited to high-frequency measurements and can
    operate over an extremely wide dynamic range, from very weak fields
    up to Earth's strong field.
 
    The twin-sensor configuration contributes to overall instrument
    reliability; if one sensor fails, field measurements can be made
    with the other sensor, with sufficient performance to achieve many
    of the major objectives of the investigation. Reliability has been
    further increased by the provision of redundant instrument power
    supplies and data processing units, and by careful selection of
    electronic components that can survive the radiation environments
    encountered during the long cruise phase of the mission and in the
    Saturnian system. The VHM provides the stability needed to maintain
    calibrations, obtained in the solar wind, whilst Cassini is inside
    the Saturnian magnetosphere for long periods during the four-year
    tour.
 
    An implicit feature of scalar or resonance magnetometers are null
    zones which arise if the ambient field falls outside a cone of 45
    degrees half angle with respect to the optical axis of the
    magnetometer. These null zones result in the signal being
    dramatically weakened, causing the absolute accuracy of the
    instrument to suffer. When Cassini is inside 4 RS a requirement has
    been placed on the mission to avoid spacecraft orientations which
    cause the planetary field to lie within the null zones of the SHM.
 
    Other features of the instrument that have been driven by the
    characteristics of the mission and by the design of the spacecraft
    are to be found in the data processing unit (DPU). The DPU contains
    a bus interface unit (BIU), provided by the Cassini project for
    interfacing to the onboard data handling subsystem (CDS) bus. In
    line with the spacecraft design, the DPU is capable of handling
    Packet Telemetry and Telecommands, and features a flexible telemetry
    storage and generation scheme to support the multiple telemetry
    modes of the spacecraft. The Tour operations concept requires that
    the DPU is able to handle trigger commands which initiate multiple
    actions within the instrument (macro commanding). Further, in order
    to optimise the analysis of discrete events such as shock crossings,
    a snapshot capability has been implemented by which up to 16 Mbytes
    of data can be stored for later downlink at higher time resolution
    than normal. This capability can be initiated by command or
    triggered by pre-defined events.
 
    Magnetic-field information is also needed by other investigations on
    the spacecraft. To this end magnetic-field data are made available
    to onboard users every second. These onboard ancillary data are raw
    and uncalibrated vectors, the data source being selectable by
    command between the two sensors.
 
    In total, the instrument consists of the two boom-mounted sensors,
    subchassis #1 (an assembly containing electronics for the FGM, VHM
    and SHM, the heater-control electronics, the power supplies and
    power-management system) and subchassis #2 (an assembly containing
    the data processing unit). Both subchassis are mounted in bay 4 of
    the Orbiter upper equipment module (UEM). The instrument
    ground-support equipment was provided by KFKI and TUB.
 
    Table I lists the main characteristics of the instrument. Power and
    data-rate values vary according to instrument mode. The values given
    in the table are for the delivered flight model where power values
    include power drawn by the Cassini-provided BIU.
 
      TABLE  I
      Main Instrument Characteristics
 
      MASS
      V/SHM Sensor                                       0.71 kg
      FGM Sensor                                         0.44 kg
      Subchassis#1 (Power Supplies, Sensor Electronics)  5.15 kg
      Subchassis#2 (DPU)                                 2.52 kg
      Total                                              8.82 kg
 
      POWER
      Sleep Mode                                         7.50 W
      Vector/Vector Mode (FGM+VHM)                      11.31 W
      Vector/Scalar Mode (FGM+SHM)                      12.63 W
 
      NORMAL DOWNLINK DATA RATE
      FGM                                 32 Vectors per second
      VHM                                  2 Vectors per second
      SHM                                    1 Value per second
      Housekeeping                           24 bits per second
      Total                                2000 bits per second
 
      DYNAMIC RANGE, RESOLUTION
      FGM                                        +/-40nT, 4.9pT
                                               +/-400nT, 48.8pT
                                             +/-10,000nT, 1.2nT
                                             +/-44,000nT, 5.4nT
      VHM
                                                 +/-32nT, 3.9pT
                                               +/-256nT, 31.2pT
      SHM
                                          256nT - 16384nT,36 pT
 
 
    THE FGM
    -------
      The FGM sensor is mounted halfway along the magnetometer boom; its
      associated analog electronics form part of the electronics
      assembly on Subchassis#1. A cable of approximately 6.5-meter
      length runs along the boom between sensor and electronics. A high
      efficiency, tuned drive design of the electronics has been chosen
      to reduce power consumption and the effect of cable loading.
 
      The FGM is similar to the Imperial College instrument flown on
      Ulysses, and to many others flown on numerous missions. It is
      based on three single-axis ring-core fluxgate sensors mounted
      orthogonally on a machinable glass ceramic block. Ceramic is
      chosen for its low thermal expansion coefficient, minimising
      misalignments between sensors due to temperature changes. In each
      sensor, a drive coil is wound around a high-permeability ring-core
      which is completely enclosed in a sense winding. The drive coil is
      driven by a crystal-controlled 15.625 kHz square wave which is
      used to generate a magnetic field that drives the core into
      saturation twice per cycle. The three drive coils are connected in
      series to simplify the cabling and circuitry. The presence of an
      ambient magnetic-field component parallel to the axis of the sense
      coil causes the saturation of the core to become asymmetrical.
      This asymmetry induces a second harmonic of the drive frequency in
      the sense coil which is proportional to the magnitude of the
      magnetic-field component along that axis. The signal is processed
      through a narrow band amplifier tuned to the second harmonic of
      the drive frequency, which attenuates harmonics other than the
      second. The result is integrated, converted to a current and fed
      back to the sensor coil to null the ambient field. The integrated
      output voltage, amplified and corrected for scale factor and
      alignment errors, is proportional to the ambient field. The three
      analogue vector components are passed to the DPU for analogue to
      digital conversion and data processing. The noise performance of
      the FGM, measured on the ground at the analogue output of the
      electronics, is better than 5 pT/Hz at 1 Hz. The electronics can
      be checked in flight using an in-flight calibration (IFC)
      capability built into the electronics and controlled by command
      from the DPU. The IFC applies a fixed offset to each of the three
      vector outputs corresponding to a signal of approximately 10 nT.
      The frequency, number of on/off cycles, of the IFC is selectable
      by command.
 
      Changing the electronics feedback path and the output
      amplification allows the sensor to be operated in one of four
      different full scale magnetic field ranges, as listed in Table I.
      The largest range (+/-44,000 nT) was included mainly for ground
      testing in the Earth's field. Switching between ranges in normal
      operations is automatic, controlled by the DPU. If the magnitude
      of any of the FGM magnetic-field components exceeds an upper
      threshold for more than a specified number of samples, the DPU
      will switch the FGM to a higher range. Similarly, if all three
      component value magnitudes fall below a lower threshold for more
      than a specified number of samples, the DPU will switch the FGM to
      a lower range. All parameters are modifiable by command and
      autoranging can also be disabled and manual range changes
      commanded.
 
      A 1W heater has been provided to maintain the FGM within its
      operating temperature range of -30 to +50 degrees C. The specially
      designed, non-magnetic unit is mounted on the ceramic sensor block
      and has control electronics on Subchassis#1. Further thermal
      control is provided by an aluminised mylar-covered fibreglass case
      over the sensor block and by three, project-provided, radioactive
      heater units mounted at equal distances around the base of the
      sensor (these units provide a total of 3W).
 
 
    THE V/SHM
    ---------
      The V/SHM sensor is the flight-spare Ulysses vector-helium
      magnetometer sensor with an added small pair of coils nested
      inside the larger Helmholtz coils used in the vector mode. The
      sensor is mounted at the end of the 11-meter magnetometer boom. A
      set of cables running the length of the boom connects it to the
      VHM and SHM electronics on Subchassis#1. The VHM electronics box
      is also the Ulysses flight-spare unit with small modifications to
      change the sensor operating ranges and to compensate for the
      different boom cable lengths. A new electronics board has been
      added to Subchassis#1 containing the electronics to operate in the
      scalar mode.
 
      The operation of the magnetometer is based on field-dependent
      light absorption (the Zeeman effect) and optical pumping to sense
      the magnetic field. Helium in an absorption cell is excited by a
      radio-frequency (RF) discharge to maintain a population of
      metastable long-lived atoms. Infrared radiation (wavelength 1083
      nm) from a helium lamp, also generated by RF excitation, passes
      through a circular polariser and the absorption cell to an
      infrared detector. The absorption (pumping efficiency) of the
      helium in the cell is dependent on the ambient magnetic-field
      direction. The optical pumping efficiency is proportional to
      cos^2(Theta) where Theta is the angle between the optical axis and
      the direction of the magnetic field. This directional dependence
      is utilised in the vector mode by applying low-frequency sweep
      fields rotating about the cell which allow the extraction of the
      three orthogonal ambient-field components. These fields are fed
      back using a set of triaxial Helmholtz coils mounted on the sensor
      housing around the cell. In the scalar mode, the directional
      dependence results in a 'field of view' restricted to a cone with
      half angle approximately 45 degrees, centred on the optical axis
      detector.
 
      Changing the VHM sweep fields allows the sensor to operate in
      different ranges. Two VHM ranges have been selected for Cassini
      (see Table I). As for the FGM, automatic ranging has been
      implemented in the DPU. The VHM electronics also have an internal
      autoranging capability (used for the Ulysses instrument). A single
      range has been implemented for the SHM. Injection of known
      currents into the Helmholtz-coil system provide an in-flight
      calibration (IFC) capability. The calibration fields apply an
      offset of approximately 1/8 of the full scale range to each vector
      component. A non-magnetic proportional heater using up to 2W is
      incorporated into the V/SHM sensor and is controlled from
      electronics built into the VHM electronics box on Subchassis#1.
      The operating temperature range of the sensor is -10 to +40
      degrees C.
 
      In the scalar mode, a weak AC field at the Larmor frequency, which
      opposes the optical pumping, is applied to the cell. This field
      causes a reduction in the transmitted light  detected by the IR
      detector. The Larmor frequency, which is proportional to the
      ambient magnetic field, is measured. In order to track the ambient
      field the applied field is frequency modulated so that the
      detector output contains a signal component harmonically related
      to the modulation frequency. The proportionality constant is the
      gyromagnetic ratio which for helium is 28.023561 Hz/nT. Detection
      and measurement of the Larmor frequency leads to a very accurate
      measurement of the ambient field magnitude. The result is passed
      as a 20-bit scalar word from the SHM electronics to the DPU. A
      more detailed description of the V/SHM may be found in Kellock et
      al. (1996).  Smith et al. (2001) provides a detailed description
      of the SHM operation and observations from the Earth Swingby in
      August 1999.
 
 
  INSTRUMENT ELECTRONICS
  ======================
    The instrument electronics are all mounted in the Upper Equipment
    Module in bay 4 and are split between two subchassis assemblies.
    Subchassis#1 contains the sensor electronics, the power supplies and
    power management system, as well as the heater control and
    instrument housekeeping electronics; its mass is given in Table I.
    On the underside of the subchassis are the power-management board,
    the FGM electronics board and the SHM electronics board. The
    power-management board contains a total of 14 non-latching power
    switches and cross-strapping circuitry for the two redundant
    secondary power supplies. The top side of the subchassis contains
    the VHM electronics box, two small boards to the left of the VHM box
    with latching relays to switch between VHM and SHM operation, two
    redundant secondary power supplies (PSU1 and PSU2), a dedicated BIU
    power supply (PSU0), and the FGM heater-control electronics and
    housekeeping circuitry. Proportional control electronics for the VHM
    heater are located within the VHM box. The power supplies and
    switches are Imperial College designs used on previous missions and
    feature built-in overcurrent trips.
 
    The basic power distribution scheme is described in Kellock et al.
    (1996). Power switches for the secondary voltage lines are
    controlled by the active processing unit and power switches for the
    power supplies and processing units themselves are controlled by
    discrete commanding from the spacecraft via the BIU and the Common
    Core (CC).
 
    Subchassis#2 contains the DPU, consisting of two redundant processor
    systems plus a small CC and the BIU. When power is first supplied
    from the spacecraft, only the BIU and the CC become active, powered
    from PSU0. The BIU allows data transfer to and from the spacecraft,
    the CC processes commands and data for power up of the secondary
    power supplies (PSU1 or PSU2) and the processors (PUA or PUB). Each
    processor system is based on an 80C86 processor with 4-MHz clock,
    32-kByte PROM, 128-kByte Hi-Rel RAM and 16-MByte state-of-the-art
    commercial DRAM. The systems normally operate singly but can be
    operated in parallel. A high-accuracy 16-bit analogue to digital
    converter (ADC) is integrated into each processor system for sensor
    data collection. Two ADC clock speeds are available, 1 MHz and 2
    MHz, the former being the default speed. Tantalum shielding has been
    used for the ADCs, DRAMs and Operational Amplifiers to reduce their
    susceptibility to radiation.  The DPU boards are folded around the
    subchassis. The electronic components face the subchassis, because
    the 2.4 mm thick 16-layer boards provide additional radiation
    shielding. The Sensor Interface Board is on the right hand side.
    The flexible connection board goes through subchassis cutouts to the
    Processor Board on the other side. The JPL-provided BIU, plus its
    associated cabling, is located on the left side.
 
    To satisfy the demands of a deep-space mission with limited
    communications, the DPU has been designed with a large measure of
    autonomy and sophisticated data-handling functions. These functions
    include the following: telecommand handling, sensor autoranging and
    IFC, sensor data collection, sensor data processing, snapshot data
    handling, telemetry generation, error correction, fault detection
    and recovery, and onboard ancillary data generation. Some of the
    functions of the DPU are described below.
 
    The DPU is designed to handle both the packet telecommand standard
    adopted by Cassini, used for normal commanding, and discrete
    telecommands used when the processor is not active. A variety of
    command functions are supported and are discussed later. The DPU
    must be able to accept telecommands at all times, in all operational
    modes. Commands may be for immediate execution, or can contain
    relative or absolute timetags for delayed execution (relative
    timetags cause execution at a fixed time with respect to reception
    of the command, absolute timetags cause execution at specific
    spacecraft times). As noted earlier, a macro commanding capability
    has been implemented whereby sequences of instrument commands can be
    stored in the DPU and the sequence started by executing a macro
    command.
 
    Hamming single-bit error correction and double-bit error detection
    is provided for all memory devices except those in the BIU. Single
    Event Upsets (SEUs) can change the content of memory cells, cyclic
    access to every memory cell corrects single-bit and reduces the risk
    of double-bit errors. Memory scrubbing is initiated every 64 seconds
    in the Hi-Rel RAM and the 16-MByte, multi-snapshot, DRAM. It takes
    about 1 hour to scrub the complete RAM. Additional memory checks
    can be initiated by command: occurences of single and double-bit
    errors are monitored. The PROMS are checked separately and contain
    a pre-defined error pattern for detection. If a permanent RAM
    problem arises, the DPU can be commanded to run its software
    directly from PROM. In-flight tests have also been implemented for
    the ADC to check the noise and conversion and settling times on each
    analogue channel.
 
    Both processor systems contain four separate dual-level latch-up
    detectors, one for the processor, one for the multi-snapshot memory,
    and one each for the ADC +/-12V supply voltages. Detectors of
    similar design have been flown on the GEOTAIL, WIND and SOHO
    spacecraft. If a latch up is detected in the processor, it will be
    immediately switched off and on again and the instrument will be
    automatically reconfigured into its previous mode. Latch ups in
    either the memory or the ADC will cause it to be immediately
    switched off and back on again. There is also a hardware watchdog
    function in the DPU which will detect problems in the DPU program
    flow and which initiates a hardware reset followed by an automatic
    instrument reconfiguration.