Data Set Overview
  =================

    The Cosmic Ray Telescope for the Effects of Radiation (CRaTER) is a
    stacked detector-absorber cosmic-ray telescope designed to answer key
    questions to enable future human exploration of the Solar System.
    CRaTER's primary measurement goal is to measure directly the average
    lineal energy transfer (LET or 'y') spectra caused by space radiation
    penetrating and interacting with shielding material.  Such measured LET
    spectra are frequently unavailable.  In the absence of measurements,
    numerical models are used to provide estimates of LET; the reliability of
    the models require experimental measurements to provide a ground truth.

    The Level 2 dataset consists of files containing data processed from the
    Level 1 primary science, secondary science, and housekeeping calibrated
    data records (CDR).  During processing, derived data records (DDR) are
    formed by combining Level 1 CDR with derived parameters such as average
    LET, detector event flags, and instrument viewing geometry data.  The DDR
    are written to files in plain text, fixed record format; each file con-
    tains DDR for a single UTC day. All times in Level 2 data products are
    reported in both spacecraft clock units and UTC.

    The Level 2 data products are intended as the primary CRaTER data source
    for further data analyses or scientific research.

    See the MISSION.CAT file for more information on the LRO mission.
    See the CRAT_INST.CAT file for more information on the CRaTER instrument.
    See SPENCEETAL2010 for detailed description of LRO flight version of the
    instrument, its operations, and data processing.

    Science Objectives and Observation Strategy
    -------------------------------------------
    CRaTER is designed to achieve characterization of the global lunar
    radiation environment and its biological impacts and potential mitigation
    as well as investigation of shielding capabilities and validation of
    other deep space radiation mitigation strategies involving materials.
    CRaTER will fill knowledge gaps regarding radiation effects, provide
    fundamental progress in knowledge of the Moon's radiation environment,
    and provide specific path-finding benefits for future planned human
    exploration.

    Parameters
    ----------
    LRO CRaTER flight instrument identification:
    --instrument model = Flight Model 1 (FM1);
    --instrument serial number (S/N) = 02;
    --FPGA revision code = 3.

    Data
    ----
    CRaTER's principal measurement is the energy deposited in the 3-pairs of
    silicon detectors by charged particles and photons passing through the
    instrument's 'telescope' unit.  Whenever the coulombic charge signal re-
    sulting from the energy deposited in a detector exceeds a predefined and
    fixed threshold, the instrument's electronics performs a detailed
    measurement of the signals from all of the detectors.  The resulting
    detector signal amplitudes are compared to the values of the 'lower level
    discriminators' (LLDs).  LLDs establish minimum amplitudes for signals to
    qualify as valid charged-particle or photon interactions.  The LLD values
    are generally set to insure that the desired charged-particle or photon
    measurements are not contaminated by system electronic noise.  Seperate
    LLD settings are required for the thick and thin detectors due to the
    difference in their sensitivities; the thin and thick detector LLD values
    are reported in the 'DiscThin' and 'DiscThick' parameters as part of the
    secondary science packet.

    In addition to the LLD settings, measurement filtering is achieved
    through detector coincidence requirements--the combination of detectors
    registering valid signals to qualify as a charged-particle or photon
    measurement 'event'.  To measure all charged particles arriving from the
    instrument's zenith or nadir directions, for example, the coincidence
    requirements would be valid signals in at least detectors 1, or 2, or 5,
    or 6.  Conversely, a coincidence consisting of valid signals in all six
    detectors would ensure only zenith- or nadir-arriving charged particles
    with high energies are reported.  For CRaTER's six axially-coaligned
    detectors there are 64 possible coincidence combinations.  The desired
    set of coincidence combinations are stored as a coincidence mask
    parameter in the instrument's memory; the coincidence mask setting is
    reported in the 'Mask' parameter as part of the secondary science packet.

    To qualify as an 'event', therefore, a charged particle or photon passing
    through CRaTER's telescope must interact and deposit sufficient energy to
    generate signals with amplitudes in excess of the specified LLDs in a
    specified combination of detectors; only data for valid 'events' are re-
    ported in the instrument's telemetry.

    The measured interaction event data is written as a series of primary
    science packets to the instrument's output telemetry buffer for the
    spacecraft to read. At ~1 second intervals CRaTER receives a timing pulse
    from the spacecraft, at which time it flushes the primary science data
    from the output buffer and writes a secondary science packet for the
    spacecraft to read.  Every 16 seconds a housekeeping packet is also
    created and written to the output buffer.

    The Level 2 data are by combining the Level 1 data with derived or
    supplemental parameters including average LET in each detector, detector
    event flags, instrument electrical power consumption, and instrument
    viewing geometry information.

    The Level 2 dataset is composed of the three types of time-sequential
    derived data records (DDR): (1) primary science, (2) secondary
    science, and 3) housekeeping.  The three types of DDR are written to
    seperate data files in plain text, fixed record format.  Each file
    contains DDR for a single UTC day.

    The Level 2 primary science data consists of a sequence of interaction
    event DDR--one DDR for each measured event.  Each DDR consists of the
    energy deposited in each of the six detectors, the resulting average LET,
    and the spacecraft time and UTC at the end of the measurement interval
    (receipt of spacecraft timing pulse).  Also included in the primary
    science DDR are two sets of flags related to the measured deposited
    energy in each detector:  one set flags deposited energies exceeding
    corresponding LLD values; the second set flags deposited energies
    approaching the saturation value for the associated amplifier-ADC strings
    (signals exceeding 95% of the ADC's dynamic range).  DDR for events
    recorded in the same measurement interval have the same time tags--the
    'SECONDS','FRACT', and 'TIME' field values.  Although numerous events may
     have the same time value, the events are recorded in the order in which
     they occurred; this relative order is captured in the DDR 'INDEX' field.

    The Level 2 secondary science DDR contain the majority of instrument con-
    figuration settings, status flags, and event counters.  Reported con-
    figuration settings include the last command sent to CRaTER, detector
    LLD settings, and coincidence mask values. The record's time tag includes
    both spacecraft time and UTC.  Status flags available in the secondary
    science DDR include detector bias status, selected pulse amplitude range
    and rate for the internal calibration pulser, and detector processing
    status. Counters report the number of 'singles' for each detector as well
    as the number of 'good', 'rejected', and total events recorded by CRaTER
    during the monitoring period.  Also included in the secondary science DDR
    is LRO's location relative to the center of the Moon; the location is
    provided as three orthogonal vectors (Px, Py, Pz) in the MOON_ME (Moon
    Mean Earth/Rotation Axis) reference frame

    The Level 2 housekeeping DDR contain measured instrument operating and
    environmental parameters used to assess the health and performance of the
    instrument, such as power supply output voltages, detector bias voltages
    and currents, pulse amplitudes from the internal calibration pulser, and
    temperatures at five locations inside of the instrument's housing.  The
    analog output signal (voltage) from radiation monitor is also included
    the housekeeping CDR.  The record's time tag includes both spacecraft
    time and UTC.  Also included in the housekeeping DDR are two status flags
    related to the relative orientation of the instrument's boresite axis:
    one flag indicates when the boresite axis does not intercept the lunar
    surface 'OFFMOONFLAG'; the second indicates when LRO and CRaTER are in
    eclipse.

    LRO's location and CRaTER's relative boresite orientation are derived
    from definitive spacecraft ephemeris ('SPK') and orientation ('CK')
    kernels, using transformation routines from the JPL NAIF (Navigation and
    Ancillary Information Facility) toolkit. The daily kernel files used
    during the Level 2 DDR processing are stored in  the EXTRAS directory.

Confidence Level Overview
    -------------------------
    An assessment of the accuracy and precision of data in the
    LRO MOON CRATER 3/4 CALIBRATED LET DATA V1.0 dataset is limited to the
    measureddeposited energy in each detector and the resulting average
    lineal energy transfer (LET or 'y') values (primary science DDR).  Gene-
    ral instrument configuration and housekeeping parameters (e.g., tempera-
    tures, voltages, currents, LLD voltages, pulser signal amplitudes, space-
    craft clock value) are provided with no statement of uncertainty--the ac-
    curacy of these parameters is assumed to be sufficient for general cor-
    relation and trending analysis.  The accuracy of the housekeeping temper-
    ature parameters has an impact on the accuracy and precision of the con-
    version from detector PHA channel numbers to deposited energy (and LET)
    values; this impact, however, is very small in comparison to other
    sources of systematic and stochastic error.

    Potential sources of instrument systematic error include signal pulse
    shaping output linearity, analog-to-digital conversion (ADC) linearity,
    electronic calibration source stability and linearity, and the accuracy
    of the gain and offset values determined for each detector-amplifier-ADC
    string.

    The linearity of the amplifier-ADC strings (i.e., pulse height
    analyzer or PHA) was established with a precision external pulser.  For a
    given pulser output setting, the variability in output pulse amplitude is
    determined to be 0.01%.  Over the pulser's full range of output pulse
    amplitude settings, the measured pulse amplitudes were found to be very
    linear, with an RMS fit residual upper limit of 0.1%.

    The external pulser was used to establish the linearity of the six CRaTER
    PHA circuits.  The precision external pulser served as a calibrated input
    charge source by coupling it (via a precision capacitor) to the base of
    each PHA circuit's preamplifier.  Each PHA circuit's response was found
    to be very linear, with RMS fit residuals significantly less than 0.1%.

    Temporal stability of the PHA circuits was established through repeated
    testing with the external pulser over an 15-month period.  Between Sep
    2007 and Jan 2009, each PHA circuit was tested five times at a fixed
    pulser output setting.  The output of each PHA circuit was determined to
    be very stable, with ~0.06% variability in the value of the center of the
    PHA peak.

    Temperature dependence of the gain of each PHA circuit was measured over
    the expected range of operating temperatures during the LRO mission.  The
    output of each PHA circuit to fixed amplitude pulses from the precision
    external pulser was measured with the CRaTER instrument operating at -30
    degrees C, -10 degrees C, +10 degrees C, and +35 degrees C (temperature
    measured inside the instrument's case close to the analog and digital
    circuit boards).  The PHA circuit gains were found to be fairly stable
    over this temperature range, with only a weak non-linear temperature de-
    pendence.  Detectors 2, 4, and 6 PHA circuits exhibited gain variations
    of ~ +/- 0.1% over the temperature range; detectors 1, 3, and 5 PHA
    circuits gains varied by ~ +/- 0.5%.

    Potential sources of stochastic error include electronic noise, uncer-
    tainty in the PHA-channel-to-deposited-energy conversion factors (i.e.,
    'calibration values'), uncertainty in actual deposited energy values
    due to digitization, and uncertainty in the derived LET values caused by
    variability in particles' paths through the detectors.

    From the standard deviations of the pulse amplitudes measured over the
    full dynamic range of each amplifier-A-to-D-converter strings, the upper
    limit on system electronic noise is approximately 0.15% of pulse
    amplitude or 0.02% of each string's maximum output value. [The system
    electronic noise measured with CRaTER operating at 10 degrees C].

    PHA channel number is converted to deposited energy by

      Ei [keV] = GiCi + Oi, where

      Ei [keV] =              deposited energy measured by detector/
                              PHA chain i,
      Ci [ADU or channel #] = output from detector/PHA chain i,
      Gi [keV/ADU] =          gain of detector/PHA chain i, and
      Oi [keV] =              offset of detector/PHA chain i.

    The calibration values Gi and Ci used to convert PHA output to deposited
    energy were determined through a combination of alpha particle exposure
    measurements and modeling of the instrument's response to moderate energy
    protons.  A more extensive description of the calibration process is
    found in SPENCEETAL2010.

    The LRO CRaTER instrument V1.0 calibration values are listed in
    SPENCEETAL2010, table 6, and reproduced here.

    Parameter      Units      D1     D2     D3     D4     D5     D6
    -----------------------------------------------------------------
    Gain, Gi       keV/ADU   76.3   21.8   78.6   21.6   76.3   21.9
    Offset, Oi     keV       105.1  50.0   152.8  74.7   119.1  46.6

    The uncertainty in the Gi and Ci values awaits further analysis.  A re-
    vision to this catalog file will be provided when the values become
    available.

    The process of converting the detector signals into digital values re-
    quires discretizing the amplifier analog output signals into one of a
    possible 4096 linearly-spaced values.  These 4096 'channel' or 'ADU'
    values correspond to ranges of ~0-300 MeV and ~0-90 MeV for the thin and
    thick detector PHA circuits, respectively.  Each PHA channel corresponds
    to a small but finite range of energies described by a probability
    distribution rather than a discrete energy value.  The calibration
    process establishes an effective energy and energy width for each
    channel. Assuming the actual deposited energy probability distribution
    for a given PHA channel is approximately flat, the average energy and
    uncertainty corresponding to the channel are the effective energy and
    energy width established through calibrations.  While the absolute
    magnitude of the uncertainty resulting from discretization is a constant
    value (one-half the gain), the relative uncertainty is a function of the
    energy corresponding to the particular PHA channel--the lower the
    channel's corresponding energy, the higher the realtive uncertainty.
    The discretization uncertainty extremes are summarized in the
    following table.

    Detector/         Energy (keV)              Energy (keV)
    PHA Chain          PHA = 0 ADU               PHA = 4095
    ----------------------------------------------------------
    D1           105.1 +/- 38.2 (36.3%)    312554 +/- 38.2 (0.012%)
    D2           50.0  +/- 10.9 (21.8%)    89321  +/- 10.9 (0.012%)
    D3           152.8 +/- 39.3 (25.7%)    322020 +/- 39.3 (0.012%)
    D4           74.7  +/- 10.8 (14.5%)    88527  +/- 10.8 (0.012%)
    D5           119.1 +/- 38.2 (32.0%)    312568 +/- 38.2 (0.012%)
    D6           46.6  +/- 11.0 (23.5%)    89727  +/- 11.0 (0.012%)

    For PHA values > 48 ADU, the relative uncertainty in the deposited energy
    due to discretization is < 1% a for all detector/PHA chains.

    The uncertainty in the deposited energy values contributes to the
    final uncertainty in the LET values.

    The LET uncertainty due to variability in particle paths through the de-
    tectors arises from the limited collimation of the particles incident on
    each detector.  Particle-detector incidence angle is not necessarily per-
    pendicular, but some value between perpendicular and the an angle deter-
    mined by the detectors defining the event's coincidence.  The result is
    that the pathlength through a given detector can be significantly longer
    than the detector's thickness, which in turn lowers the LET value for a
    given deposited energy value.  The particle pathlengths throught the de-
    tectors--instead of a being known and fixed values--must instead be de-
    scribed by probability distributions.  The resulting LET value is there-
    fore also described by a probability distribution.  In practice, LET is
    computed using the most likely pathlength value (as determined analyti-
    cally or through Monte Carlo modeling), while the probabilitic nature of
    the pathlength is included in the value's overall uncertainty.  The path-
    length uncertainty is greatest for particles which deposit energy only in
    the outer-most pairs of detectors, and decreases as particles deposit
    energy in the remaining pairs of detectors.

    This overview has identified, described, and where possible enumerated
    the various error/uncertainty components.  The confidence levels for the
    total cumulative uncertainty in the measured deposited energies and
    derived LET values awaits further analysis.  When the values become
    available a revision will be provided to this catalog file.

    Review
    ------
    A minimal set of automated quality control steps are used by the data
    processing system to verify the integrity of the data during the initial
    creation of the L0 data files.  Each raw data packet's CCSDS header is
    checked for format and content. Packets are discarded if their headers
    are corrupted, incorrectly formatted, or containing invalid values.  All
    packets are sorted into time order and checked for temporal gaps.  Dupli-
    cate packets are also discarded.  Metrics plus any detected anomalies are
    written to process log files for review by scientists and engineers from
    the instrument team.  Anomalies noted during the processing are investi-
    gated.  Anomalies due to missing input files (e.g., instrument science
    and housekeeping data files, spacecraft housekeeping data files,
    spacecraft ephemeris kernels, and ancillary files such as leap second and
    spacecraft clock kernels) are corrected by locating the missing input and
    reprocessing the data.

    All data is periodically analyzed using graphical and statistical methods
    to check for out-of-range values as well as anomalous trends that may
    indicate detector and/or amplifier-ADC string degradation.

    Data Coverage and Quality
    -------------------------
    The start date for the initial version of the LRO-L-CRAT-3/4-DDR-
    PROCESSED-V1.0 archival volume is 2009-06-29T00:00:00.000. This date/time
    is the beginning of the first full day following completion of LRO lunar
    orbit insertion (LOI) and transition to the nominal nadir-pointing
    observation attitude.  It is also the first day for which complete
    re-constructed ephemeris ('SPK') data was provided by the LRO Mission
    Operations Center.  There is only limited re-constructed ephemeris data
    currently available for the period between initial instrument power-up
    (2009-06-20) and LOI completion and transition to the nominal observing
    attitude.  CRaTER data obtained during Cruise Phase (instrument power-up
    - 2009-06-23), Lunar Orbit Acquisition (2009-06-23), and initial
    Commissioning Period (2009-06-23 - 2009-06-28) will be included when more
    complete ephemeris data from this early part of the mission becomes
    available.

    Data gaps are identified during initial data processing. The gap start
    and stop times are recorded in gap files stored in the DOCUMENT directory
    there are seperate gap files for the primary science, secondary science,
    and house-keeping data sets.  Each gap file contains a cumulative listing
    of the missing data up to and including the days for the data current
    volume.  Description of overall data coverage and quality. This section
    should include information about gaps in the data (both for times or re-
    gions) and details regarding how missing or poor data are flagged or
    filled, if applicable.  The minimum duration between successive data
    packets to qualify as a data gap is specified during data processing. The
    default durations are 2 seconds for both primary and secondary science
    data packets, and 20 seconds for housekeeping data packets.  These values
    may be over ridden at the time of data processing, however.  The actual
    durations used while processing a specific set of data are recorded in
    the corresponding process log file; the log files are found in the DATA
    directory with their corresponding data products.

    Aperiodic episodes of sporadic, significant elevation in the thick detec-
    tor (D2, D4, and D6) singles rates have been observed during all phases
    of mission phases. The elevated singles rates most commonly occur in
    detector D2, but have also been observed in detector D6; a detector's
    singles rate may increase by a factor of 20 or more.  During these
    periods increases may occur in both the 'reject' and 'good' event rates.
    Episodes tend to last for three to five weeks, followed by extended
    periods with nominal singles rates.  During an episode singles rates
    vary sporadically between nominal and extremely elevated levels, although
    there seems to be a general gradual build-up and decline in the peak
    magnitude of the singles rates over the course of an episode.  Despite
    intensive analysis, the cause for the periods of elevated singles rates
    has not yet been determined.  No correlation has been found with
    spacecraft location, local space and spacecraft environment conditions,
    instrument boresite direction, or spacecraft and instrument operations.
    Users are urged to first plot the detector singles rates and 'good' and
    'reject' event rates as a function of time to identify periods with
    elevated singles rates which may impact their particular use of the data.

    Limitations
    -----------
    The LRO MOON CRATER 3/4 CALIBRATED LET DATA V1.0 data set includes all
    data obtained by the CRaTER instrument, including data from periods when
    the instrument was placed into special configurations.  Special configur-
    ations include the instrument start-up tests that occur whenever the in-
    strument is power cycled to (e.g., initial instrument start-up, recovery
    following spacecraft transition to sun-safe mode) as well routine cali-
    brations (90-degree off-nadir GCR background measuerments, internal pul-
    ser sweeps, LLD zero crossing measurements, and LLD sweeps).  These per-
    iods can be detected by monitoring the values in the 'CalLow', 'CalHigh',
    'DiscThin', and 'DiscThick' fields in the secondary science DDR.

    Timing resolution for the set of events recorded between two successive
    timing pulses (buffer readouts)is limited to the corresponding spacecraft
    times.  If, for example, 560 particle 'events' are measured between two
    successive timing pulses, the exact time of each event's occurrence is
    unknown--all that is known is that event was measured between the times
    of the two timing pulses. The sequence in which the events were measured,
    however, is preserved-for a given time interval, the first reported event
    was measured before the second reported event, etc.

    The maximum rate at which detector measurements can be reported in the
    primary science data is ~1200 events per second; the true number of
    events in each time interval is reported in the secondary science DDR.

    Users should be aware of the impact of the LLD settings on the primary
    and secondary science data.  The LLD settings establish the minimum
    amplitudes of the amplifier output pulse heights (i.e., minimum deposited
    energies) to qualify as a valid signal and trigger the ADC process.  In
    addition to determining the lower limit of the PHA and LET spectra, the
    choice of LLD values directly affects the number of 'good' and 'reject'
    events reported in the secondary science data DDR.  For a given set of
    incident charged-particle energy spectra, as the LLD values increase,
    the 'good' and 'reject' event rates will decrease.  Users analyzing the
    temporal variability of 'good' and 'reject' event rates should ensure the
    LLD settings do not change over the analysis period.  The nominal
    instrument operating mode maintains constant LDD settings.  Modes using
    varying LLD settings, however, occur during instrument power-up tests
    and routine calibration procedures.  In addition, as the mission
    progresses changes in noise levels due to instrument component aging may
    require adjustments to the baseline LLD settings.