(The following paper has been modified in format from the original and
was spell-checked after transfer to ASCII format.  The format for
references has been changed to be consistent with PDS standards.)
 
Reprinted by permission of Kluwer Academic Publishers,
 
------------------------------------------------------------------------
Space Science Reviews 60: 305-315, 1992.
Copyright 1992 Kluwer Academic Publishers. Printed in Belgium.
------------------------------------------------------------------------
THE GALILEO HEAVY ELEMENT MONITOR
 
T. L. GARRARD(1), N. GEHRELS(2), and E. C. STONE(1)
 
1 California Institute of Technology, Pasadena, CA 91125, U.S.A
2 NASA Goddard Space Flight Center, Greenbelt, MD 20771, U.S.A.
 
 
Abstract. The Heavy Ion Counter on the Galileo spacecraft will monitor
energetic heavy nuclei of the elements from C to Ni, with energies from
~6 to ~200 MeV nucl^-1. The instrument will provide measurements of
trapped heavy ions in the Jovian magnetosphere, including those
high-energy heavy ions with the potential for affecting the operation of
the spacecraft electronic circuitry. We describe the instrument, which
is a modified version of the Voyager CRS instrument.
 
 
                                1. Introduction
 
 
The Heavy Ion Counter (HIC) is included on the Galileo spacecraft
primarily for the purpose of monitoring the fluxes of energetic heavy
ions in the inner Jovian magnetosphere and high-energy solar particles
in the outer magnetosphere in order to characterize the ionizing
radiation to which electronic circuitry is most sensitive. The
measurements performed will also be of scientific interest, since the
instrument's large geometry factors and extended energy range will
provide spectral information for ions from 6C to 28Ni with energies of
~6 to > ~200 MeV nucl**-1. In this article we will concentrate on Jovian
magnetospheric science. We review here previous scientific results
concerning trapped high-energy heavy ions, describe anticipated new
findings with Galileo, and provide a brief instrument description.
 
 
                               2. Previous Results
 
 
During the Voyager encounters with Jupiter, it was discovered
[KRIMIGISETAL1979A], [KRIMIGISETAL1979B], [VOGTETAL1979A],
[VOGTETAL1979B] that a major component of the trapped radiation in the
inner Jovian magnetosphere is energetic heavy ions. The dominant heavy
species inside ~ 10 R[J] in the MeV nucl**-1 energy range were found to
be oxygen and sulfur, with sodium also present. The abundances in the
inner magnetosphere indicate that the source of the ions is the surface
or atmosphere of Io.
 
The phase space density of the energetic oxygen and sulfur ions has
positive radial gradient (i.e., increasing outward) in the inner
magnetosphere [GEHRELSETAL1981], implying that the diffusive flow at
these energies is inward. On the other hand, the density of the oxygen-
and sulfur-rich plasma in the inner magnetosphere is highest at the Io
plasma torus and decreases outward [SISCOEETAL1981]. Barbosa et al.
[BARBOSAETAL1984] have proposed that charge exchange in the plasma torus
produces fast neutrals which escape to the outer magnetosphere where ~
0.2% are re-ionized by solar ultraviolet or electron impact and
recaptured Some of the newly created ions are subsequently energized by
stochastic acceleration caused by magneto- hydrodynamic waves, producing
ions with large magnetic moments that adiabatically diffuse inward,
some attaining energies in excess of 30 MeV nucl**-1 in the inner
magnetosphere.
 
As the energetic ions flow inward, they are lost from the
magnetosphere, presumably by pitch-angle scattering of the
mirroring particles into the loss cone (see, e.g., [THORNE1982]). The
strength of this loss mechanism determines the rate at which the heavy
ions precipitate into the Jovian atmosphere, exciting ultraviolet and
X-ray auroral emissions.  At the time of the Voyager encounter in March
1979, it appeared losses due to pitch-angle scattering were occurring at
nearly the maximum rate, resulting in the precipitation of ~ 10**24 ions
s**-1 above ~70 MeV nucl**-1 G**-1, with an extrapolation down to 10 MeV
nucl**-1 G**-1 suggesting the possibility of a loss rate of ~10**26 ions
s**-1 and an auroral power of ~10**13 W [GEHRELS&STONE1983].  Major
losses are occurring in the magnetic moment (energy) range appropriate
to HIC where the total inward flow rate of oxygen ions with magnetic
moments greater than several thresholds is given as a function of radial
distance (L) from Jupiter.  The number of inflowing ions first decreases
inside 15 R[J] and falls off sharply inside 10 R[J].  Only a small
fraction of the trapped particles survive to produce the trapped
radiation at 5 R[J]. As a result, changes in the rate of pitch-angle
scattering can result in significant changes in the flux of particles
reaching 5 R[J]. It is therefore important to better characterize not
only the flux at 5 R[J], but also the nature and any long-term
variability in the loss mechanism operative outside of 5 R[J].
 
 
                            3. Anticipated New Results
 
 
The HIC instrument should provide new information on the spectra of
heavy ions at higher energies than previously possible.  Fluxes at these
energies are based on extrapolations of spectra measured at lower
energies and as a result are very uncertain.
 
Measurements with the HIC instrument will significantly reduce the
uncertainty in the high-energy fluxes. The energy of a 3070 MeV nucl**-1
G oxygen ion is within the HIC energy range into 5 R[J]. The geometry
factor for E > 20 MeV nucl**-1 oxygen ions (inside 8.5 R[J] for 3070
MeV nucl**-1 G**-1 ions) is ~4 cm**2 sr compared with 0.88 cm**2 sr for
the Voyager CRS instrument. More importantly, the livetime for measuring
 oxygen and sulfur ions with E >= 50 MeV nucl**-1 will be essentially
100% because a polling priority system (see instrument description
section) will discriminate against the large fluxes of protons and
electrons that dominate the CRS analysis rate. As a result, a flux of
only 10**-5 cm**-2 s**-1 sr**-1 should produce 1 analyzed event in
10 hours, the time that Galileo is inside 7 R[J].
 
The oxygen ions which have >= 100 MeV nucl**-1 at 5 R[J] have >= 10
MeV nucl**-1 at 10 R[J]. Thus, as Galileo makes repeated orbital
passes in the vicinity of Europa, it will be possible to monitor the
fluxes of particles with M >= 3000 MeV nucl**-1 G**-1 throughout the
mission. Such long term information is especially important, since there
are a number of reasons why the fluxes might vary. For example, the flux
of high-energy ions could be measurably affected by changes in the
density of the Io plasma torus which is the source of the escaping
neutral atoms and by changes in the solar ultraviolet which re-ionizes
the neutrals in the outer magnetosphere. The HIC instrument will measure
any time dependence of the energetic heavy ion fluxes and permit
correlative studies of any associated changes in auroral emissions or
diffusion processes.
 
The fluxes at 5 R[J] depend strongly on the loss processes
occurring inside of ~15 R[J]. It is postulated that the losses are due
to pitch-angle scattering of the mirroring particles into the loss cone,
and that the rate of scattering is close to the strong pitch-angle
diffusion limit in the inner magnetosphere ([THORNE1982],
[GEHRELS&STONE1983]). At this limit there is sufficient pitch-angle
scattering to refill the loss cone as rapidly as particle precipitation
empties it, resulting in a nearly isotropic pitch-angle distribution.
Since the HIC detectors will view nearly perpendicular to the spacecraft
spin axis, essentially complete coverage of the pitch-angle distribution
will be possible. Investigation of any time dependence of the loss
process will also be possible as Galileo repeatedly passes through the
radial range between 15 and 10 R[J] where the radial profiles indicate
that significant losses occur.
 
 
                        4. Instrument Description
 
 
The Galileo Heavy Ion Counter (HIC) consists of two solid-state
detector telescopes called Low Energy Telescopes or LETs. Use of
these two telescopes over three energy intervals provides the geometry
factor and energy range necessary to determine the fluxes of the heavy,
penetrating radiation to which solid state memories are most sensitive.
Heavy collimation and high discrimination thresholds on all detectors
provide the necessary immunity to accidental coincidences from the large
proton background. Three-parameter analysis provides additional
rejection of background.
 
Two separate telescopes are included in order to cover a wide
energy range while minimizing pulse pile-up through the optimum
selection of detector and window thicknesses. The LET E is optimized for
the detection of nuclei with energies as high as 200 MeV nucl^-1,
requiring thicker detectors. Thick windows protect this system from
low-energy proton pile-up, but also exclude lower energy oxygen and
sulfur nuclei. The second telescope, LET B, has a substantially thinner
window so that it can detect lower energy nuclei (down to ~ 6 MeV
nucl^-1), especially in the outer magnetosphere. The properties of the
two telescopes are indicated in Table I and discussed below.
 
 
 
                                 TABLE I
                          Nominal LET parameters
------------------------------------------------------------------------
Detector   Radius      Thickness        Threshold
           (cm)        (micro-meter)    (MeV)
 
 
------------------------------------------------------------------------
LB1        0.95           32.1            0.3
LB2        0.95           29.6            0.4
LB3        1.13          421              3.7
LB4        1.13          440              2.0
LE1        0.95           30.4            9.3(1.4)**a
LE2        0.95           33.4            2.0(0.3)**a
LE3        1.13          463             25(5.0)**a
LE4,5      1.66        ~2000            117(23)**a
------------------------------------------------------------------------
**a High gain.
 
 
In the LETB telescope ions which penetrate LB1 and LB2 and stop in
the thicker LB3 detector are analyzed. Detector LB4 is used in
anti-coincidence. (If desired, the command system can be used to allow
analysis for ions which penetrate LB1 and stop in LB2, or for ions which
stop in LB1 . Also detector LB4 can be turned off.) Lighter nuclei
(especially hydrogen and helium) are rejected by 'slant' discrimination
with a weighted sum of the signals from the front three detectors,
 
 
                 SLB = LB1 + 0.42 LB2 + 0.2 LB3
 
 
required to be above about 9.6 MeV.
 
The thin window (25 micro-meter Kapton) serves for thermal control
and for protection from sunlight. All detectors are surface- barrier
type. An important feature is the use of keyhole detectors for LB1 and
LB2, which define the event geometry. The active area of these detectors
 excludes the nonuniform edge of the silicon wafer through the use of
keyhole- shaped masks during the deposition of the Au and Al contacts.
 
This telescope is an improved version of the Voyager CRS LETs
(Stone et al., 1977) which have demonstrated charge (Z) resolution
of 0.1 charge units at oxygen under solar flare conditions (Cook et al.,
1980) and 0.2 charge units at ~ 5 R[J], Voyager 1's deepest penetration
of the Jovian magnetosphere [GEHRELS1982]. Additional collimation and
the thicker window decrease the HIC LET's response to background protons
in the Jovian magneto- sphere. The flux of protons in detector L1, for
example, should be reduced by at least a factor of 10 from that observed
on Voyager 1.
 
In order to measure the flux of heavy nuclei at higher energies,
LE4 and LE5 are each 2000-micro-meter lithium-drifted detectors. The
front detectors, LE1, LE2, and LE3, are identical to their LET B
counterparts, providing spectral continuity and overlap. The collimator
housing is relatively thick to provide background immunity and has a
large opening angle to provide a large geometrical factor. The windows
are 76 micro-meter Kapton and 254 micro-meter Al.  As in LET B, low
charge (low Z) events are recognized and rejected by a slant
discriminator,
 
            SB = LE1 + 0.5 LE2 + 0.1 LE3 + 0.04(LE4 + LE5),
 
where SB must exceed 9.6 MeV to allow analysis.
 
For intermediate energies, a narrow-angle geometry is defined by
LE1 and LE2 for particles stopping in either LE2 or LE3. For the
highest energies, where the fluxes are exceedingly small, measurements
are made with the wide- angle geometry defined by LE2, LE3, LE4, and the
collimator. LE1 is not required but its discriminator is recorded as a
tag bit. LE5 distinguishes stopping and penetrating events. The maximum
energy observed is determined by the LE2 discriminator threshold. This
maximum is ~185 MeV nucl**-1 for oxygen as noted in Table II. To allow
detection of penetrating galactic cosmic-ray nuclei at higher energies
in the outer magnetosphere, the LET E pre-amplifier gains can be
increased by a factor of 5 to 7 by command.
 
With the exception of spacecraft interface circuitry, all of the
electronics were originally part of the Proof Test Model of the Voyager
Cosmic-Ray Science Instrument (CRS), and additional detail may be found
in Stone et al. [STONEETAL1977B] and Stilwell et al. [STILWELLETAL1979B].
With the adjustment of amplifier gains and discriminator thresholds and
the incorporation of thicker detectors, collimators, and windows, it
has been possible to develop an instrument which is optimized for the
measurement of high-energy heavy ions trapped in the Jovian
magnetosphere. Minor modifications to the logic allow the instrument
to recognize events of the various types mentioned above, which are
summarized in Table II.
 
Event data are stored in buffers which are read according to a
polling scheme which prevents domination of the telemetry by any one
type of event. The buffer polling logic cycles through the five buffers
listed in Table II, reading out one each minor frame (2/3 s) and
stepping to the next non-zero buffer on the subsequent minor frame. If a
particular type of event occurs less than 0.3 times per second (i.e., is
rare) then all of that type will be transmitted regardless of activity
in other event types. If a particular type of event occurs more often
than 0.3 times per second, it will be readout at least 0.3 times per
second and more of ten if the other event buffers are empty.
 
Telemetry of counting rates and pulse height analyzed events is
rapid compared to the nominal 3 rpm spin rate of the spacecraft; thus
pitch-angle distributions of the trapped radiation can be measured.
Both telescopes have their axes oriented near the spin plane for this
purpose. The time resolution of the HIC is in the range from 2/3 to 2 s,
implying an angular resolution in the range from 12 deg. to 36 deg.,
which is to be compared to the telescope opening angles of 25 deg. in
narrow geometry mode and 46 deg. in wide geometry mode.
 
Many of the functions of the coincidence logic and the
buffering/readout scheme can be modified by command to optimize the
instrument for changing environments or partial failures. As noted
above, commands can also be used to change gain on the LET E
pre-amplifiers.
 
 
                                  TABLE II
                               Analysis modes
------------------------------------------------------------
Name           Requirement         Geometry    Z range
                                   factor
                                   (cm^2 sr)
------------------------------------------------------------
                         ___
LETB         LB1.LB2.LB3.LB4       0.44        C to Ni
                     ___
DUBL         LE1.LE2.LE3           0.44        C to Fe
                         ___
TRPL         LE1.LE2.LE3.LE4       0.44        C to Ni
                         ___
WDSTP        LE2.LE3.LE4.LE5       4.0         C to Fe
 
 
WDPEN        LE2.LE3.LE4.LE5       4.0         C to Fe
 
 
WDPEN        LE2.LE3.LE4.LE5.HG    4.0         Li to O
(High gain)
------------------------------------------------------------
 
 
                               TABLE II cont.
-----------------------------------------------------------------
Name         Oxygen          Sulfur Signals   Signals
             energy range    energy range     telemetered
             (MeV nucl^-1)   (MeV nucl^-1)
-----------------------------------------------------------------
 
 
LETB          6 to 18         9 to 22         LB1,LB2,LB3
 
 
DUBL         16 to 17        24 to 25         LE1,LE2
 
 
TRPL         17 to 27        25 to 38         LE1,LE2,LE3
 
 
WDSTP        26 to 46        37 to 70         LE2,LE3,LE4
 
 
WDPEN        49 to ~185      >= 70            LE2,LE3,LE4+LE5
 
 
WDPEN        49 to ~500      -                LE2,LE3,LE4+LE5
(High gain)
-----------------------------------------------------------------
 
                             Acknowledgements
 
A special acknowledgement is due A. W. Schardt, whose untimely death
prevented his co-authorship of this paper. This project was made
possible by the development of the CRS instrument under the leadership
of R. E. Vogt. W. E. Althouse (Caltech) and D. E. Stilwell (GSFC) have
provided invaluable engineering and programmatic support and advice. We
also are pleased to acknowledge the contributions of A. C. Cummings at
Caltech; M. Beasley, W. D. Davis, J. H. Trainor, and H. Trexel at GSFC:
and D. R. Johnson at JPL. This work was supported by a number of NASA
contracts and grants.
 
                                References
 
Barbosa, D. D., Eviatar, A., and Siscoe, G. L.: 1984, J. Geophys. Res.
89, 3789.
 
Cook, W. R., Stone, E. C., and Vogt, R. E.: 1980, Astrophys. J. 238,
L97.
 
Gehrels, N.: 1982, 'Energetic Oxygen and Sulfur Ions in the Jovian
Magnetosphere', CIT Ph.D. Thesis.
 
Gehrels, N. and Stone, E. C.: 1983, J. Geophys. Res. 88, 5537,
 
Gehrels, N., Stone, E. C., and Trainor, J. H.: 1981, J. Geophys. Res.
86, 8906.
 
Krimigis, S. M., Armstrong, T. P., Axford, W. I., Bostrom, C. O., Fan,
C. Y., Gloeckler, G., Lanzerotti, L. J., Keath, E. P., Zwickl, R. D.,
Carbary, J. F., and Hamilton, D. C.: 1979a, Science 204, 998.
 
Krimigis, S. M., Armstrong, T. P., Axford, W. I., Bostrom, C. O., Fan,
C. Y., Gloeckler, G., Lanzerotti, L. J., Keath, E. P., Zwickl, R. D.,
Carbary, J. F., and Hamilton, D. C.: 1979b, Science 206, 977.
 
Siscoe, G. L., Eviatar, A., Thorne, R. M., Richardson, J. D., Bagenal,
F., and Sullivan, J. D.: 1981, J. Geophys. Res. 86, 8480.
 
Stilwell, D. E., Davis, W. D., Joyce, R. M., McDonald, F. B., Trainor,
J. H., Althouse, W. E., Cummings, A. C., Garrard, T. L., Stone, E. C.,
and Vogt, R. E.: 1979, IEEE Trans. Nuc. Sci. NS-26, 513.
 
Stone, E.C., Vogt, R.E., McDonald, F.B., Teegarden, B.J., Trainor, J.H.,
Jokipii, J.R., and Webber, W. R.: 1977, Space Sci. Rev. 21, 355.
 
Thorne, R. M.: 1982, J. Geophys. Res. 87, 8105.
 
Vogt, R. E., Cook, W. R., Cummings, A. C., Garrard, T. L., Gehrels, N.,
Stone, E. C., Trainor, J. H., Schardt, A. W., Conlon, T., Lal, N., and
McDonald, F. B.: 1979a, Science 204, 1003.
 
Vogt, R. E., Cummings, A. C., Garrard, T. L., Gehrels, N., Stone, E. C.,
Trainor, J. H., Schardt, A. W., Conlon, T. F., and McDonald, F. B.:
1979b, Science 206, 984.
---------------------------------------------
End of Space Science Reviews Article
 
 
Changes to the instrument in Phase 2A
---------------------------------------------
 
In Phase 2, HIC realtime data are collected by a CDS program for a
period of time that depends on readout rate:  50 RIM (at 1 bps), 25 RIM
(2 bps), or 10 RIM (5 bps), where 1 RIM = 91 minor frames and 1 mf = 2/3
sec.  At the end of the collection period, another portion of the
program compresses the data into their output format.
 
The output (before packet headers are added) consists of up to 375 bytes
of binary data in two blocks.  A dump of simulated data is attached as
Appendix A. For details about packet headers, see MOS-GLL-3-310
[ECR35559], Flight Software Requirements, Appendix, p. 5-47; and
625-610: SIS 2244-05 P2, Instrument Packet File [SIS2244-05].  The
present description is of the HIC data only.
 
The first block, of 143 bytes, contains rate data arranged in a
predetermined format.  There are 57 rate ''words'' of 2 1/2 bytes each,
and 1/2 byte of filler (0x0) at block's end.  The first byte of each
rate word is a counter of the number of times that rate was read out.
The rest of the word (1 1/2 bytes) gives the sum of the rate counts from
those readouts, in log-compressed form.  The compression scheme is the
same as that in the SRD.
 
Several types of rate are subdivided, i.e., have more than one rate word
in the output block.  Each successive rate word within a type represents
data taken in a later portion of the collection period.  The 57 rate
words appear in the following order:
 
10 words of DUBL  (''A'' rates, for ten successive time divisions)
 6 words of TRPL  (''B'' rates, for six successive time divisions)
 6 words of WDSTP (''C'' rates, for six successive time divisions)
 6 words of WDPEN (''D'' rates, for six successive time divisions)
10 words of LETB  (''E'' rates, for ten successive time divisions)
 6 words of LE1   (''G'' rates, for six successive time divisions)
 1 word  of LE5   (''F'' rates, MUXN=10)
 1 word  of LE3   (''F'' rates, MUXN=11)
 1 word  of LE4   (''F'' rates, MUXN=12)
 1 word  of LE2   (''F'' rates, MUXN=13)
 6 words of LB1   (''H'' rates, MUXN=10,
                                for six successive time divisions)
 1 word  of LB2   (''H'' rates, MUXN=11)
 1 word  of LB3   (''H'' rates, MUXN=12)
 1 word  of LB4   (''H'' rates, MUXN=13)
 
(Slant data, from ''F'' and ''H'' rates with MUXN<10 or MUXN>13, are
discarded.)
 
Thus for the longest collection period (50 RIM), we can distinguish
~5-min changes in the DUBL and LET B rates and ~9-min changes in the
five rates that have six divisions each.  At higher data rates
(collection periods of 25 RIM or 10 RIM), the time resolution is
proportionally better.
 
The second block, of up to 232 bytes, contains event data arranged in a
flexible format.  The contents of this block can vary considerably
depending on the number and kind of events that were observed.  Events
have no time divisions.
 
The events are divided into fourteen types and kept, during the
collection period, in fourteen different arrays.  Each collection array
can hold up to 20 or 32 events, depending on type.  The array numbers
determine the order in which events are output.  Event types are
distinguished in the output block not by absolute position (as are the
rates) but by header words.
 
Each string of event information begins with a one-byte header.  The
first half- byte contains the event array number, i.e., the type.  The
second half-byte is a counter that gives the number of events in the
string for that type, less one; i.e., counter = counts - 1.  If no
events were observed for a given type, no string is output.
 
Each event type has a characteristic word length, so the header also
gives the length of the event string.  Each type also has a
characteristic meaning assigned to each bit in its word.  For details,
see Appendix B.
 
After all the output events comes an event counter array, which consists
of a leading 'f' followed by six numbers of 1 1/2 bytes each.  These
show the total number of events counted in the collection period for
each of six kinds of event:  DUBL, TRPL, WDSTP, WDPEN, LETB, and null
(tag word = 0).
 
The fourteen event types are distributed among the five non-null kinds
as follows:  DUBL, type 9;  TRPL, types 5 and 12;  WDSTP, types 1, 6,
13, and 14; WDPEN, types 7 and 8;  and LETB, types 2, 3, 4, 10, and 11.
Counts in this array include ''caution'' events, those whose caution bit
in the tag word is set.
 
When rates are very low, all observed events are output and the event
block is very short.  When rates are very high, details of some events
will be lost because of the limited size of the output block:  232 bytes
will hold only about 60 events plus the counters.
 
If the event arrays are quite full, the program outputs sixteen events
from each type, starting with number one, until it has no more room.
For this situation, only events of types 1-5 will be output in detail;
the rest will be lost, as in the Appendix A dump.  Since the program
outputs the first sixteen events of each type, events from the beginning
of the collection period are favored ever more heavily as rates climb.
 
------------------------------------------------------------------------
Appendix A:  hex dump of output block for sparse events (event block
short)
 
------------------------------------------------------------------------
 
Bar = first rate counter of each series, or event type's header.
 x = filler nybble (always 0x0)
 
octal      data*
address
         __
0000000  8880 8988 1897 8179 8818 9881 8978 1798 8189 8818
                     __
0000024  9781 7988 18ed 7edf d7fd fd7f dfd7 fdfd 7fdf c7fc
         __                                   __
0000050  ed73 1fd7 3dfd 73df d73d fd73 dfc7 3ded 76df d77d
                                  __
0000074  fd77 dfd7 7dfd 77df c77c 8872 a977 3c98 73e9 873e
                                              __
0000120  9773 c987 3e98 73e9 773c 9873 e987 3eed 6b1f d6bd
                                  __    __    __    ___
0000144  fd6b dfd6 bdfd 6bdf c6bd 5e76 b5e6 815d 68b5 d697
         __                                   __    ___
0000170  0f85 20f8 5210 8601 0860 1086 0108 605e 6b05 d6ba
         __    x__                               __
0000214  5d6c 5012 ab9b 9c65 ab9b 9c65 ab9b 9c65 525c 66d5
                               __
0000240  4e5c 66d5 4e5c 66d5 4e62 9cfc f652 9cfc f652 9cfc
              __                   x__                   x
0000264  f652 72fc a9af ca9a fca9 a082 bcab 5bca b5bc ab50
         __                                             __
0000310  924c 24c2 f0fb 794c 24c2 f0fb 794c 24c2 f0fb 79c2
                                       __
0000334  3185 a89d 3185 a89d 3185 a89d d217 2670 ca17 2670
                     __                               _
0000360  ca17 2670 cae2 19c3 d8a5 19c3 d8a5 19c3 d8a5 f003
                           x
0000404  0060 0c00 6001 17b0
 
 
* From Octal Address 0000 to 0216 rate data are provided. Higher
  addresses contain event data.
 
-------------------------------------
Appendix B:  event construction
-------------------------------------
 
The table below gives details of how each event type is compressed into
its word.  The tag word is discarded for all types but #9 since the
array number gives much of the relevant information.  The ''maximum
number of events'' is the highest number that the program will put into
that collection array.
 
A ''small'' event has all zeroes in the first half-byte of each of its
PHA words. Other events are ''big.''  A ''caution'' event is one whose
caution bit (last bit of tag word) is set.
 
Consult the SRD, Table 6 (p.8), for detector correspondence to PHA words
for the various event types.
 
 
Array  Word   Max #   Event Type              Bit source
No.    Len.   Evts.                      PHA3     PHA2     PHA1
----------------------------------------------------------------
 1      32     32   big LE1 WDSTP       top 11   top 10   top 11
 2       8     32   LET B single        --       --       top 8
 3      20     20   big LETB double     --       top 10   top 10
 4      32     20   big LETB triple     top 10   top 11   top 11
 5      32     20   big TRPL            top 10   top 11   top 11
 6      32     20   big !LE1 WDSTP      top 11   top 10   top 11
 7      20     20   LE1 WDPEN           top 10   top 10   --
 8      20     20   !LE1 WDPEN          top 10   top 10   --
 9      48     32   DUBL and ''caution'' all     all      all
                                                          plus tag word
10 (a)  20     20   small LETB double   --       bot 10   bot 10
11 (b)  32     20   small LETB triple   bot 10   bot 11   bot 11
12 (c)  32     20   small TRPL          bot 10   bot 11   bot 11
13 (d)  32     20   small LE1 WDSTP     bot 11   bot 10   bot 11
14 (e)  32     20   small !LE1 WDSTP    bot 11   bot 10   bot 11
 
Note that the following pairs of event types have the same construction
(there are only 9 different schemes):  1 & 6;  4 & 5;  7 & 8;  11 & 12;
13 & 14.