Galileo Plasma Wave Receiver is described by

  Gurnett, D. A., W. S. Kurth, R. R. Shaw, A. Roux, R. Gendrin, C. F. Kennel,
F. L. Scarf, and S. D. Shawhan, The Galileo Plasma Wave Investigation,
Space Sci. Rev., 60, 341-355, 1992.

Scientific Objectives:

The basic objective of this investigation is the study of plasma
waves and radio emissions in the magnetosphere of Jupiter. The
Voyager 1 and 2 flybys of Jupiter have now clearly shown that many
complex types of plasma wave and radio-emission phenomena occur in the
Jovian magnetosphere. These include electromagnetic whistler mode
emissions called chorus and hiss, electromagnetic continuum radiation
trapped in the magnetospheric cavity, electrostatic waves associated
with harmonics of the electron cyclotron frequency, and a wide variety
of escaping radio emissions. Some of these waves, such as the whistler
mode emissions, are believed to play an important role in the dynamics
of the magnetosphere by controlling the pitch-angle scattering and loss
of energetic charged particles. In other cases plasma waves provide an
important diagnostic tool by revealing various characteristic
frequencies of the plasma, from which quantities such as the electron
density can be computed.

Since the Galileo spacecraft will be the first orbiter of Jupiter,
this spacecraft will provide a much more comprehensive study of the
Jovian magnetosphere than was possible with the previous Pioneer and
Voyager flybys of Jupiter. Specifically, the orbit of Galileo will
provide a survey of the magnetotail at distances of up to 150 RJ over a
range of local times near local midnight, a region that has never
previously been explored; repeated passes through the plasma sheet, and
the tail lobes; and numerous close flybys of the Galilean satellites.
Of particular importance will be a very close pass by the satellite Io.
The Voyager flybys showed that volcanic gases escaping from this moon
are the main source of plasma in the Jovian magnetosphere. The primary
energization of plasma in the Jovian magnetosphere is believed to occur
in a dense plasma torus that surrounds Jupiter near Io's orbit. This
energization is associated with many complex plasma wave phenomena,
including the generation of intense kilometric and decametric radio
emissions.

In addition to exploring regions never previously investigated,
Galileo, by virtue of its long lifetime in orbit around Jupiter, also
provides a unique new capability for carrying out studies of temporal
variations on time scales that cannot be investigated with a single
flyby. For example, it is known that the kilometric and decametric
radio emissions associated with Io and its plasma torus have temporal
variations on time scales of weeks and longer. With Galileo these
temporal variations can be monitored over periods of several years and
compared with other remote sensing instruments. These measurements
should be able to tell us, for example, whether the variations are
associated with changes in the volcanoes on Io. Considerable interest
also exists in searching for evidence of magnetospheric substorm
phenomena, possibly comparable to auroral substorms in the Earth's
magnetosphere. With the Galileo plasma wave instrument, it should be
possible to provide remote sensing of substorms in a manner comparable
to the remote sensing of terrestrial auroral kilometric radiation,
which is known to be closely associated with terrestrial substorms.

To carry out comprehensive studies of plasma waves and radio
emissions at Jupiter, the Galileo plasma wave instrument incorporates
several new features that provide improvements over the previous
Voyager 1 and 2 measurements. These improvements include (1) nearly
simultaneous electric and magnetic field measurements to distinguish
electrostatic waves from electromagnetic waves, (2) direction finding
measurements to determine source locations, and (3) better frequency
and time resolution to resolve fine structure in the plasma wave and
radio emission spectrum. The main instrument package and the electric
dipole antenna system were designed and constructed at the University
of Iowa, and the search coil magnetic antenna was provided by the
Centre de Recherches en Physique de l'Environnement Terrestre et
Planetaire (CRPE).

Calibration:

An extensive series of calibrations and performance checks were
performed on the plasma wave instrument both on and off the spacecraft.
Since the logarithmic compressors used in the spectrum analyzers do
not give a true logarithmic response, the transfer function of the
logarithmic compressors must be calibrated. Because of the large
number of channels, it is not practical to calibrate each frequency
channel separately. Instead, the transfer function is measured for
each logarithmic compressor, and a frequency response calibration is
performed at a fixed amplitude for all channels using that compressor.
This procedure provides accurate calibrations for each frequency step
because each band of the receiver uses a single filter and logarithmic
compressor.

A look-up table can be constructed which converts the telemetry
data number to input signal strength. When combined with the overall
frequency response across each band, these calibrations are sufficient
to determine the signal strength in all channels served by this filter
band and compressor.

In addition to the amplitude response of the compressors, a
frequency response is also performed for each frequency channel. All
frequency channels are checked to confirm that the filter bands have
the proper shape and no spurious responses. The effective noise
bandwidths are measured by stimulating the instrument with a white
noise signal of known spectral density.

For the electric field antenna, the electric field strength is
computed by assuming that the antenna has an effective length of
Leff = 3.5 meters. This length is the distance between the geometric
centers of the two dipole elements. For the search coil magnetic
antennas, the magnetic field sensitivity and frequency response was
calibrated in the IPG magnetic field observatory at Chambon La Foret,
France. These calibrations were performed using a Helmholtz coil
driven by a known AC current source. The absolute accuracy of the
sensitivity calibration is estimated to be about 3 percent. The
magnetic noise levels were measured by placing the search coils in a
mu-metal chamber, which shields the sensors from external noise
sources.

Operational Considerations:

Nominally, the instrument is operated any time low rate science
(LRS) or greater data rate capability is available. When LRS is the
maximum data rate, the instrument is operated in its power-up mode,
with the SA and SFR toggling back and forth between the electric and
magnetic antennas. When wideband data can be recorded or transmitted
to the Earth, then the wide range of instrument modes and antenna
configurations are utilized based on the science objectives for a given
time interval. The UVS instrument has a stepper motor that drives its
grating which is a major source of magnetic interference in the
frequency range from about 100 Hz to 2 kHz. Every attempt is made to
work with the UVS team to minimize the times during which the grating
is moved while PWS is observing on the magnetic antenna. In many
cases, this requires a time- sharing arrangement which allows for some
percentage of magnetic viewing time in an interference-free
environment, but which also allows UVS to observe with a moving grating
in order to achieve its science objectives.

Detectors:

The plasma wave sensors on Galileo consist of one 6.6 meter
tip-to-tip electric dipole antenna and two search coil magnetic
antennas. The electric dipole antenna is mounted at the end of the
magnetometer boom approximately 10.6 meters from the spacecraft, and
the search coil magnetic antennas are mounted on the high gain antenna
feed. The electric antenna consists of two graphite epoxy elements
with a root diameter of 2.0 cm, tapering to 0.3 cm at the tip. The
dipole elements are mounted perpendicular to the magnetometer boom to
minimize electric field distortion effects due to the spacecraft
structure. The antenna axis is also oriented perpendicular to the
spacecraft spin axis in order to permit direction finding. Each
element is hinged 1.8 meters from the tip so that the antenna can be
folded for launch. A housing at the base of the dipole elements
contains two preamplifiers. These preamplifiers provide low impedance
signals to the main electronics package. Each element is grounded to
the spacecraft structure through a 250 MegOhm resistance to limit
differential charging effects.

The search coil magnetic antenna consists of two high permeability
rods, 25.5 and 27.5 cm long, one optimized for low frequencies, 5 Hz to
3.5 kHz, and the other optimized for high frequencies, 1 kHz to 50 kHz.
The winding on the low frequency search coil consists of 50,000 turns of
0.07 mm diameter copper wire and the winding on the high frequency
search coil consists of 2000 turns of 0.14 mm diameter copper wire.
The two search coils are mounted orthogonally to minimize the electrical
coupling between the sensors. Both search coils are mounted
perpendicular to the spacecraft spin axis. The high frequency sensor is
perpendicular to the electric dipole antenna and the low frequency
sensor is parallel to the electric dipole antenna. Two preamplifiers
mounted in a housing near the search coil are used to provide low
impedance signals to the main electronics package. Frequencies below
2.4 kHz are obtained from the low frequency search coil, and
frequencies above 2.4 kHz are obtained from the high frequency search
coil.

Electronics:

All of the signal processing for the plasma wave experiment is
performed in a single main electronics package. The main electronics
package is mounted in the spacecraft body near the base of the
magnetometer boom. Signals from the electric dipole antenna and the two
search coils are processed by a wideband receiver and three spectrum
analyzers: a high frequency spectrum analyzer also called the High
Frequency Receiver (HFR), a medium frequency spectrum analyzer also
called the Sweep Frequency Receiver (SFR), and a low frequency spectrum
analyzer also called simply the Spectrum Analyzer (SA). The HFR
provides 42 frequencies from 100.8 kHz to 5.645 MHz with a fractional
frequency spacing of delta-f/f ~ 10.0% and a bandwidth of 2 kHz. One
spectral sweep is provided every 18.67 seconds with a dynamic range of
100 db. The SFR provides 112 frequencies from 40 Hz to 160 kHz with a
fractional frequency spacing of delta-f/f ~ 8.0%. This analyzer gives
one spectral sweep every 18.67 seconds with a dynamic range of 100 db.
The low frequency SA provides 4 logarithmically spaced frequency
channels from 5.62 Hz to 31.1 Hz. All four channels are sampled once
every 2.67 seconds with a dynamic range of 110 db. The data from the
HFR, SFR, and SA (and survey wideband data as described below) are
transmitted to the ground via the low rate telemetry at a bit rate of
240 bits/sec.

The wideband waveform receiver provides waveform measurements in
three frequency bands, 5 Hz to 1 kHz, 50 Hz to 10 kHz, and 50 Hz
to 80 kHz. The frequency band to be used is controlled by the
spacecraft Command and Data Subsystem (CDS). An automatic gain control
(AGC) circuit is used to control the amplitude of the output waveform.
The AGC time constant is 0.1 seconds in the two high frequency bands
and 1.0 second in the low frequency band. The waveform from the
wideband receiver is digitized by a 4-bit analog-to-digital converter
(ADC). The sample rate of the ADC is fixed at either 3,150, 25,200, or
201,600 samples per second, depending on the frequency band selected.
The waveform data can be transmitted in real time or recorded on the
spacecraft digital tape recorder.

The plasma wave instrument has several modes of operation and
methods of data transmission. These modes are also controlled by
the spacecraft CDS. The medium and low frequency spectrum analyzers and
the wideband waveform receiver can be connected to either the electric
dipole antenna or the search coil magnetic antennas. In the normal
mode of operation, the SFR and SA are cycled between the electric and
magnetic antennas so that alternate electric and magnetic spectrums are
obtained. Since the search coils do not provide signals in the
frequency range covered by the HFR, this analyzer is always connected
to the electric antenna. In the cycling mode of operation, the time
required for a complete set of electric and magnetic field spectrums is
37.33 seconds. The SFR and SA can also be locked on either the
electric or magnetic antennas to provide improved time resolution at
the expense of complementary electric and magnetic field coverage. In
all cases the HFR, SFR, and SA outputs consist of an 8-bit binary
numbers that are approximately proportional to the logarithm of the
received signal strength. In the ground data processing the data from
the HFR, SFR, and SA will be displayed in the form of color
frequency-time spectrograms. The frequency scale of the Galileo
spectrograms will extend from 5.6 Hz to 5.65 MHz, and variable time
scales will be available, ranging from 30 minutes to more than 24
hours, depending on the application. Normally, 24-hour spectrograms
will be used to survey the plasma wave data. These survey spectrograms
will be used to select specific intervals for more detailed analysis,
such as comparison with charged particle or magnetic field data, or
direction-finding analyses.

The greatest flexibility in the operation of the plasma wave
instrument is available in the wideband waveform receiver. This
receiver provides very high resolution measurements of electric and
magnetic field waveforms during times of special interest, such as the
pass through the Io torus and satellite encounters. The waveform data
provide the highest possible frequency and time resolution, subject
only to the constraints of Fourier analysis, delta-f*delta-t ~ 1.
Although the waveform receiver has only three frequency bands, with bit
rates of 12.6, 100.8, and 806.4 kbits/sec, several spacecraft modes are
available for recording and transmitting the data to the ground. In
the highest time resolution mode, an essentially continuous sample of
the electric or magnetic field waveform can be obtained over a
bandwidth of 50 Hz to 80 kHz for periods of up to 18 minutes (the time
required to fill the spacecraft tape recorder).

On the ground the waveform data will be Fourier transformed in
discrete packets, usually consisting of 1024 samples, and
displayed in the form of a frequency-time spectrogram. These
frequency-time spectrograms provide the highest time resolution data
available from the Galileo plasma wave instrument. In certain modes of
operation, such as MPW, XPW, and PW4, the duration of the wideband
recording can be extended at the expense of reduced duty cycle,
frequency coverage, or analysis bandwidth. To provide some wideband
telemetry even when the high rate telemetry link is not available, a
waveform survey output is included in the regular low rate telemetry
data. This waveform survey output provides one block of 280 waveform
samples every 18.67 seconds in two frequency bands, 5 Hz to 1 kHz and
50 Hz to 10 kHz.

Filters:

The following three tables describe the 158 frequency channels
which make up the low rate science portion of the Galileo PWS.

Table 1. Spectrum Analyzer (SA) Channels

Channel MOD(mf,4) Center Frequency (Hz) Bandwidth (Hz)
1 0 (4) 5.62 0.832
2 3 10.0 1.86
3 2 17.8 2.75
4 1 31.1 4.79

mf is the minor frame counted from 1 through 28.

Table 2. Sweep Frequency Receiver (SFR) Channels

Chan mf Freq. (Hz) Bandwidth (Hz)

Band 1
1 1 42.1 4.26
2 2 45.6
3 3 49.0
4 4 52.5
5 5 56.0
6 6 59.6
7 7 66.7
8 8 70.4
9 9 77.7
10 10 81.5
11 11 89.0
12 12 96.7
13 13 104.5
14 14 112.5
15 15 120.6
16 16 128.9
17 17 137.3
18 18 150.2
19 19 158.9
20 20 172.5
21 21 186.4
22 22 200.7
23 23 215.5
24 24 235.9
25 25 251.7
26 26 268.0
27 27 290.6
28 28 314.1

Band 2
29 1 337. 6.76
30 2 364.
31 3 392.
32 4 420.
33 5 448.
34 6 476.
35 7 534.
36 8 563.
37 9 622.
38 10 652.
39 11 712.
40 12 774.
41 13 836.
42 14 900.
43 15 965.
44 16 1.031k
45 17 1.098k
46 18 1.201k
47 19 1.272k
48 20 1.380k
49 21 1.491k
50 22 1.606k
51 23 1.724k
52 24 1.887k
53 25 2.013k
54 26 2.144k
55 27 2.325k
56 28 2.513k

Band 3
57 1 2.70k 120.
58 2 2.91k
59 3 3.14k
60 4 3.36k
61 5 3.58k
62 6 3.81k
63 7 4.27k
64 8 4.50k
65 9 4.98k
66 10 5.21k
67 11 5.70k
68 12 6.19k
69 13 6.69k
70 14 7.20k
71 15 7.72k
72 16 8.25k
73 17 8.78k
74 18 9.61k
75 19 10.17k
76 20 11.04k
77 21 11.93k
78 22 12.85k
79 23 13.79k
80 24 15.09k
81 25 16.11k
82 26 17.15k
83 27 18.59k
84 28 20.10k

Band 4
85 1 21.6k 1520.
86 2 23.3k
87 3 25.1k
88 4 26.9k
89 5 28.7k
90 6 30.5k
91 7 34.2k
92 8 36.0k
93 9 39.8k
94 10 41.7k
95 11 45.6k
96 12 49.5k
97 13 53.5k
98 14 57.6k
99 15 61.7k
100 16 66.0k
101 17 70.3k
102 18 76.9k
103 19 81.4k
104 20 88.3k
105 21 95.4k
106 22 102.8k
107 23 110.3k
108 24 120.7k
109 25 128.9k
110 26 137.2k
111 27 148.8k
112 28 160.8k
(Note that the same bandwidth applies to the entire set of channels in
band.)

Table 3. High Frequency Receiver (HFR) Channels

HFR Center Frequency
Channel mf (MHz)
1 1, 2 0.1008
2 5, 6 0.1134
3 9, 10 0.1260
4 13, 14 0.1386
5 17, 18 0.1512
6 21, 22 0.1638
7 25, 26 0.1764
8 3, 4 0.2016
9 7, 8 0.2268
10 1, 12 0.2520
11 5, 16 0.2772
12 9, 20 0.3024
13 3, 24 0.3276
14 7, 28 0.3528
15 1 0.4032
16 5 0.4536
17 9 0.5040
18 13 0.5544
19 17 0.6048
20 21 0.6552
21 25 0.7056
22 2 0.8060
23 6 0.9070
24 10 1.008
25 14 1.109
26 18 1.210
27 22 1.310
28 26 1.411
29 3 1.613
30 7 1.814
31 11 2.016
32 15 2.218
33 19 2.419
34 23 2.621
35 27 2.822
36 4 3.226
37 8 3.629
38 12 4.032
39 16 4.435
40 20 4.838
41 24 5.242
42 28 5.645
(The bandwidth for all channels is 1340 Hz)

Mounting Offsets:

The electric antenna is mounted at the end of the magnetometer
boom such that its effective axis is parallel to the spacecraft X
axis (perpendicular to both the magnetometer boom and the spacecraft
spin axis. The low frequency magnetic search coil is mounted with its
effective axis parallel to the spacecraft X axis and the high frequency
search coil is parallel to the spacecraft Y axis (perpendicular to the
X axis and to the spacecraft spin axis.

Field of View:

The field of view of the PWS, whether from the electric dipole
antenna or one of the magnetic search coils is a standard dipole
antenna pattern which has a maximum sensitivity to the field along the
axis of the sensor. For radio waves which propagate above the
characteristic frequencies of the plasma and which do not interact with
the local plasma, this means maximum sensitivity is to sources in a
plane perpendicular to the antenna axis, since the electric and
magnetic fields of a radio wave are normally perpendicular to the
propagation vector.

Data Rates:

The basic low rate science (LRS) data rate of the instrument is
240 bps. Wideband waveform receiver data rates range from 19.2
kbps to 806.4 kbps, depending on the telemetry mode. Rates of 94._
kbps and lower can be either recorded on the spacecraft tape recorder
or transmitted directly to the ground; rates of 403.2 and 806.4 kbps
can only be recorded onboard for later playback at lower rates.

Instrument Modes:

The PWS has several modes of operation. The SA and SFR can either
monitor only the electric antenna, only the magnetic antenna, or toggle
back and forth between the two, obtaining a complete spectral scan from
each antenna before switching to the other. This toggling mode is the
most commonly utilized mode. There is also a mode which enables the
magnetic search coil calibration tone. When enabled, the calibration
signal is excited for 56 mf (37.33 sec) at the beginning of each
MOD(RIM,8)=0 until disabled.

The wideband receiver has three basic modes (analysis bandwidths)
and can be attached to either the electric antenna or the magnetic.
The three modes provide analysis bandwidths of 10 kHz, 80 kHz, and
1 kHz although there is an additional mode which toggles between the
10 kHz and 1 kHz mode. In this mode, the waveform data is collected
for 14 mf (9.33 sec) in one bandwidth and then for 14 mf in the other
bandwidth.

A wide range of telemetry formats are available for the wideband
data. For any one of the three wideband modes, the instantaneous
data rate is fixed (806.4 kbps for the 80 kHz mode, 100.8 kbps for the
10 kHz mode, and 12.6 kbps for the 1 kHz mode. However, the different
telemetry modes differ primarily in the number of consecutive samples
collected during an RTI. This results in a variation in the duty cycle
depending on the data rate allocated to this data stream in the
selected telemetry mode.

Phase 2 Software Implications:

In response to the failure of the high gain antenna and the
resulting reduction in downlink telecommunications capabilities for the
Galileo spacecraft, the Galileo Project undertook a massive
re-programming of onboard software in order to enable science
observations in Jupiter orbit. Coupled with flight software changes,
modifications to the Deep Space Network were also undertaken to improve
the overall downlink capability from the spacecraft. Together, these
actions increased the actual bit-to-ground capability from a maximum of
about 10 bps to a maximum of 160 bps. In addition, onboard data
compression was implemented which increased the information content of
the downlink by roughly a factor of 10.

The PWS instrument has no microprocessor; all of its functions
are hardwired, hence, no reprogramming of the instrument itself was
possible. One result of this is that the basic timing of the
instrument is identical to that described in the section above.
However, significant software additions in the CDS and AACS were
implemented which enable the PWS to perform its basic observations at
dramatically lower data rates. These changes can be categorized simply
as editing and compression. The LRS observations are edited as
described below to reduce the observations retained for downlink from
240 bps to 65 bps. The remaining 65 bps LRS data stream is then
compressed using the same integer cosine transform (ICT) algorithm as
used for the Solid State Imaging (SSI) data. The severity of this
compression is variable and can range from 40 bps to 5 bps, with 5 bps
being the most often utilized data rate. The wideband data are not
compressed and are minimized through editing functions only. In
addition, a new wideband telemetry mode was developed which uses bit
allocations in the original LRS telemetry frame originally reserved for
Golay encoding to produce a mode with drastically reduced data rate
requirements.

Realtime Science

LRS Editing and Compression: The original PWS LRS data
stream is edited to reduce the data rate before compression from 240
bps to 65 bps. First, the 120 bits of waveform survey data are edited
out of the data stream. Second, all housekeeping and status
information is removed; if one assumes that the sequenced commands are
executed properly and the instrument timing is maintained, the status
of the instrument can be unambiguously determined from the most recent
mode command and the current spacecraft clock. A recorded mode like
the original LRS but now called LPW preserves the full, original PWS
LRS data stream and can be used to compare realtime science with the
full data steam at limited times in case questions arise about the
assumptions of correct command execution or instrument timing. Third,
only one sample per channel is preserved in a given 18.67-second
instrument cycle. Accordingly, only the first sample of the four SA
channels are retained and only the first sample of the lower frequency
HFR channels are retained. Further, since there is an overlap between
the upper frequency range of Band 4 of the MFR with the lower range
of the HFR, and since the HFR has better sensitivity over this range,
the highest frequency 6 channels of MFR band 4 are edited out. What
remains is a single sample of each of 152 of the original 158 channels
in the low rate portion of the instrument. This is the data stream
which is forwarded to the AACS for ICT compression.

The ICT compression works on 8x8 pixel blocks of an image.
Since The PWS dynamic spectrogram can be thought of as an
image, the compression algorithm can be used to compress the
spectrogram prior to transmission to the ground. One complication is
that the PWS generally alternates between a magnetic and electric
spectrum (at least for the SA and MFR channels) and the resulting
alternating spectra add entropy to the 'image' and thereby reduce the
compressibility of the data set. Therefore, the electric and magnetic
spectra are separated or 'unzipped' prior to compression. Also, it is
necessary to build up 8x8 blocks of the spectrogram. 152 channels can
be broken into 19 8-channel segments, hence, 8 electric and 8 magnetic
spectra are accumulated in order to make an 8x152 pixel strip (19 8x8
blocks) of electric field data and a similar strip for the magnetic
data. (Note that the electric HFR data is included with the magnetic
strip so that the 18.67-sec samples of the HFR channels are preserved.)
Hence, data is collected on 16 x 18.67 second time intervals (about 5
minutes) in order to build the two 8x152 pixel strips. Each of the
strips is compressed individually with all 19 8x8 blocks in each being
used to generate a downlink packet. Since transmission errors will
make decompression impossible for all data following the error, a
5-minute gap will appear for any packet with a telemetry error.
Fortunately, telemetry errors are very infrequent and the data which
reaches the ground intact is virtually immune from the spikes typical
of bursty bit errors in an uncompressed telemetry stream.

Once on the ground, the electric and magnetic strips are
decompressed and 'zipped' back together in the original time order.
This allows sequential electric (or magnetic) spectra to be maintained
together in the event the instrument is not cycled between E and B
sweeps. The net result of this compression/decompression scheme is that
the full temporal and spectral resolution of the PWS instrument is
maintained even though the real data rate to the ground is as low as 5
bps. Of course, as in any lossy compression scheme, information is
lost. By maintaining the spectral and temporal resolution, the loss in
the resulting data set is in amplitude inaccuracy. Based on both
ground experimentation and analysis of the Jupiter data, we believe the
amplitude errors at 5 bps are no more than about 6 dB and are much less
at the higher data rates (less severe compression). At 5 bps these
amplitude errors appear as 'tiling' in which all pixels in a given 8x8
block have similar values but are different from adjoining 8x8 blocks
or else the 8x8 blocks take on a checkerboard appearance. The errors
do not appear to be systematic, hence, we believe that averaging pixels
in frequency and time over regions of the spectrogram where the
spectrum appears to be 'simple' could provide a better estimate of the
true signal strength. Generally, however, the 5 dB accuracy will
enable a wide range of studies without need to know the absolute
amplitude better than a few dB.

Recorded Data

LRS data: When PWS data are recorded, the full LRS data stream
is recorded and played back. This includes all status bits,
additional samples of multiply-sampled channels (e.g. SA channels), and
the waveform survey data. The discussion of data in the original
instrument description is fully applicable in this case.

Wideband data: The wideband data suffered the most through the
process of reducing the downlink requirements. Nevertheless, a minimal
wideband capability was retained. Beginning with the Io flyby, a new
data mode was introduced which replace the Golay encoding bits formerly
used for the LRS data format with PWS wideband data in the LPW format.
This mode is called LPW and is usually modified with the term Golay
bits to distinguish these data from the low rate data. All three of
the bandwidths (wideband modes) can be used with the LPW format and for
each, a total of 832(????) contiguous 4-bit samples are acquired.
However, these 832 samples are recorded only once per 2 minor frames,
or 2.67 seconds, hence, the best spectral temporal resolution (temporal
resolution of Fourier transforms assuming 1 transform per set of
contiguous samples) is 2.67 seconds. The spectral resolution (416
spectral components is very competitive with the HPW (94._ kbps mode)
but the temporal resolution is poorer by a factor of about 50. It
is this mode, however, which is used for virtually all of the orbital
tour recording.

Some of the original wideband telemetry modes were eliminated
and those which remain (in addition to the LPW/Golay bits) are MPW,
MPP, and HPW. These provide 7.68, 19.2, and 94._ kbps, respectively.
In practice, the LPW/Golay bits provides superior spectral resolution
over the MPW mode at significantly reduced bit rate (due to the
poor temporal resolution) and the MPW mode is not utilized in the
tour. Limited use of the MPP and HPW modes is included in the tour
data set, however.

For all the wideband telemetry modes, an additional capability
for reducing the downlink requirements was implemented. This is
an editing function often referred to as '1 of n-line editing' and
consists of returning 1 of every n sets of contiguous samples recorded.
Most of the tour data were returned with n = 2 or n = 4 so that
the temporal resolution of the Fourier transformed spectra are a factor
of 2 or 4 poorer than the original recorded data. For example, in
the LPW/Golay bits mode, if n = 2, the time between returned spectra
is not 2.67 seconds, but 5.33 seconds. Reducing the number of samples
in a contiguous set of samples by returning every nth sample was
never considered because this would under sample the waveform and
lead to aliasing problems. Likewise, truncating the number of
samples in a set would reduce the spectral resolution, thereby
defeating the remaining attribute of the wideband data."