Galileo Magnetometer, Ida High-Resolution Data - Additional Documentation ========================================================================= [These data were extracted from the separate catalog files originally provided with the HRSC and HRGEOPH data files.] Dataset Overview ================ This dataset contains data acquired by the Galileo Magnetometer during the Ida flyby on Aug 28, 1993. Limited space on the tape recorder forced the magnetometer team to limit their highest time resolution observations to a ~30 minute interval beginning after closest approach to Ida. In order to acquire the high time resolution mag data at Ida necessary to look for magnetic signatures similar to those observed at Gaspra, a new method of using the instrument optimal averager section was tested. This method used the spacecraft Command and Data System (CDS) computer to sample the MAG data buffer memory to retrieve vectors every 4/3 second. CDS stored the data in internal data buffers (DBUM1A and DBUM1B). The data were downlinked back to earth in near real-time via CDS memory readouts (MROs). The data were highly over-filtered by the instrument but many of the effects of over-filtering could be removed in ground processing by knowing the filter response function. Details of this process are described below in the data processing section. High resolution magnetometer data are provided in four coordinate systems. Of the two data files included, one contains data in Inertial Rotor Coordinates (IRC = despun spacecraft). This file includes the spacecraft attitude and spin phase angles from the AACS system and the sensor zero levels. The other data file contains magnetic field data in Ida-centric Solar Ecliptic (IdaSE), Earth Mean Equatorial equinox 1950 (EME-50), and Heliographic (RTN) coordinates. Trajectory data are provided as a separate dataset for each of the geophysical coordinate systems. Primary Dataset Reference: Kivelson, M.G., Z. Wang, S. Joy, K.K. Khurana, C. Polanskey, D.J. Southwood, and R.J. Walker, 'Solar Wind Interactions with Small Bodies: 2. What Can Galileo's Detection of Magnetic Rotations Tell Us About GASPRA and IDA', Advances in Space Research, 1994 Primary Instrument Reference: Kivelson, M. G., K. K. Khurana, J. D. Means, C. T. Russell, and R. C. Snare, 'The Galileo magnetic field investigation', Space Science Reviews, 60, 1-4, 357, 1992. Data Columns (Spacecraft coordinates IRC): Fortran format (1X, A24, 11F11.3) time S/C event time (UT) given in PDS time format YYYY-MM-DDThh:mm:ss.sssZ rotattt Rotor twist angle (EME-50) rotattd Rotor attitude declination (EME-50) rotattr Rotor attitude right ascension (EME-50) spinangl Rotor spin angle - inertial S/C coordinates spindelt Rotor spin motion delta screlcon Rotor-Platform relative cone angle screlclk Rotor-Platform relative clock angle Bx_sc Magnetic field X component in S/C (IRC) coordinates By_sc Magnetic field Y component in S/C (IRC) coordinates Bz_sc Magnetic field Z component in S/C (IRC) coordinates Bmag |B| Magnitude of B Data Columns (Geophysical Coordinates): Fortran format (1X, A24, 10F11.3) time S/C event time (UT) given in PDS time format YYYY-MM-DDThh:mm:ss.sssZ Bx_IdaSE X component of B in IdaSE coordinates (towards Sun) By_IdaSE Y component of B in IdaSE coordinates (towards dusk) Bz_IdaSE Z component of B in IdaSE coordinates (|| to ecliptic normal) Bx_eme50 Magnetic field X component in EME-50 coordinates By_eme50 Magnetic field Y component in EME-50 coordinates Bz_eme50 Magnetic field Z component in EME-50 coordinates Br Magnetic field radial component in RTN coordinates Bt Magnetic field tangential component in RTN coordinates Bn Magnetic field normal component in RTN coordinates Bmag |B| Magnitude of B Missing data value = 99999.999 Data Acquisition: The data for this dataset were acquired by the outboard magnetometer (flip left position) in the low field (+/- 32 nT) mode (ULLR). This mode has a digitization step size of 0.0625 nanoTesla. However, these data are acquired at 30 vectors/second and then recursively filtered in the instrument. The high rate data that are recorded to tape have a sample rate of 4.5 vectors/second. If there is sufficient variation in the 6-7 input samples that make up a single output sample, then the effective digitization step size becomes much smaller. The data are next passed to an onboard processing algorithm which corrects the data for offsets, gains, and geometry. The data can now be sent to the tape recorder or passed to the optimal averager section of the instrument. Optimal averager data are decimated to 1 vector every minor frame such that only the first sample of the minor frame is retained. These data are then despun into Inertial Rotor Coordinates (IRC) and passed to another recursive filter operation. The filter used in the Ida high resolution (4/3 sec) data was: Bo(i) = (15/16) * Bo(i-1) + (1/16)*Bi(i) where Bo(i) is the output of the i-th sample and Bi(i) is the new input vector at the time of the i-th sample. Data were sampled by the CDS every 4/3 second. Data are not averaged any further. The time tag associated with each data record is the sample time (UT) on the spacecraft (SCET). Coordinate System: The Galileo magnetometer data are being archived in 4 coordinate systems. The first coordinate system is referred to as inertial rotor coordinates (IRC). This coordinate system has the Z axis along the rotor spin axis, positive away from the antenna, and the X and Y axes in the rotor spin plane. In a crude sense, when the spacecraft is far from Earth, positive X points south, normal to the ecliptic plane, positive Y lies in the ecliptic plane in the sense of Jupiter's orbital motion and positive Z is in the anti-Earth direction. The spacecraft antenna (negative Z direction) is kept earthward-pointing to about +/- 10 deg accuracy. Ida-centric Solar Ecliptic (IdaSE) is an Ida-centered coordinate system defined by the primary vector along the instantaneous Ida->Sun (GSun) line with the Earth's ecliptic north pole (ENP) as the secondary vector. In this coordinate system: X is the GSun unit vector taken to be positive towards the Sun. Y is the vector formed by the unitized cross product ENP x GSun Z completes the right-handed set (Z = X x Y) such that the X-Z plane contains the ecliptic north pole. The Earth Mean Equatorial equinox of 1950 (EME-50) coordinate system is an inertial reference system. The primary vector in this system points from the Earth towards Aries at the reference epoch. The secondary vector is along the Earth's rotational axis positive towards its north pole. In this coordinate system: X is a unit vector in the direction of Aries at the equinox of 1950 Y is a unit vector in the direction of Z x X Z is a unit vector in the direction of the Earth's equatorial north pole at the equinox of 1950. The EME-50 coordinate system is directly supported by SPICE as the 'FK4' inertial reference frame. The heliographic (RTN) coordinate system is centered at the Sun. The primary vector in this system points from the center of the Sun to the spacecraft (R). The secondary vector in this system is the Sun's north rotational axis (Omega). R is along R, positive away from the Sun. T is along the cross product Omega x R N completes the right-handed set (R x T) Data Processing: The tape recorded portion of this dataset was processed according to the standard methods employed for LRS data. These methods have been well described in the Venus, Earth, and Gaspra documentation. A 1-D spacecraft interference function was removed from the LRS data during their processing. The 1-D and 2-D methods and coefficients are described in the Earth 1 and 2 data processing descriptions. Recorded data cover the time interval 16:53:20.026 through 17:21:36.691. Data sampling in this interval is the normal 3 vectors per minor frame with sample times offset from the start of the minor frame by (0.0, 0.2333, 0.2333) seconds. The remainder of the dataset was acquired by using the optimal averager and CDS in a special procedure used only at Ida. The highest normal data acquisition rate of the optimal averager section of the instrument is 1 vector per spacecraft frame (RIM). In order to achieve the higher desired data rates, the spacecraft CDS reads out the instrument memory where the recursive filter stores its vectors every other minor frame (4/3 sec). The data were stored in CDS data buffers DBUM1A and DBUM1B and then returned by CDS memory readouts (MRO). The readout of the CDS memory did not go smoothly and had to be done several times due to telemetry errors. In the end, pieces of memory were reconstructed by hand with various address ranges coming from various MRO files. Even after hand splicing segments of data together, many small segments of data contained dropouts and spurious values. In order to recover from the effects of over-filtering of the data, the data were deconvolved with the filter response function in the frequency domain. This process is not tolerant of data spikes or gaps. In regions of lost or contaminated data, data were interpolated from the nearest good values. The interpolation scheme used was unable to seamlessly replace missing data. The effect of this procedure is to introduce ringing in these areas during the deconvolution. Filter ringing also occurs at the ends of the interval in the FFT when the data are transformed back to the time domain. The crossover from CDS DBUM1A to DBUM1B occurred at closest approach. At the crossover, a minor frame timing shift occurred. In other words, even though all of the data in DBUM1A and DBUM1B are sampled at 4/3 second, the separation between the two buffers is 2 seconds. The FFT method does not tolerate uneven sampling. In order to get around this problem, and not have FFT edge effects at closest approach, the entire dataset was resampled to 2/3 second by linear interpolation for the filter recovery and the data from both buffers were deconvolved together. After the proper signal amplitude and phase were recovered, the offsets were removed. The instrument removes the best guess of the sensor zero levels prior to despinning the data and passing it to the optimal averager. Any residual DC offset is then spun up to the spin frequency by the despinning process. In order to remove the offsets from the despun data, the spin plane components were notch filtered in the frequency domain. The data were Fourier transformed and the background spectra was determined. In the frequency band affected by spin tone, the real and imaginary components of the transform were separately set to match background levels. The data were then transformed back into the time domain. The spin aligned sensor offset was determined and removed using the values determined from the cruise data. After the data were transformed back into the time domain, the data that were added during the interpolation to even 2/3 second sampling were removed. Only values at times when data were actually take remain in the dataset. Data at the ends of the dataset where strong FFT ringing occurs have been excluded. Attitude control data necessary to transform the magnetometer data into geophysical coordinates were acquired only for the few seconds during the taking of each of the Ida images. These very sparse AACS data were linearly interpolated between values. Since the data were already despun, spin phase information was not required. CONFIDENCE_LEVEL_NOTE These data have been highly processed to obtain the final dataset. Both the magnetometer data and the AACS data required to transform the mag data into geophysical coordinates are of lower than average quality. The dataset should not be used in wave studies. Much of the processing has been done in the frequency domain and the spectra clearly show the effects. Overall, the field values are probably correct to 0.01 nT and the AACS angles to better than a degree. The segment of data that were recorded to tape (16:53 ->17:22) do not carry as strong a warning but users should be wary of power at the spin period or its harmonics. The recorded data interval provided a mechanism for verifying the filter recovered data properly reflect the source signal at the sampling resolution. Both types of data were simultaneously acquired for about 30 minutes. When superimposed on each other, the filter recovered data accurately trace the high resolution data but they appear to have a high frequency cutoff that is less than would be expected from 4/3 second samples.