Overview of Signal Path, Block Diagram and Science Quantities
This section provides a brief overview of the signal paths and processing through the EFW instrument. A more detailed discussion of the design and functionality appears in subsequent sections.
The potential differences between each of the six spherical sensors and the spacecraft (labeled V1 through V6) are driven by unity gain preamplifiers near the sphere and the signals are sent down their respective boom cables to the Instrument Data Processing Unit (IDPU) where they enter the Boom Electronics Board (BEB). These signals are transferred directly to the Digital Filter Board (DFB) where AC coupled versions of the signals are generated in analog circuitry. The sensor signals from opposing booms are also differenced on the DFB board to provide AC (>10 Hz) and DC coupled versions of the three components of the electric field. The analog signals are digitized with 16-bit cross-strapped A-D converters on the DFB. Higher frequency (10 Hz–400 kHz) wave electric field signals over the amplitude range (±40 mV/m) are transferred to the EMFISIS wave instrument from the BEB via an analog buffer for wave analysis in that instrument. The digitized sensor and electric field measurements made on the DFB are transferred to the Digital Control Board (DCB) for formatting into Survey and Burst telemetry formats. The DFB provides the three components of the DC electric field from opposite sensors, denoted E12dc, E34dc, E56dc, which are sampled at 32 samples/s and digitally filtered at 10 Hz with a dynamic range of ±1V/m and the single ended probe-spacecraft potential measurement (V1 through V6) sampled at 16 samples/s with a dynamic range of ±225 volts. These single sensor measurements when differenced on the ground can provide electric field measurements with a dynamic range up to 4 V/m. The waveform data are telemetered to the ground continuously over the entire RBSP orbit in a survey mode. The DFB also calculates for inclusion in the survey mode a variety of spectral and filter bank products. These include the complex Fast Fourier Transform (labeled SPEC) of selectable EFW field measurements over a frequency range of 1 Hz to 6.4 kHz. In the default mode, a spectrum is telemetered once every 8 seconds and is organized into 64 pseudo-logarithmic frequency bins. The cross spectra between two selected measurements (XSPEC) is also calculated and telemetered to the ground in 64 pseudo-logarithmically spaced frequency bins. The two default measurements used in this calculation are the AC coupled differential electric field measurement from the 1–2 boom pair (E12ac) and the component of the search coil along the spin axis (SCM_W). Continuous measurements of rapid variations in average power and peak values of waves, is provided in survey mode for two selectable quantities via a bank of broadband filters. These quantities are denoted FBK_1_av, FBK_1_pk, FBK 2_av, and FBK_2_pk respectively. In the default mode, these broad band filters are sampled at a cadence of 8 times per second in seven pseudo-logarithmically spaced frequency bins from 1 Hz to 6.5 kHz.
High time resolution measurements of waveforms, which cannot be continuously telemetered within the allocated EFW telemetry, are provided through two burst modes using two independent burst memories involving different programmable modes of collection and playback controlled by the DCB. The Burst telemetry nominally includes three high frequency versions of the differential electric field measurements, the six sensor-spacecraft potential measurements, and the three components of the search coil magnetometer. These modes will be described in more detail in later sections.
Sensor System and Measurement Accuracy
Spin Plane Sensors
The spin plane sensor and cable system is illustrated in Fig. 9. The sensor is a conducting metal sphere of radius 4 cm coated with DAG 213 in order to minimize work function variations over the sphere and from sphere to sphere. In the deployed configuration, the sphere is connected to a high input impedance preamplifier via a thin 3 m long conducting wire. The sensors are current biased to control their floating potential and to minimize their sheath impedance. Theoretical calculations and comparison of measurements from biased and unbiased probes on ISEE-1, CRRES and Polar have shown that current biasing can reduce errors due to variations in floating potential by three orders of magnitude. In a manner similar to CRRES, Polar and Cluster, diagnostic sweeps in voltage will be used to determine the optimum value of the bias current. The sheath impedance coupling the electric field sensors to the plasma is adjusted through biasing to be <107 ohms which is much smaller than the input resistance of the PMI OP-15 pre-amplifier input stage (1011 ohms). Similarly, the capacitive impedance of the 3-meter bare wire portion of the sensor dominates over the input impedance of the OP-15 preamplifier input stage. This sensor design dramatically limits both DC and AC voltage divider effects on input signals producing a near unity gain out to high frequencies. The frequency response of the electric field instrument ranges up to 400 kHz. The OP-15 was used on CRRES and on THEMIS. It is radiation tolerant to 50 krads of total dose and is tantalum shielded to 100 krads to meet the 2-year mission specification. The spin-plane boom cables carry power supply and biasing voltages out to the pre-amplifiers. They carry the sensor voltages measured by the pre-amplifiers back to the main electronics box.
An important error source in the measurement of electric fields is associated with the spurious photocurrents flowing between the spherical sensors and neighboring boom and sensor elements. These photocurrents produce fluctuations in the floating potential of the sphere surface relative to plasma potential. The small surface area of the very thin 3 m long wire connecting the sphere to the pre-amplifier is designed to limit the magnitude of such photo-currents to the sensor. In addition, the flow of photocurrents is controlled by voltage biasing of neighboring surfaces relative to the sphere potential. The biasing is controlled by the DCB microprocessor. The in-board surface called the guard is nominally biased at a constant value of ∼5 volts negative relative to the sphere in order to limit the flux of photoelectrons to the sensor from the nearby cable and the large spacecraft surface. The out-board surface is typically biased ∼1 volt negative to limit the outflow from the pre-amplifier housing to the sensor. Precise values will be determined through processor controlled bias sweeps and evaluation of the measurement accuracy at different points along the orbit during commissioning, as well as periodically during the mission.
The fact that the spin axis of the spacecraft is pointed nearly towards the Sun contributes to an especially sensitive electric field measurement in the spin-plane. This orientation results in nearly uniform solar illumination of the spin plane sensors over a spacecraft rotation. Consequently, the photocurrent to the spin plane sensors is also nearly constant over a spacecraft rotation. The electric field data will be less accurate when (1) the spacecraft is shadowed by the Earth so that the photo-currents necessary to produce a stable potential reference for the probes are not present, (2) during very brief periods when the thrusters on the spacecraft are firing and the spacecraft is surrounded by a thruster plume, (3) and during periods after attitude maneuvers when the boom cables are oscillating about their equilibrium points.
The spin-axis electric field is about an order of magnitude less accurate than the spin plane measurements since the spin axis booms are shorter by a factor of ∼7 and have a comparatively greater error contribution from asymmetries in the spatial configuration of the spacecraft potential. A new feature of the spin-axis stacer booms on RBSP is that the length of the booms can be controlled such that, after on-orbit calibration, the sensors are positioned on the same equipotential of the spacecraft potential reducing offset errors to the electric field. Thus, the booms are not “popped” outward to a pre-determined length under the spring force of the stacer, but are restrained during deployment by the tension in the cable, which is slowly played out in a controlled fashion by a motor. The motor does not have a retract capability, but does allow the lengths of the two opposing spin axis booms to be independently adjusted.
Accuracy of the measurement along the spin axis booms is affected if the anti-sunward spin-axis boom is partially shadowed by one of the four solar panels as the spacecraft rotates. At these times, because there will be perturbations per spin period, the spin axis electric field component will be calculated from samples obtained from those intervals when the spin axis boom is not shadowed. In addition, spacecraft attitude is controlled by operators to minimize this shadowing by maintaining at least 15 degree offset of the spin axis from sun pointing.
On CRRES, during some of most intense electron injection events, the spacecraft differentially charged hundreds of volts relative to the plasma. These voltages exceeded the power supply rails of the pre-amplifier electronics and, as a consequence, the electric field instrument sometimes saturated during interesting electron acceleration events. On RBSP, the project instituted a rigorous program of electrostatic cleanliness to insure that the surface of the spacecraft was conducting and that conductive paths tied all major exterior surfaces together including the solar panels and thermal blankets. The electrostatic cleanliness program was designed to attempt keep spacecraft surfaces electrical equipotentials to within 1 volt and avoid differential charging.
The EFW instrument provides a measurement the spacecraft potential which can be used to estimate of the thermal plasma density in the plasmasphere and identify interval of spacecraft charging. The EFW instrument telemeters “single probe potentials”, V1s,V2s,…,V6s, which consist of the measured potential differences between individual probe and the spacecraft. The spacecraft potential is calculated on the ground by summing the single probe potentials from sensors on opposite sides of the spacecraft. This sum removes the differential electric signal due associated convection, waves, and other ambient plasma processes. In lower density cold plasmas (<100 cm−3), the spacecraft potential relative to the ambient plasma is primarily determined by the balance between photoemission associated with solar illumination and the thermal current associated with thermal electrons in the plasma. The current-biased probes provide a stable reference (within 1–2 volts of plasma potential) for the measurement of the spacecraft potential relative to the ambient plasma. Interpretation of the spacecraft potential in terms of the properties of the thermal electron plasma must include the fact that the spacecraft is typically in the Langmuir probe “focusing regime” (Pedersen 1995; Pedersen et al. 2008). In these circumstances the spacecraft potential scales as the logarithm of the electron density with a small contribution from temperature effects. The spacecraft potential is typically calibrated periodically against other measures of plasma density. As discussed in the Introduction, Fig. 6 presents a calibration of the spacecraft potential versus density as determined from measurements of the upper hybrid frequency by the EMFISIS instrument (Kletzing et al. 2013) over the density range from 1 to 10−3 during one orbit. The spacecraft potential measurements provide a higher time resolution (DC—100 Hz), but less accurate, measurement of the density and thermal plasma structure than the one obtained from the upper hybrid frequency measurement. The spacecraft potential algorithm for calculating density is inaccurate during intervals of negative spacecraft charging induced by very strong fluxes of 10 eV to 1 keV electrons near apogee. Periods of strong spacecraft charging may be identified by comparing the spacecraft potential to the HOPE measurement of H+ and O+ over the 10 eV to 5 keV range (Reeves et al. this issue).
Sensor Frequency Response
The frequency response of the spin-plane sensor system is governed by the properties of the plasma sheath around the sensors and by the ability of the preamplifier to drive the long 50-meter cables. The source impedance of the plasma sheath surrounding the sensor forms a voltage divider with the input impedance of the preamplifier system. The switch from resistive to capacitive coupling is associated with a roll-off in frequency response from unity near DC to 0.6 at ∼100 Hz. A second higher frequency roll-off in the sensor/cable system occurs as a consequence of the capacitive loading of the preamplifier by the long spin-plane cables. This occurs near 300–400 kHz and defines the upper end of the frequency response for the signals sent to the EMFISIS instrument.
Figure 10 presents plots of the gain and phase shift of the spin plane and spin axis boom pre-amplifier system and cables at the entrance to the EFW IDPU. Note that the spin axis booms have a higher frequency response as a consequence of the decreased attenuation due to the lower capacitance of the shorter spin axis boom. Additional phase shifts for these measurements are associated with the 5 pole Bessel anti-aliasing filters. For Bessel filters the phase shift is linear as a function of frequency. For time domain signals this results in a constant time delay for a given roll-off frequency. Table 3 provides the time delays for different anti-aliasing frequencies for the 5 poles Bessel filters appropriate for all quantities in the burst and survey data.
On-orbit Calibration and Analysis of the Electric Field Measurement
Calibrations of the electric field instrument can be provided in several ways. During quiet times when geophysical electric fields are small, the electric field measurement should approach E=−V
sc×B where V
sc is the spacecraft velocity relative to the rest frame of the Earth and B is the measured magnetic field. For electric fields that are constant over a spin period, the measured signal from orthogonal booms pairs should be 90 degrees out of phase and scale in magnitude with 100 m boom lengths. During periods of strong flows in the near Earth plasma sheet, the velocity moments, V
, from the HOPE plasma instrument can be used to estimate E=−V
Historically it has been found that the most accurate quasi-static electric field determinations are provided by least-squares spin-fits to the electric field measured by the spin plane sensors in the rotating frame of the spacecraft. In the rotating system, a constant electric field in inertial space appears as a sinusoidal signal. This fit determines the optimum values of the amplitude and phase of this sinusoidal signal, determining the magnitude and direction of the electric field projected into the spin plane. Such spin fits are especially accurate because they use the large number (484) of measurement points gathered over one spin period and because they remove work function voltage differences between the probes and other errors which appear as constant offsets to the sinusoid. After one fit, the fit may be further optimized by removing noise points far from the fit value (i.e. more than 2 standard deviations) and repeating the process as needed to obtained the desired accuracy. Perturbations to the sinusoidal electric field signal observed by the rotating sensors due to angle dependent spurious photo-currents or wake effects can removed by selectively masking data points over those rotation angles most susceptible to the perturbations.
Analysis of data from the CRRES spacecraft, which also had a sun-pointing spin axis, shows that the dominant error source for the quasi-static spin plane electric field is typically due to the effect of attitude uncertainties in the Lorentz transformation from the spacecraft frame to the inertial frame of the Earth. The attitude uncertainty of <30 will allow EFW to meet or exceed the spin-plane boom measurement sensitivity requirement above 3 R
E radial distance (either of 0.3 mV/m or 10 % of the amplitude of the electric field, which ever is larger). The anticipated Spin plane electric field sensitivity after spin fits and other ground analysis efforts is expected to be 0.1 mV/m (or 10 % of the amplitude) based on CRRES measurements (Rowland and Wygant 1998).
Spin Axis Electric Field Measurement
The spin axis measurement is especially accurate for higher frequency wave measurements (>100 Hz to 400 kHz) where it provides for a full three-dimensional electric field measurement. However, for quasi-static and low frequency measurements, the measurement is less accurate than the spin plane booms since the spin axis booms provide a shorter measurement baseline. In addition, the spin axis sensors are closer to the spacecraft and sample a larger fraction of the spacecraft charging structure especially at apogee where the Debye length is larger than the spacecraft dimensions and the charging structure has a slower fall-off with radial distance. Since this structure is asymmetric with respect to the Earth–Sun line, a portion of this charging structure appears as a differential signal in the low frequency electric field. These asymmetric contributions to the spin axis measurement are partially mitigated on RBSP, since the lengths of the individual spin axis booms may be incrementally adjusted (outward only) during on-orbit calibrations to remove the contribution from the spacecraft potential to the electric field. Finally, the spin axis booms may be shadowed by solar panels, which rotate and intermittently shadow the anti-sunward boom varying the photoemission and potentials of booms surfaces. This shadowing can produce periodic spikes repeating at the spin period lasting <1/16 second, which must be removed from the data. This photoemission variability is strongest for the anti-sunward spin-axis boom when the spin axis of the spacecraft points directly at the Sun. This shadowing near the electric sensors is mitigated on RBSP through on-orbit attitude maneuvers that control the angle between the spin axis of the spacecraft and the Earth sun-line so that it is always larger than 15 degrees. Measurement of the quasi-static spin-axis electric field requires extensive ground analysis and validation. The quasi-static spin axis electric field measurement is most useful for large electric fields associated with ULF waves, shocks, and injection events. In addition, it is designed to provide useful information on the structure of the large-scale convection electric field during major geomagnetic storms when the field is especially strong and penetrates to lower L-value, and higher plasma densities (Wygant et al. 1998; Rowland and Wygant 1998). It is complemented when possible by the derivation of the spin axis electric field from the measurement of the two spin plane components of the electric field and the full three dimensional magnetic field measurement (from EMFISIS) and the constraint that (for large scale slowly varying fields) E•B=0.
IDPU Electronics: Boom Electronics Board
The signals from the six electric field sensors enter the IDPU onto the Boom Electronics Board (BEB). At this point, the DC signal has a gain of near unity and a dynamic range of ∼225 volts. The sensor electric field signal is transferred to the Digital Filter Board (DFB) for signal processing and A-D conversion. Signals from opposing booms are subtracted and sent as buffered wave electric field signals with a frequency response up to 400 kHz to the EMFISIS instrument. The required noise level of the signal to the EMFISIS instrument at the output of this buffer is 10−14 V2/m2 Hz at 1 kHz and 10−17 V2/m2 Hz at 100 kHz. This was verified at EFW instrument level testing and also during interface testing with the EMFISIS instrument itself. The required minimum dynamic range is ±30 mV/m. This value is met and exceeded by the actual dynamic range of ∼40 mV/m.
One of the principle functions of the BEB is to generate and transfer the different control voltages that are used for current biasing of the probe and to control the potential of surfaces near the sensors. These signals are transferred to each of the six boom deployment units and then out via wires in the boom cables to each of the sensors. These bias control circuits consist of the current bias circuit, the usher bias circuit, and the guard bias circuit. Each of the sensors is independently controlled by a set of bias circuits.
The current bias circuitry results in the injection of a microprocessor-controlled bias current from the sensor surface into the plasma to control the sensor floating potential and the plasma sensor-sheath resistance. This is achieved by setting the operational point on the current-voltage curve of the sensor/plasma sheath. In low-density plasmas, the bias current is generally adjusted to be a significant fraction of the total photocurrent to the probe. The injected bias current may be adjusted over a range between ±500 nA to an accuracy of 0.2 % by the bias circuitry. The bias circuit has a high frequency roll-off (6 dB) at about 300 Hz. The optimum bias current will be determined by on-orbit bias sweeps during the commissioning phase and also at a lesser cadence throughout the mission. Depending on on-orbit calibration results, the bias current may have different values in the low-density plasma near apogee and in the higher density plasma deep within the plasmasphere.
The usher and guard bias circuits generate voltages that equal the sum of a sphere output plus a microprocessor-controlled voltage offset. The roll-off for the usher signal is 300 Hz and that for the guard circuit is 100 Hz. These circuits control the perturbing sensor-boom photocurrents over the frequency range from DC to somewhat less than 100 Hz where the sensor-plasma sheath impedance is resistive. The usher and guard voltages are transferred via the boom cable to their respective conductive control surfaces on the preamplifier sensor housing. Thus, the potential of these control surfaces relative to the sensor are a constant microprocessor-controlled value. The voltages may be adjusted to a constant value over a range of ±40 volts with a 16-bit DAC. The usher and guard bias potentials will be optimized during the commissioning phase and periodically throughout the mission. Typical values range between zero and several times the photoelectron e-folding energy or ∼5 volts. Typically, once adjusted to their optimum values, the guard and usher voltages remain constant for months to years. The commanded values of the bias, guard, and usher voltages are included in EFW housekeeping telemetry, as are the parameters of the bias sweeps.
Search Coil Signals
The three components of the wave magnetic field from the EMFISIS search coils (Kletzing et al. 2013) are provided to the EFW instrument via analog lines where they are digitized along with the electric field signals. The EFW instrument science measurements focus primarily on the lower frequency (<250 Hz) portion of the wave signal. The maximum response for the EFW search coil signal is at about 100 Hz but on-orbit measurements indicate that large amplitude waves can be detected in the >2 kHz range. Search coil data is incorporated into EFW spectral products and also burst data collections. The gain (left hand label) and phase (right hand label) responses of the search coil were calibrated relative to the electric field signals during bench tests and also on the spacecraft. The results of these calibrations are shown in Fig. 11. The blue and red traces of the phase response simply refer to the nature of the signal from the signal generator that provided the spectral power for optimum calibration. A sinusoidal signal was used at low frequency (blue) and a broad band “white noise” signal was used for high frequency calibration (red).
Digital Filter Board
The Digital Filter Board (DFB) and associated firmware was designed, fabricated, and tested by a team from the Laboratory for Atmospheric and Space Physics (LASP) at the University of Colorado led by Robert Ergun. The digital filter board is responsible for analog processing of sensor signals, A-D conversion, and anti-aliasing, as well as onboard calculation of a variety of spectral and cross-spectral products.
Time Domain measurements
As indicated by Fig. 8 diagram, and as previously discussed, the analog signals from each of the six spheres are passed to the DFB where the signals from opposing boom pairs are differenced and filtered in analog circuitry to provide three analog signals proportional to the electric field. In addition, higher gain AC coupled versions of the analog signals are generated to more sensitive wave measurements.
The analog filtering roll-off is performed by 5 pole Bessel filters with a linear phase shift with frequency. This linear phase shift in frequency space corresponds to a constant time delay for signals in the time domain and gives an undistorted waveform.
After A-D conversion, signals are digitally filtered with 5 pole low pass Bessel filters at an adjustable Nyquist frequency. The Bessel filters have a near linear phase shift as a function of frequency. This results in a constant group delay over almost the entire band-pass, which minimizes distortion of the waveform of the analog signal. This property is particularly useful when performing interferometric timing between opposite probes. The digital filtering is implemented in FPGA based firmware. The filtering algorithm consists of a recursive algorithm that allows filtering with programmable (2N) Nyquist frequencies. The maximum filter frequency is 6.5 kHz. The total group delays for different sampling frequencies are presented in Table 3.
There are three contributions to phase shifts between the “natural” and telemetered electric field waveforms. One is the phase shift due to the sensor-plasma sheath, one is the sensor-cable frequency response/phase shift, and the other is associated with the anti-aliasing filters of the DFB board. The first two contributions are shown in Fig. 10. The last contribution consisting of the constant time delays, due to the linear phase shift of the analog 5 pole Bessel filters set at 6.5 kHz, and also the time delays due to the FIR 5 pole Bessel filters discussed above and presented in Table 3 should be included during ground analysis.
The higher frequency digitized electric field waveforms (up to 16.4 k samples/s) are transferred to the Data Controller Board (DCB) for waveform recordings in either of two burst memory systems. The burst data sampling rates, duration of the burst, and selection of specific time intervals to record into memory, as well as, which of the intervals in memory to telemeter to the ground, are all controlled by the microprocessor in the DCB.
The DFB produces the three components of the 16-bit digitized electric waveform sampled at 32 samples/s (E12_SVY, E34_SVY, and E56_SVY). It provides six single-ended measurements of the potential difference between the individual electric sensors and the spacecraft sampled at 16 samples/s (denoted V1_SVY through V6_SVY). These quantities are described in Table 4. In addition, it provides higher time resolution data inputs for either of the two instrument burst modes. The burst 1 mode includes three components of the electric field (E12_B1, E34_B1, E56_B1), six components of the spacecraft–sensor potential (V1_B1 through V6_B1), and three components of the AC magnetic field from the EMFISIS search coil magnetometer (SCM_U_B1, SCM_V_B1, SCM_W_B1). The burst 2 mode data returns a similar complement of electric field (E_12ac_B2, E34ac_B2, E56ac_B2), search coil (SCM_B2, SCM_2B2, and SCM_B2) and single ended measurements (V1ac_B1 through V6ac_B2) with the exception that in the default mode the single ended and electric field signals are AC coupled with a higher gain.
The DFB receives signals from the EMFISIS search coil magnetometer via an analog interface to provide measurements of the 3-D wave magnetic field. These data may be used as input to the EFW spectral products or filtered and sent to the DCB to be included in EFW time domain burst recordings.
The DFB also receives DC fluxgate magnetometer analog signals from the EMFISIS instrument in order to allow for rotation of EFW data into a magnetic field aligned coordinate system on-orbit. This data stream is not normally included in EFW telemetry. However, for purposes of redundancy, the digitized fluxgate data can be included in EFW telemetry in the event of failure of the EMFISIS digital section. This redundancy is motivated by the important role that the DC magnetometer data play in the RBSP spacecraft attitude determination and analysis of field aligned coordinates for the different science instruments. There is a similar redundant data path for EFW electric field data above 10 Hz via the EMFISIS instrument. The EFW preamplifiers and signal buffers to the EMFISIS instrument will remain powered in the event of a failure of the EFW digital section.
Survey Mode Spectra and Cross Spectra
In the survey mode, the DFB provides a continuous stream of electric and magnetic field spectra over the frequency range of ∼1 Hz to 6.5 kHz at a cadence of one spectrum every 8 seconds. It calculates 2048-point complex FFT spectra and cross spectra every 0.125 second over a frequency range from 1 Hz to 6.5 kHz. Prior to the FFT calculation, the waveform data are multiplied by a Hanning window. Subsequent to the FFT calculation, in order to save telemetry, the real (R) and imaginary (I) parts of FFT frequency components are pseudo-logarithmically compressed into 36, 64 or 112 frequency bins. In the default mode, the spectra accumulated over 1/8 second are averaged together over an 8 second period to provide the real and imaginary components of selected wave forms in 64 frequency bins every 8 seconds. The quantities that can be selected for the spectra and cross-spectra are presented in Table 5.
In addition, the complex cross-spectra are calculated for four selected pairs of quantities. This allows us to calculate, on the ground, spectra of the estimated complex E/B ratio, the magnetic field-aligned Poynting flux, wave polarization, or, the correlation (or anti-correlation) between density and the magnetic field magnitude fluctuations. These calculations are especially designed to routinely determine the properties of whistler waves which provide strong signal to noise ratios and have pulse widths nearly comparable to the 125 ms spectral sampling interval The results of these calculations can be compared for brief selected periods of time to similar calculations using the full three dimensional electric and magnetic field obtained from the burst recordings. The number of frequency bins, the sampling interval may be optimized on the basis of these comparisons.
The DFB board uses a CORDIC algorithm to rotate wave electric and magnetic field data into a magnetic field aligned coordinate system based on the direction of the average magnetic field. The rotation is performed in the spin plane of the spacecraft to obtain E
perp, which is the component of the electric field perpendicular to the magnetic field direction in the spin plane and E
par or the component of the electric field closest to parallel to the magnetic field within the spin plane. The spin axis electric field component is not included in the rotation. The wave magnetic field data is also rotated in order to determine the projection of the wave magnetic field orthogonal to the background magnetic field direction and orthogonal to the direction of E
perp. The data from this field-aligned coordinate system may be input into the spectral or cross spectral products in survey mode to provide spectral information on wave Poynting flux parallel to the average magnetic field and also the velocity of propagation of waves along the magnetic field. CORDIC rotated data may also be incorporated in burst formats for playback to the ground. A cross-correlation algorithm allows calculation of the phase-lag between opposite boom pairs, which can provide information on the velocity of propagation of small-scale plasma structures propagating from one sensor to the opposing sensor with a baseline separation of 100 meters. This technique is most feasible for structures with scale sizes of several kilometers or less and travelling at less than 800 km/s. The fidelity of these calculations will be determined by comparison to the full three dimensional electric and magnetic field wave forms which are downloaded from burst mode recording.
Survey Mode Filter Bank Data
The DFB also provides measurements of the average power and peak amplitudes from either 7 (default) or 13 logarithmically spaced band-pass filters, implemented in FPGA based firmware, between 0.8 Hz and 6.5 kHz sampled at a cadence of 8 samples/s. Details of the filter banks are summarized in Table 6. Two quantities may be input into the filter banks. In the default mode, V12ac, the AC coupled measurement of the electric field from sensors 1 and 2 in the spin plane is input into filter bank 1 And SCMw, the spin axis component of the search coil magnetic field serves as input into the second filter bank. The broadband filters provide a continuous stream of survey mode measurements over the entire orbit providing information on rapid time variations and bursty wave activity over a broad frequency spectrum. These measurements are also used as an input to burst trigger algorithms.
Data Controller Board and Burst Memory Modes
The Data Controller Board (DCB) is responsible for the reception of instrument commands, timing and status information from the spacecraft command and data handling system. The DCB implements commands, keeps track of internal data acquisition timing and monitors spacecraft status. It controls instrument data compression, telemetry formatting and sends all EFW telemetry to the spacecraft for transmission to the ground. It controls instrument operational modes including bias sweeps on the BEB board. It also handles burst memory sampling formats, triggering, recording, playback, and memory management for the DFB. It controls all boom deployments. It calculates space weather products including electric field spin period sine wave fits and the spacecraft potential. It also controls the redundant magnetometer back-up mode in the event of an EMFISIS instrument failure.
The DCB is based on a general purpose microprocessor, the Z80, an 8-bit processor implemented as FPGA firmware. The FPGA chosen for flight is an Actel RTAX2000S. It includes a Z80 processor as an instantiation of the CAST Inc. Z80 IP-core. The CPU in the FPGA is supported by external 32K×8 boot PROM, 128K×8 SEU-immune static RAM and 128K×8 EEPROM. Additional logic within the FPGA handles the processor bus control and provides registers for accessing the various sections of memory. Also included in the FPGA are the instrument interfaces, the SDRAM controller, the FLASH memory controller, error detection and correction logic for SDRAM and FLASH memories, the spacecraft interface logic, DMA and data management control, analog housekeeping control, and timing/time-tagging support.
Like the other EFW IDPU boards, the DCB resides on a 6U VME board, connecting to a custom instrument backplane. Power is received through the backplane connection, which is also used to communicate with the other IDPU boards: the LVPS, the BEB and the DFB (described elsewhere). The DCB receives its operating power from the LVPS as well as a number of analog housekeeping values. The DCB controls the boom deployment power switching in the LVPS. The DCB controls sensor biasing and modes in the BEB. The DCB controls the operating modes of the DFB and directs DFB data products via DMA to DCB memory.
The DCB serves as the EFW interface to the spacecraft command and data handling system. The interface includes a UART based telemetry interface (115.2 KBaud), a timing signal (1 Hz Clock and “Spin Pulse”), and a UART (115.2 Kbaud) based command interface.
For Burst-1 data storage, 32 GBytes of non-volatile FLASH memory is installed on the DCB. This bulk memory is composed of eight 4 GByte memory modules. Each memory module includes eight 512M×8 Micron FLASH Memory devices. Each of the 4 GB modules is separately powered. FPGA based logic is used to streamline the intensive DMA data-transfer operations between FLASH and the DCB-based random access memory banks including error detection and correction. The CPU is responsible for continuous FLASH memory module management.
For Burst-2 data storage, 256 MBytes of local dynamic RAM memory is installed on the DCB. The SDRAM can be powered on and off to save power and clear problems. FPGA logic manages the SDRAM refreshing.
Burst Memory and Operations
The EFW instrument has two independent burst memory systems, which focus on different aspects of high frequency waves and involve different modes of data selection. The measurement quantity formats, data rates, modes of burst playback and triggers are programmable and are controlled by the DCB. Analog filtering, analog to digital conversion, and digital anti-aliasing occur on the DFB. The solid-state burst memories are located on the DCB. Tables 4–7 present the format, sampling, signal ranges, and A/D resolution of science quantities for burst 1 and burst 2 modes respectively.
Burst Mode 1
The Burst-1 mode uses the 32 GB flash memory for the recording and playback of high time resolution waveform data. Telemetry quantities and sampling rates are summarized in Table 7. In the default mode, the sampled quantities include three components of the DC coupled electric field, three components of the search wave magnetic field from the EMFISIS search coil, and the 6 values of the potential differences between each of the 6 sensors and the spacecraft. All these quantities are sampled at 512 samples/s in and low pass filtered at ∼200 Hz with a constant time delay 5 pole anti-aliasing filter in the Burst-1 default state.
An alternative format that could be selected includes ac-coupled versions of the signals with higher gains.
The sampling rate is programmable and with selectable values ranging between the default of 1 sample/s to 16.4 k samples/s in factors of two increments with adjustable low pass anti-aliasing filters near the 80 % of the Nyquist frequency or 6.4 kHz at the highest sampling rate of 16.4 k samples/s.
At the nominal rate with continuous sampling, the 32 GB memory would be filled in ∼20 days. The telemetry allocation for playback is 3932 bits/s. This allows for about 4 % of the orbit averaged total non-compressed burst 1 data to be played back. This corresponds to about 40 minutes of data at the default Burst-1 sampling rate. The minimum size burst is 2 Mbytes or about 5 minutes in the default mode. Higher sampling rates may be used and the corresponding fraction of the data played back is less.
Data compression algorithms implemented in the DCB flight software increase the actual volume of B1 and B2 recorded data played back on orbit by factor of 2–4.
The selection of the “most interesting” time intervals for high time resolution bursts and their relation to time intervals of strong particle energization and to intervals when high time resolution data are not obtained is of crucial importance to the EFW science goals.
In the default Burst-1 playback mode, the time intervals of the Burst-1 mode data to be returned to the ground for analysis may be selected by designated EFW scientists at the EFW SOC and uplinked to the EFW instrument for playback of Burst-1 data intervals on subsequent passes. This “human intervention” allows EFW scientists to identify interesting time intervals on the basis of EFW survey data and any data available from other instruments on the RBSP spacecraft. In addition, information from ground based investigations, other spacecraft, or the BARREL investigation will be incorporated into the burst selection decision as available.
Alternatively, the time intervals of the Burst-1 playback data may be autonomously determined by software algorithms in the DCB flight software. Some of the information from the EFW instrument to be used to decide the burst intervals of the most value include the DC electric and magnetic fields, information on intervals of intense wave activities using spectral data and also power in electric fields and magnetic fields measured by broad-band filters which are sampled at 8 samples/s.
Candidate burst intervals of special interest include: (1) Waves observed during substorm injections and dipolarizations; (2) Waves observed during interplanetary shocks impacts on the magnetosphere; (3) Waves observed during observations of energetic electron microburst and other intervals of electron flux enhancements or loss revealed by the RBSP ECT MAG-EIS high time resolution measurement of energetic electrons; (4) Intervals of magnetic conjunction with the BARREL balloon campaign designed to investigate microbursts and other modes of electron precipitation and loss; (5) Periods of high wave activity when the two RBSP spacecraft are “closely spaced” sampling the same wave field; and, (6) Science targets of opportunity including compressed front-side magnetopause crossings during major storms when energy transfer into the magnetosphere is the largest.
Burst Mode 2
The second burst waveform mode (Burst-2 or interferometric mode) in the default mode typically samples each of the six individual electric field sensors at nominal rate of ∼16.4 k samples/s. Telemetry formats and sampling options for this mode are summarized in Table 8. The signals from the individual probes provide interferometric timing information of small scale structures as they move over the spacecraft. These measurements allow the determination of the direction and velocity of propagation of the structure. Similar bursts have proven useful in determining the properties of ion phase space holes and other small scale structures (Ergun et al. 2001; Dombeck et al. 2001; Cattell et al. 2001) as they propagate along the magnetic field.
Burst-2 data collection is autonomously triggered with an algorithm in flight software on the DCB that continuously evaluates a quality factor which is a weighted linear combination of a variety of measured wave electric and magnetic field parameters. The most important of these parameters are the measured power and the peak in the electric and magnetic fields in the DFB broadband filters. The weights for the different field parameters used to compute the quality factor are stored in look-up tables that can be updated by ground command. Similar automated burst trigger algorithms have been used by electric field instruments on the FAST, Polar, Cluster, and THEMIS missions.
Burst-2 data are stored in 256 MB of rad-hard SDRAM memory. In the EFW default telemetry mode, this burst mode can playback data covering 0.1 % of the orbit period or for about 300 seconds without data compression. The burst information is played back in a prioritized list based on the value quality factor. This insures that the “best” events are telemetered to the ground. Burst 2 data can also be collected and played back on the basis of time tagged commands.
Low Voltage Power Supply Board
The function of the Low Voltage Power Supply (LVPS) board is to generate voltages for the analog and digital sections of the IDPU and to controlling boom deployment motors. Receiving an unregulated supply voltage from the spacecraft power bus, the LVPS provides stable fixed (referenced to signal ground) voltages to power analog and digital circuitry throughout EFW. It also produces ±225 volts for the floating ground reference and ±15 volts floating power for the sensor pre-amplifiers. In addition, the LVPS receives two additional switched power lines from the spacecraft power bus to allow for deployment of the EFW mechanisms. When the spacecraft activates these primary circuits, the EFW DCB controls power switches within the LVPS to pop pin-pullers, fracture frangibolts and activate deployment motors.
The dynamic range of the electric field measurement in the spin plane must accommodate voltage swings that include the amplitude of the largest anticipated electric fields (1 V/m) integrated over the boom separation distances (50 m) plus the largest (slowly varying) floating potential of spacecraft that is expected in low density plasmas (40 volts) or during electron injection events (<10 volts). This motivates the use of a floating power supply system in which the sensor pre-amplifiers power supplies (±15 volts) are referenced to a floating ground reference driven by a low pass (<500 Hz) signal at the potential of the pre-amplifier output. This allows low frequency excursions of ±225 volt power supply which track the sphere outputs for potential variations below 500 Hz. Higher frequency signals are tracked by the ±12 volt power supply.
This power supply system is used to power the pre-amplifiers at the ends of the booms. Special care in LVPS layout has been taken to insure that the power supplies are especially low in noise. This includes placement of traces and ground planes. The power converter frequency is fixed at 200 kHz so that noise is generated only at this frequency and its integer multiples.
The power supplies for the sensor pre-amplifiers and the buffers to the EMFISIS system are designed to continue to function in the event of a failure of the EFW IDPU digital section. (As discussed, similar precautions have been taken by the EMFISIS team to provide a back-up analog data stream for the fluxgate magnetometer data via the EFW instrument).
Spin Plane Booms and Deployment Units
The electric field spin plane boom deployment units were designed, constructed and tested by the Space Science Laboratory at the University of California at Berkeley. Members of the boom engineering team have been responsible for the booms on the Air Force S3-3, ESA GEOS 1 and 2, NASA/ESA ISEE-1, Swedish Viking, Swedish Freja, NASA/USAF CRRES, NASA Polar, FAST, and THEMIS, and ESA/NASA Cluster spacecraft, as well as boom systems on dozens of sounding rockets.
The RBSP spin plane booms consist of two pairs of centripetally deployed booms in the spin plane of the spacecraft. As shown in Fig. 9, the booms are terminated with spherical sensors, which are separated by 100 meters. The boom deployment units are mounted on the periphery of the spacecraft at 90 degree intervals. In the stowed configuration, the spin plane boom cable is wound around a spool within the deployment unit and is deployed under the influence of the “centrifugal force” associated with the spacecraft rotation. The mass of one of the spin plane boom deployment units including sensor and cable is ∼2.0 kg. The linear mass density of the boom cable is ∼3.6 gm/m. Each deployment unit contains a rotating wire storage spool, a metering wheel based cable deployment assembly, a DC brush motor, over tension and end of wire indicators. A micro-switch is pulsed on each turn of the cable spool allowing flight software in the DCB to monitor and control the deployed length and to match the deployed lengths of opposing pairs of booms.
At the end of the deployed multi-conductor cable, a preamp enclosure is attached and contains the preamp electronics board, as well as acting as two of the photoelectron control surfaces (usher and guard). A fine wire exits the preamp housing and connects to a spring-loaded spool enclosed within a spherical probe of radius 4 cm. Three meters of fine wire are wound onto the spool mechanism inside the spherical probe and the wire is intended to deploy by centrifugal force as the boom elements are deployed.
The motor in the base of each unit is shielded with mu metal to limit leakage of stray magnetic fields. Magnetic shielding is external to the motor, and does have some effect on performance, by its magnetic field short-circuiting effects. The drive mechanism is fully enclosed to provide EMI shielding, as well as to keep debris out. EMI filters are used on both power lines.
Spin Axis Booms
The electric field axial boom deployment units were designed, constructed and tested by the Space Science Laboratory at the University of California at Berkeley. Members of the boom engineering team have been responsible for the booms on the Air Force S3-3, ESA GEOS 1 and 2, NASA/ESA ISEE-1, Swedish Viking, Swedish Freja, NASA/USAF CRRES, NASA Polar, FAST, and THEMIS, and ESA/NASA Cluster spacecraft, as well as boom systems on dozens of sounding rockets.
The RBSP axial booms consist of one pair of booms to be deployed along the spacecraft spin axis. As shown in Fig. 12, the booms can be deployed to a maximum length of 7 meters each and are terminated with spherical sensors with internal preamplifiers giving a maximum tip-to-tip separation of ∼15 meters. The boom deployment units are mounted back-to-back within a carbon composite tube installed along the center line between the top and bottom decks of the spacecraft structure. In the stowed configuration, the axial probe is held against the spacecraft deck in a caging mechanism. When released, the probe and whip assembly rotate up along the spacecraft spin axis. At that point the whip/probe assembly can be deployed along the spin axis at the end of a stacer boom element. The stacers are spring loaded helical beryllium-copper coils of metal with an outward spring force. Unlike designs for previous spacecraft, the deployed length of this boom is adjustable. The purpose of this capability is to optimize the position of the sensors relative to the asymmetric electrical equipotential associated with the spacecraft. By placing the sensors on the same spacecraft equipotential surface, error offsets to the electric field measurement may be removed. In this design, the outward spring force of the stacer is restrained by the outer Kevlar braid of a cable through the center and along the length of the stacer tube. The in-board end of this cable is wound around a motor driven spool. Powering the motor allows the spool to unwind under the outward spring force of the stacer element. As a result, the deployed boom length can be adjusted with a resolution of 0.5 cm but only in one direction—there is no retraction capability. The lengths of the booms will be adjusted in stages during the commissioning phase of the mission.
The sensor consists of a 75 cm-long, tapered (4.8 to 7.0-mm) graphite-coated (DAG-213) whip with 4 cm radius sphere at its tip. The preamp enclosure is similar to that found on the spin plane booms, and contains the preamp electronics board, as well as acting as two of the photoelectron control surfaces (usher and guard). The preamp and whip assembly is mounted to the outboard end of the graphite coated (DAG-214) main stacer. A two-stage deployment assist device (DAD) is spring-loaded, and serves to start the stacer deploy. The stacer element is formed during deployment through two sets of roller nozzles to provide for lateral stability of the stacer when fully deployed.
As with the spin plane booms, the motor in each unit is shielded with mu metal to limit leakage of stray magnetic fields. The drive mechanism is fully enclosed to provide EMI shielding, as well as to keep debris out. EMI filters are used on both power lines.