The FIELDS Instrument Suite for Solar Probe Plus

2.1.1 The V1–V4 Electric Antennas

Four voltage sensors (V1–V4) are deployed in nearly orthogonal, co-linear pairs slightly behind the plane of the spacecraft heat shield (the Thermal Protection System or ‘TPS’), as shown in Fig. 4. To function properly as a double-probe electric field measurement, these sensors must be coupled to the plasma through a photoelectron current; i.e. they must be in sunlight. This orientation also keeps the V1–V4 antennas out of the spacecraft wake and ensures a minimal perturbation to the spacecraft interaction with the solar wind.

Figure 6 shows a CAD drawing of one of the units; each unit is identical, with some small differences in the mounting hardware. The primary sensor consists of a 2 m long, \(1/8''\) diameter Niobium C-103 thin-walled tube (called the ‘whip’). All but the last 8 cm of the whip are exposed to full sunlight and will reach high temperatures (>1300 °C) at SPP perihelion. The whip is clamped to a 30 cm Molybdenum ‘stub’ element which acts as an electrical and thermal isolator and the signal from the whip is fed through with a molybdenum wire to the hinge and preamp below. A chevron shaped C-103 heat shield covers the final 8 cm of the whip and the entire stub. This creates a substantial shadowed area that radiates excess heat into space. Additionally, the shield, whip, and stub are all isolated from each other with sapphire, a good thermal insulator. Both of these features greatly reduce the thermal input into the base, where conductors and other materials must be below 230 °C. This design has been modeled and verified in the laboratory. The preamp sits near the bottom of the antenna mechanism, to minimize stray capacitance from the cables. The antenna system is stowed back against the spacecraft for launch and deploys by spring force and is rate-limited by a flyweight brake system. After release, it will take a few seconds to deploy to the final configuration. Each sensor will be deployed individually, with the FIELDS electronics operating, to aide in calibration and characterization of the measurement.

Open image in new window

2.1.2 The V5 Voltage Sensor

A simple voltage probe (‘V5’) will be mounted on the SPP magnetometer boom, deployed in the umbra behind the spacecraft (and is therefore coupled to the plasma through thermal electrons rather than photoelectrons). While this sensor sits in the spacecraft plasma wake (described above) and will see the low frequency signatures of that interaction, it will provide good capacitively-coupled voltage measurements of the radial electric field \(E_{||}\) present in plasma waves and will help constrain the knowledge of the electrostatic center of the spacecraft.

The V5 design is shown in Fig. 7. Two short sensor elements extend from a preamplifier box and are electrically isolated from the box by PEEK fittings. The two tube elements are tied together at the preamp (i.e. it is not a differential measurement).

Open image in new window

2.1.3 Electric Preamplifiers

The four main antennas are connected individually to the V1–4 preamps and the fifth antenna on the mag boom is connected to the V5 preamp. The preamps provide low-noise high-impedance inputs, voltage gain, and low impedance outputs. As shown in Fig. 8, the V1–4 preamps provide three outputs: the HF 20 MHz bandwidth output to the Radio Frequency Spectrometer (RFS), the MF 1 MHZ output to the Time Domain Sampler (TDS), and the LF 64 kHz output to the Antenna Electronics Board (AEB) and Digital Fields Board (DFB). The HF amplifier chain is a new design consisting of a FET input buffer followed by a wide bandwidth op amp providing gain and driving a 50 ohm terminated coax output. The LF and MF signals are provided by a second unity gain op amp, as used in legacy designs (e.g. THEMIS, RBSP). This op amp is powered by a supply referenced to a “floating ground driver” on the AEB (see Sect. 2.2.1 below), and provides a signal range of ±70 V from DC to 300 Hz and ±10 V from 300 Hz to 1 MHz. The V5 preamp does not include the HF chain and consists of the single legacy op amp design providing LF and MF outputs.

Open image in new window

2.1.4 Fluxgate Magnetometers

The two fluxgate magnetometers (MAGs) for SPP are similar to the triaxial, wide-range, low-power, and low-noise magnetometers built by Goddard Space Flight Center (GSFC) for MAVEN, Van Allen Probes, STEREO, etc. This line of flight magnetometers totals 79 instruments to date starting with IMP-4, launched in 1966. For SPP, the MAGs will provide data with bandwidth of ∼140 Hz, sampling at 292.97 Sa/s, as part of the FIELDS instrument suite. The measurement dynamic range is ±65,536 nT with a resolution of 16 bits. The primary science objectives addressed by the magnetometers are determining the structure and dynamics of the magnetic fields at the sources of the fast and slow solar wind, contributing to the study of the coronal processes that lead to heating of the solar corona, and exploring the roles of shocks, reconnection, and turbulence in accelerating energetic particles.

The sensor design for the SPP MAGs (Fig. 9) will provide maximum thermal isolation from the boom, which will undergo considerable temperature variation, depending on its location in the umbra or in sunlight. Kinematic mounts limit the heat transfer across the feet of the sensor. The heritage of the kinematic mounts derives from the Juno magnetometers, where they were used to ensure that the sensor temperature variation was limited. Heater power is provided by a proportional AC heater to reduce the temperature variations of the sensors and to provide survival heating. The heater is synchronized to a frequency provided by the MEP. The sensor mass requirements have lead to use of a lightweight composite base.

Open image in new window

Each MAG sensor has a corresponding electronics board in the SPP FIELDS MEP. For enhanced reliability, the outer MAG board is controlled by the DCB, and the inner MAG board by the TDS. The control component on each board is the rad-hard Aeroflex FPGA that contains all of the MAG logic functions and on-board SRAM. These functions include command handling, telemetry packet formatting, ADC readout, auto-ranging algorithm, DRIVE clock generation, and housekeeping readout. MAG produces a data product at the required cadence of 1 message per 0.874 seconds. The ranging algorithm selects one of the four ranges (±1024, ±4096, ±16,384, ±65,536 nT) based on the ambient magnetic field. The sensor connection to the electronics boards is accomplished by a tuned circuit, including the harness, which is calibrated at the GSFC Acuña Magnetometer Test Site. Because the sensor temperatures will be substantially lower than standard GSFC magnetometer sensors and will vary significantly due to changes in the spacecraft Mag boom orientation relative to the Sun, calibration will be performed over a wide range of temperatures.

The MAG sensors are mounted on the SPP magnetometer boom (Fig. 10), where their relative separation and close proximity to the spacecraft compromise their gradiometric functionality, and thus the ability for an accurate removal of any spacecraft field contamination at the outer MAG. This enhances the importance of the magnetic cleanliness testing being implemented for the spacecraft and payload.

Open image in new window

2.1.5 Search Coil Magnetometer

The search coil magnetometer (SCM) will measure all three components of the AC magnetic signature of solar wind fluctuations, from 10 Hz up to 50 kHz and a single component from 1 kHz to 1 MHz. The wide bandwidth and dynamic range allows FIELDS to investigate transients caused by interplanetary shocks and reconnection, the turbulent cascade beyond the electron kinetic scale, but also numerous plasma wave modes.

The SCM instrument consists of a triaxial search-coil that has a solid technical heritage in several past missions (Dudok de Wit et al. 2011). Nearly identical instruments are being built for the Taranis and Solar Orbiter missions. Each sensor consists of a magnetic core with a winding whose voltage is proportional to the time-derivative of the magnetic field (Seran and Fergeau 2005). Two sensors of SCM cover the ELF/VLF frequency range from 10 Hz–50 kHz. The third one is a dual-band sensor that covers both the ELF/VLF and the LF/MF (1 kHz–1 MHz) ranges. The three sensors, each of which is 104 mm long, are mounted orthogonally on a non-magnetic support, see Fig. 11.

Open image in new window

Two challenges raised by Solar Probe Plus are the low temperature environment on the magnetometer boom, and the unusually large dynamic range of the instrument. The latter is needed to accommodate both small-amplitude fluctuations of solar wind turbulence, and large transients near the Sun. Peak values, as scaled from observations made by Helios at distances from the Sun down to 0.29 AU, may reach 3000 nT in the ELF/VLF range. Thanks to a careful design, the dynamic range of the instrument has been increased from past models by several tens of dB, now reaching 160 dB in the ELF/VLF range, and 130 dB in the LF/MF range. SCM will be located in the shade of the spacecraft, at the end of the magnetometer boom, and thus needs a heater to keep it above deep space temperatures. To mitigate thermal losses, the instrument will be wrapped in an insulating MLI layer, with very compact design. The SCM design is very compact; in particular, the preamplifier has been miniaturized by 3D Plus, and will be located inside the foot of the instrument.

The sensitivity and instrument response of the SCM are illustrated in Fig. 12. The sensitivity is sufficient to observe small-amplitude solar wind turbulence in the inner heliosphere, and properly distinguish Elsässer variables, while also capturing large transients (hence the low gain −50 dBV/nT in the ELF/VLF range). The analog signals in the ELF/VLF range will be processed by the Digital Fields Board (DFB), which will deliver either spectra or continuous waveforms up to 150,000 Sa/s. The LF/MF signal will be processed by any of the RFS, DFB, or TDS receivers. The survey data products will be spectral matrices, which give access to the polarization, and waveforms of up to 293 Sas/s for all three components. The latter will be merged with the DC magnetic field, as measured by MAG, into one single composite magnetic field product.

Open image in new window

2.2.1 Antenna Electronics Board (AEB)

The signals from the four V1–V4 electric field antennas and the V5 sensor enter the MEP on the Antenna Electronics Boards (AEB1 and AEB2). At this point, the DC signal has a gain of near unity and a dynamic range of 115 volts. The sensor electric field signals are transferred to the Digital Filter Board (DFB), the Time Domain Sampler (TDS) and the Radio Frequency Spectrometer (RFS) for signal processing and digitization. These two AEB units are functionally almost identical with AEB1 on the FIELDS1 side of the instrument and AEB2 on FIELDS2 (see block diagram above). AEB1 processes signals from V1, V2, and V5, while AEB2 processes V3 and V4 signals.

One of the principle functions of the AEBs 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 five sensor units. On the forward sensors (V1–V4), these bias control circuits consist of the whip current bias circuit, and the stub and shield voltage bias circuits. On the tail sensor (V5), these bias control circuits consist of the antenna bias current circuit and the stub (box) voltage bias circuit.

The current bias circuitry implements 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 broad range of illumination levels achieved on SPP, ranging from 16 to over 500 times the solar constant at 1 AU, require the implementation of three ranges for the bias current on the illuminated sensor surfaces (V1–V4). These three ranges are controlled to 12-bit accuracy (0.025 % of full range) as follows: ±802 nA, ±14.1 μA, ±414 μA.

The current bias circuit has a bandwidth (3-dB) of 450 Hz. The current bias circuit on V5 only implements the lowest range (±802 nA). The voltage bias circuits (stub and shield on V1–V4; box on V5) allow for ±40 V of DC offset relative to the sensor potential to be applied to the given surface for sensor-to-spacecraft potential differences of ±60 V DC, primarily for photo- and secondary electron control. The voltage bias circuits have the same bandwidth (3-dB) as the current bias circuits (450 Hz).

The optimum bias currents and voltages will be determined by on-orbit bias sweeps during both the commissioning phase near 1 AU (for comparison with prior mission results) and also at a lesser cadence throughout the perihelion passes. It is expected that the bias currents and voltages will be changed using explicit commands during the inbound and outbound legs of the perihelion passes. In addition, an on-board automated bias current adjustment system successfully used in flight on the NASA Van Allen Probes mission (Wygant et al. 2013) may be used if perihelion conditions prove to be too dynamic for fixed bias settings to encompass.

The commanded values (digital to analog converter settings) controlling the bias currents of the whips and the bias voltages of the stubs and shields are included in instrument housekeeping telemetry as are the parameters of the bias sweeps. In addition to generating and controlling the sensor bias currents and voltages, the AEB also houses the floating ground driver and floating power supplies used to power the sensor preamplifiers (LF/MF stage and HF front-end), as well as additional regulation for the HF backend. This floating preamp design allows for the use of low-voltage, low-noise, low-leakage parts in the preamp design while still allowing the system to handle the tens of volt quasi-DC offsets between sensor and SC potentials due to current biasing and differences in sensor and spacecraft illumination. The floating ground driver has a dynamic range of ±100 V (required range is ±60 V), with a bandwidth (3-dB) of 450 Hz.

2.2.2 Digital Fields Board (DFB)

The Digital Fields Board (DFB) is responsible for conditioning, digitization, and processing signals from the five voltage sensors and four search coil magnetometer (SCM) coils over a frequency range of DC to ≈60 kHz, as well as production of a calibration signal to the SCM. These nine analog inputs are processed by the DFB into twenty-five digital data streams, which are then used to produce a range of time- and spectral-domain data products.

The signals measured by the FIELDS five voltage sensors are amplified by the preamplifiers and are passed to the AEB boards for low frequency signal conditioning before entering the DFB. Four analog voltage signals arrive directly from the SCM sensor, three from the low frequency (LF) SCM windings, and one from the medium frequency (MF) winding. These nine inputs are processed by the DFB into twenty-six digital data streams, which are then used to produce a wide range of time-domain and spectral-domain data products. A more detailed description of the DFB for Solar Probe Plus can be found in Malaspina et al. (2016).

Voltages measured by the five FIELDS antennas encounter the FIELDS preamplifiers and the AEB boards before entering the DFB. The four SCM signals arrive directly from the SCM sensor, three from the low frequency (LF) SCM coils, and one from the medium frequency (MF) coil. DFB analog conditioning includes separation into DC-coupled and AC-coupled signals, production of differential signals, application of anti-aliasing filters, and application of gain stages. Figure 14 summarizes these steps graphically.

Open image in new window

To digitize twenty-six signals at 150 kS/s and maintain a low-power, low-mass, radiation-tolerant design, the DFB makes use of the Teledyne SIDECAR application specific integrated circuit (ASIC) (Loose et al. 2005). The SIDECAR was designed to support compact focal plane electronics for space telescopes. A SIDECAR ASIC is currently operating on the Hubble Space Telescope Advanced Camera for Surveys, and the James Webb Space Telescope mission will use make use of the SIDECAR. The DFB does not engage most of the SIDECAR’s focal plane-specific capabilities, instead treating it as a bank of 32 16-bit ADCs (analog to digital converters).

After analog to digital conversion, digital signal processing occurs within a Microsemi RTAX4000 FPGA, converting the twenty-six available data streams into a range of time-domain and spectral-domain data products. Immediately after analog-to-digital conversion, all DC-coupled single-ended and differential signals have their cadences reduced to 18.75 kS/s using an 8-point boxcar average, and average voltage signals (DC- and AC-coupled) in the plane of the heat shield are calculated digitally: \(V_{Avg} = (V_{1} + V_{2} + V_{3} + V_{4}) / 4\).

All signals then enter one of two cascading digital filter banks (low-speed for 18.75 kS/s data, high-speed for 150 kS/s data). The filter banks apply a fifth-order finite impulse response (FIR) Bessel filter to the incoming data. They then apply an averaging FIR and decimate the data by a factor of two. This process operates recursively, resulting in multiple low-speed waveform data streams at sample rates of \(18.75 / 2^{N}\ \mbox{kS/s}\) for \(N\) from 0 to 14. High-speed waveform data streams have sample rates of \(150 / 2^{N}\ \mbox{kS/s}\) with \(N\) from 0 to 6. SCM-LF data streams start in the high-speed digital filter cascade and are passed to the low-speed digital filter cascade upon reaching 18.75 kS/s in order to make the SCM-LF data available to all low-speed and high-speed time- and spectral-domain data products. Further details on the digital filters appear in Cully et al. (2008) and Ergun et al. (2014). The DFB uses the output of these cascading digital filter banks to generate survey waveform data, which has a low cadence (relative to the ADC sampling frequency), but which has continuous coverage over the Solar Probe Plus orbit.

The DFB also produces competitively selected snapshots of high-speed waveform data using the DFB burst memory (DBM). DBM snapshots are \(N x\) ∼3.5 s long and are comprised of six (selectable) waveform channels captured simultaneously, where \(N\) is determined by the selected sample rate \(150 / 2^{N}\ \mbox{kS/s}\). The DBM is capable of producing far more data than the DCB internal memory can store. Therefore, data is down-selected using a competitive buffer scheme. In this scheme, there is one circular ingress buffer constantly being filled, several hold buffers which retain high-quality data, and an egress buffer which is read out to the DCB. Because emptying the egress buffer is a slow process relative to filling the ingress buffer, data in the ingress buffer is periodically assigned a buffer quality. Each time an ingress buffer quality is calculated, it is compared with the hold buffer quality values. If the ingress buffer quality exceeds a hold buffer quality, the ingress buffer is promoted to hold buffer status. The displaced hold buffer data will either be demoted to a lower hold buffer, or else it will be discarded. Whenever the egress buffer is emptied (completely read out to the DCB), the hold buffer data with the highest quality is promoted to egress status. In this way, DBM data quality is competitively assessed, and approximately 1 % of the highest quality data is sent to the DCB for storage and potential telemetry to the ground.

The DBM has six hold buffers, split into two parallel competitive paths. Side A contains hold buffers which compete on the basis of the Coordinated Burst Signal (CBS). Side B contains hold buffers that compete on the basis of DFB-assigned quality. These two sides alternate in their access to the egress buffer. For Side B, buffer quality is assigned by the DFB as follows. One of the six selected DBM channels is designated the ‘trigger’ channel. For every \(m\) samples of the trigger channel (called a slice), the eight points with the largest absolute value amplitudes are selected. The buffer quality is then assigned using one of three schemes: (1) the maximum of these eight points, (2) the average of these eight points, or (3) the average of the lower seven of these eight points. This third configuration is intended to reduce the number of dust impact voltage spikes captured by the DBM. See Zaslavsky et al. (2012) for a description of dust impact voltage spikes observed by the STEREO spacecraft in the solar wind and Malaspina et al. (2014) for a voltage spikes observed by the Wind spacecraft in the solar wind. As the ingress buffer fills, slice quality flags are calculated continuously. While successive slice quality flag values are increasing, designation of the ‘largest’ quality flag is reserved until slice quality begins to decrease. In this way, the DBM can be re-triggered by later, more interesting data.

The DFB also produces several spectral data products. The first of these is the band-pass filter bank (BP) data. The BP data products on Solar Probe Plus are direct decedents of those produced on THEMIS (Cully et al. 2008) and the Van Allen Probes (Wygant et al. 2013). BP data is produced from band-passed waveform data, generated by taking the difference between two adjacent digital filter bank cascade outputs. The low-speed and high-speed digital filter bank cascades produce 15 and 7 streams of BP waveform data (respectively). For each \(Q\) samples of bandpass waveform data, the maximum absolute value and the average absolute value are calculated (\(Q\) is configurable), resulting in 15-bin low-speed BP spectra and 7-bin high-speed BP spectra. The BP spectra have coarse frequency resolution but high time resolution. Up to 4 low-speed and 4 high-speed channels of BP data are selectable, with any low-speed or high-speed (respectively) waveform channel as a source. BP data is used as one DFB input to the coordinated burst signal.

The DFB also produces power spectra and cross spectra using a windowed FFT (Fast Fourier Transform) algorithm, for up to 4 low-speed and 4 high-speed channels. In all cases, transforms are calculated on 1024-point segments of waveform data (9.375 kS/s for low-speed, 150 kS/s for high-speed), with a Hanning window applied. The FFT algorithm produces both real (\(R\)) and imaginary (\(I\)) spectral components. Given input signals 1 and 2, power spectra (\(P\)) are calculated as \(P1_{k} = (R1_{k}^{2} + I1_{k}^{2})\) and \(P2_{k} = (R2_{k}^{2} + I2_{k}^{2})\) for frequency bin k. Real and imaginary cross-spectral components are calculated as \(RX_{k} = (R1_{k} R2_{k} + I1_{k} I2_{k})\) and \(IX_{k} = (R2_{k} I1_{k} - R1_{k} I2_{k})\). Real and imaginary cross spectral components can be used to derive the coherence and phase between signals 1 and 2.

Power spectra and cross-spectra results are averaged into pseudo-logarithmically spaced frequency bins, with either 56 or 96 frequencies. With 56 frequencies, the bin width (\(df / f\)) varies between 6 % and 12 %. With 96 frequencies, \(df / f\) varies between 3 % and 6 % per bin. Narrow band signals can be excluded by applying a spectral mask prior to frequency binning. This capability may be exercised to exclude, from spectral products, narrow-band electromagnetic interference due to the spacecraft or other instruments.

DFB power spectra and cross-spectra can be time-averaged, where the amount of time-averaging is configurable and may be set independently for low-speed and high-speed products. Time-averaging sets the number of 1024 point spectra to average when reporting a single spectral result. Low-speed spectral products have full data coverage, while high-speed spectral products can sample up to 12.5 % of the incoming data.

The DFB also generates a calibration signal for the SCM. The calibration signal consists of the sum of two sine waves, where the frequency of one sine wave is fixed at 9.6 kHz and the frequency of the second wave changes every 64 wave cycles. This calibration signal is designed to enable in-flight characterization of the gain and phase response of the SCM across the lower range of its sensitive frequency range.

This configuration would give good waveform coverage of electric and magnetic fields and density fluctuations to the convected ion scales over the whole orbit, and spectral density measurements to the electron cyclotron frequency.

In Burst Mode, the DFB will provide fast (up to 150 kS/s) of similar waveform quantities (3E, and 3 \(\delta\)B) and also transmit the cross-spectral matrices described above. These data would be in addition to Survey Mode data.

2.2.3 Time Domain Sampler (TDS)

As described above, in order to increase overall mission reliability, the FIELDS instrument was split into two parts, FIELDS1 and FIELDS2. Each half has some instrumentation, control over some of the FIELDS antennas and a power supply. The core of FIELDS2 is the TDS. As originally planned, the TDS subsystem was a single board processor controlled burst acquisition system designed to collect transient wave phenomena from the FIELDS electric and magnetic sensors. The new design added an interface to the SPP spacecraft command and data handling system, control of one of the two FIELDS DC magnetometers (the in-board MAGi), control of one of the two FIELDS antenna electronics boards (AEB2 controlling antennas V3 and V4) and control of one of the two FIELDS power supplies (LNPS2). In addition, the TDS also maintains a communications interface with the SWEAP instrument. The FIELDS TDS derives heritage from a similar instrument on the STEREO spacecraft (Bougeret et al. 2008).

As the core of the FIELDS2 side of FIELDS, the TDS performs a number of typical data processing functions. Especially crucial is the way the TDS keeps track of time. In normal circumstances, the spacecraft interface provides the TDS with precise information as to the mission elapsed time (MET). The TDS can use this time in its time-stamping of data. However, in order to synchronize the two halves of FIELDS, the TDS also receives information for the FIELDS1 side (from the DCB) indicating the precise MET as it was acquired by the DCB. In normal circumstances, the TDS uses the MET received from DCB as the source of its internal clock. In addition, in order to reduce noise and coordinate measurements, the various clocks throughout the FIELDS suite are synchronized. The FIELDS1 DCB produces a set of master clocks, all derived from a single internal high frequency master clock. These clocks go to the data acquisition samplers and to the various chopping power supplies. In normal circumstances, the TDS operational clocks are taken from the DCB. In the case where the DCB clocks are unavailable, the TDS also includes its own internal clock. The TDS block diagram is shown in Fig. 15.

Open image in new window

The Time Domain Sampler instrument (TDS) makes rapid samples of waveforms for the study of high frequency waves. The rapid simultaneous sampling of five channels to nominally include two orthogonal electric dipoles, a single ended electric monopole, the radial (sunward) component of the electric field provided by V5, and a single axis from the search coil magnetometer allows the study of waveforms, their distortions, and, through ground-based spectral analysis, a frequency determination which is far more accurate than an on-board filter analysis system.

The highest sampling rate is about two million samples per second (1.92 MSa/s giving a ∼1 MHz Nyquist frequency) with several lower commandable rates (e.g. 480 kSa/s, 120 kSa/s and so on). The sampling clock is derived from the FIELDS master clock provided by the DCB. At the top sampling speed, the TDS has 160 Mbps of total throughput while its nominal share of the FIELDS downlink rate is only of the order of a few hundreds of bps. The TDS achieves this large reduction in bit-rate while maintaining high scientific return by selecting events for transmission to the DCB and the ground intelligently.

Waveform events are usually triggered with a peak value in the center of the event. The duration of time-series events is commandable with a typical length of 65,536 samples or about 33 ms. Once an event is acquired, TDS flight software evaluates the event ‘quality’. Events with the highest quality will be selected for transmission to the DCB and then possibly to the ground. Each of the five channels is filtered and digitized simultaneously. The analog-to-digital converters (ADCs) provide 16-bits of dynamic range and are linear over the range. The noise level is less than 30 μV RMS (at 100 kHz) at the input to the preamplifier. The largest signal obtained is about 1 V RMS. The commercial ADCs are protected from latch-ups by a circuit breaker which shuts off power when high current is detected, allowing parasitic currents within the device to dissipate. The TDS processor will normally turn the converters back on after a programmable cooling-off period (nominally 5 seconds).

The TDS also produces a steady stream of high-value information with very little impact on FIELDS telemetry. Once per minute, TDS will give the peak value observed on each channel (during the preceding minute), the mean value, the RMS power value and zero-crossing counters which give an indication of wave frequency and the number of dust impacts in the period. 100 % of the incoming data stream (at 1.92 MSa/s) will be processed here. Examination of this low bit-rate stream will allow after-the-fact selection of events saved in the massive FIELDS burst memory.

While the TDS is a “wave” instrument, as part of the TDS, one of the data acquisition channels will be devoted to counting particles. The TDS includes an interface to the SPP SWEAP instrument. SWEAP will provide the TDS with a signal line indicating particle counts. Each pulse received will indicate a particle collected by a part of SWEAP and will result in incrementing a counter in the TDS. Each time a TDS waveform voltage is sampled (at 1.92 MSa/s), the incrementing counter will be latched and sampled. In this way, TDS events will provide a snapshot of voltage as a function of time as normal and, in addition, will provide a high resolution picture of the simultaneous particle flux as a function of time. This will give an unprecedented view of wave particle correlation.

In addition to the SWEAP particle counts, FIELDS2 and SWEAP will exchange timing and messages to allow SWEAP data acquisition to be synchronized with the FIELDS master clock. The messages will allow both SWEAP and FIELDS to collect and identify periods of high activity and interest.

2.2.4 Radio Frequency Spectrometer (RFS)

The RFS is a dual channel digital spectrometer, designed for both remote sensing of radio waves and in situ measurement of electrostatic fluctuations. The RFS receives inputs from the V1–V4 electric field antennas, using the high frequency output of the FIELDS electric field preamplifiers. Both RFS channels are digitally sampled simultaneously, allowing for calculations of auto spectra for each channel and cross spectra between the two channels. Using multiplexers to select antennas, each RFS channel can use as input either the difference between any two antennas (dipole mode) or the difference between any antenna and spacecraft ground (monopole mode). In addition to the electric field antennas, the single axis MF winding from the search coil may also used as an input to the RFS.

The RFS analog electronics are physically located on an isolated segment of the FIELDS DCB circuit board. RFS digital signal processing (DSP) is handled by the DCB FPGA and the DCB flight software. A functional block diagram of the RFS is shown at the top of Fig. 17 as a subsystem contained within the overall DCB block diagram. In the figure, one RFS channel (Channel 1) is shown, while the identical channel is hidden below.

The sensitivity and dynamic range requirements of the RFS are driven by the expected levels of the input signals. Through careful component selection and board layout, the RFS design has been optimized for low noise performance, with the aim of observing the galactic synchrotron spectrum above the instrumental noise level. Observation of the galaxy will offer an absolute calibration source and enable accurate intercalibration of the RFS with other spacecraft.

In the inner heliosphere, the intensity of solar radio emissions will be greatly enhanced, as the Solar Probe Plus spacecraft will be much closer to the source regions of the radio emission. The expected intensity of the largest Type III radio bursts, the strongest radio signals in the RFS frequency range, determines the highest anticipated signal levels seen by the RFS. Together, the galactic signal and the largest expected Type III radio bursts determine the required dynamic range of the RFS receiver. The large dynamic range is achieved with a 12-bit ADC and the ability of the RFS to operate in both high and low gain modes, for small and large signals respectively. In addition to the remote sensing signals (radio bursts and the galaxy), the RFS makes in situ measurements of the quasi-thermal noise spectrum (QTN), generated primarily by ambient electrons. Analysis of the QTN spectrum allows for very accurate determination of the total electron density as well as estimates of electron temperature and other plasma properties (Meyer-Vernet and Perche 1989).

The RFS samples the HF preamp output at the FIELDS master clock frequency \(f_{s} = 38.4\ \mbox{MSa/s}\) (note that \(38.4\ \mbox{MSa/s} = 150\ \mbox{kHz} \times 256\)), yielding a Nyquist frequency of 19.2 MHz. This cadence is determined by the EMC control plan for the SPP spacecraft. The EMC plan requires that power converters operate at specific, well-controlled frequencies, using multiples of 50 kHz starting at 150 kHz. This frequency specification limits the power supply-generated noise to specific frequency channels, enabling a “picket fence” scheme (Bougeret et al. 2008) where noise-free measurements can be made between the narrow-band power supply frequencies and their many higher order harmonics, as described above.

To measure the weakest type III radio bursts, FIELDS needs sensitivity down to the level of the galactic synchrotron spectrum (Novaco and Brown 1978; Manning and Dulk 2001), which is approximately a few \(\mbox{nV/}\sqrt{\mathrm{Hz}}\) at ∼1 MHz. In general, conducted and radiated noise generated by spacecraft and instrument subsystems exceeds these levels by orders of magnitude. This can be seen in Fig. 16, which shows several estimated signal levels together with the spacecraft electric radiated emission (RE02) requirements. Estimates of interplanetary type III emissions, plasma quasi-thermal noise, and the galactic spectrum all fall 30 or 40 dB below the peak RE02 levels. The required measurement levels are achieved by levying a spacecraft-wide electromagnetic cleanliness (EMC) program that includes maintaining frequency control on all DC-DC power converters.

Open image in new window

Open image in new window

The DSP signal chain for the RFS starts with a time series of voltage measurements recorded by the ADCs. These time series contain the physical signals of interest as well as the power supply harmonics, which must be removed via spectral processing. Ordinary fast Fourier transform (FFT) techniques are subject to spectral leakage, which would cause the power supply noise to spread from its narrow frequency peaks and overwhelm the signal throughout the RFS frequency range. To reduce this spectral leakage, the RFS implements a polyphase filter bank (PFB) digital signal processing algorithm (Vaidyanathan 1990), followed by a standard FFT. The PFB algorithm weights the data with a windowing function, splits the data into \(N\) blocks of equal size (“taps”), then adds the taps together to produce a single time series which can then be passed to a standard FFT algorithm. The combined PFB-FFT process optimizes the leakage response of the resulting spectra, effectively isolating the noise in specified frequency bins and preserving the noise-free gaps necessary for making physics measurements.

The nominal length of a RFS sample is 32,768 samples. Using an 8 tap PFB, this results in a 4,096 point time series for the FFT, which in turn yields 2048 positive frequencies. The full resolution spectra would be too large to store and telemeter, and so select frequencies are extracted from the full resolution spectra and stored in memory for downlink. Both autocorrelation and cross correlation measurements are produced from the selected bins of the spectra.

The RFS operational frequency range is 10 kHz–19.2 MHz. This frequency range is subdivided into the LFR (10 kHz∼2.4 MHz) and HFR ranges (∼1.6 kHz–19.2 MHz), with the primary science of the LFR consisting of the in situ QTN measurement, while the HFR focuses on remote sensing. The LFR sampling cadence is reduced from 38.4 MHz to \(f_{s} = 4.8\ \mbox{MHz}\), using a Cascade Integrator Comb (CIC) filter to anti-alias and downsample by 8. Because the frequency resolution of FFT algorithms is equal to \(f_{s}/N\), the lower \(f_{s}\) allows for better frequency resolution for LFR frequencies while using an identical DSP signal chain. For both the LFR and HFR, the chosen frequencies allow for a relative frequency spacing \(\Delta f/f\) of approximately 4.5 % throughout their respective frequency ranges.

2.2.5 Data Control Board (DCB)

The DCB is the primary controller of the FIELDS instrument. It serves as the primary link between the spacecraft (S/C) and the FIELDS instrument, receiving, decoding and distributing S/C commands to the FIELDS subsystems. Note that as described in Sect. 2.2.3, the TDS subsystem can recover some C&DH capability should the DCB or its power supply fail. The DCB operates autonomously using both Absolute and Relative Time Sequences (ATS and RTS), loaded from the ground prior to each solar encounter. The DCB processes instrument data into two streams, called Survey and Burst, storing the latter in a 32 GB Flash memory. At commanded downlink rates approaching 245 kbps, the DCB selects, compresses and mixes Survey and Burst data into the S/C telemetry stream.

The DCB consists of an embedded processor (CPU) with associated CPU memory (PROM, SRAM, EEPROM) and a dedicated bulk memory array (Flash). The DCB block diagram is shown in Fig. 19. The CPU is a Coldfire 32-bit IP-Core implemented in a radiation-hard RTAX-4000 FPGA. The startup Flight Software (FSW) is stored in a radiation hard 32 kB PROM. Operational software, on-board scripts, tables and other parameters reside in a 512 kB EEPROM, and are transferred to the 2 MB SRAM at start up. Instrument data messages are transferred at 4 Mbps to the SRAM via Direct Memory Access (DMA). The analog housekeeping, instrument control, S/C Interfaces, bulk memory controller subsystems and Radio Frequency Spectrometer logic are also implemented in the same FPGA.

The FIELDS instrument operates synchronously with the DCB managing the clocks and timekeeping. The 38.4 MHz Master Clock, resident on the DCB, provides all FIELDS instrument system clocks as well as the power supply synchronization and interface signal timing. The 38.4 MHz operation, common across the FIELDS subsystems, results in deterministic noise bands. Much of this ‘picket fence’ noise is either removed by selective filtering or recognized by the upstream data processing algorithms.

DCB FSW interfaces with the AEB, RFS, LNPS, DFB, TDS and MAG boards, routing commands from internal ATS and RTS scripts as well as commands from the S/C. FSW collects completed CCSDS packets from the DFB and TDS, then routes these by APID to either the telemetry (Survey) or to the Flash memory (burst). FSW controls the AEB and LNPS, collecting housekeeping analog data, limit monitoring and telemetering them. FSW collects MAG vectors, averages them and generates CCSDS packets at commanded data rates. FSW operates the RFS FPGA logic (as detailed in Sect. 2.2.4), collecting waveforms and performing FFTs. FSW performs median-filtering and production of CCSDS packets with High and Low Frequency spectra, cross spectra, phase and coherence data. FSW also performs Peak Tracking of the Low Frequency data to determine the plasma frequency and thus select the proper high resolution channels to telemeter.

The DCB exchanges burst status, electric and magnetic field information with the S/C and keeps track of Mission Elapsed Time (MET) via S/C messaging. An important part of FIELDS operations is the coordination of bursts by FSW. In order to maximize the collection of the most important and interesting scientific data, the FSW uses a linear combination of incoming data from DFB, TDS, RFS and SWEAP and a table of commandable weighting factors to determine when FIELDS should collect and save burst data in concert. This Coordinated Burst Signal (CBS) is calculated at 4 times/(NY second) and sent to both FIELDS1 and FIELDS2 and the S/C within the second.

The internal solid state recorder allows for the continuous collection of high-data rate acquisitions, the most interesting of which are selected for downlink via ground command. The 32 GB memory is composed of four separate Flash memory modules, each independently switched. FSW is capable of concurrently writing data to, and reading data from, the Flash memory at 2 Mbps. Individual Flash blocks which repeatedly fail are marked ‘off’ the FSW and skipped in subsequent processing. FSW references the Flash memory via a Virtual-to-Physical memory map, allowing large bad sections to be mapped out of the address space. A hardware-based scrubbing subsystem automatically corrects single-bit errors and flags multi-bit upsets. FSW tracks bad-blocks and provides Flash Read and Write pointers in telemetry.

DCB components have been selected to survive the seven year minimum mission length in the harsh solar environment which imposes significant radiation exposure and thermal stresses. High-quality parts (radiation hard FPGA and single-event upset (SEU) immune CPU memories) minimize the probability of upset during the crucial observation periods and maximize the probability of successful operation over the course of the mission.