Originally published in:
Space Sci. Rev., 71, pp. 563-582, 1995.
The Magnetometer on the Polar Spacecraft is a high precision instrument designed to measure the magnetic fields at both high and low altitudes in the polar magnetosphere. This instrument will be used to investigate the behavior of field-aligned current systems and the role they play in the acceleration of particles and the dynamics of the fields in the polar cusp, magnetosphere, and magnetosheath. It will measure the coupling between the shocked magnetosheath plasma and the near polar cusp magnetosphere where much of the solar wind magnetosphere coupling is thought to take place. Moreover, it will provide measurements critical to the interpretation of data from other instruments. The instrument design has been influenced by the needs of the other instruments for immediately useable magnetic field data and high rate (100+ Vectors/Sec) data distributed on the spacecraft. The design provides a fully redundant instrument with enhanced measurement capabilities depending on available spacecraft power.
While no region of the Earth's magnetosphere is completely unexplored, some regions are poorly understood. One of the most poorly understood regions is the high altitude polar magnetosphere and polar cusp. Previous magnetic investigations of this region included those on the HEOS-2 spacecraft [Hedgecock, 1975] and HAWKEYE [Van Allen, 1992]. These spacecraft contained restricted payloads and provided low data rates. Nevertheless,these missions and previous missions at lower latitudes have allowed us to develop a qualitative picture of how the solar wind interacts with the high latitude magnetosphere, as sketched in Figure 1.
A cutaway sketch of the principal regions of the magnetosphere illustrating the current systems (solid arrows) and flows (dashed lines).|
The magnetic field configuration and the current systems are linked by Maxwell's laws. Specifically, the vector curl of the magnetic field is proportional to the current density. If these current systems are stationary or move slowly with respect to the velocity of the spacecraft, and if they are planar, we can use the change in magnetic field to establish the current density. At low altitudes the velocity of the spacecraft is certainly much faster than the velocity of the current system. However, at high altitudes we must rely on measurements such as those of the plasma velocity to determine the speed of boundaries and current sheets. The necessity on this mission of measuring both small and large fields very precisely and to provide data to the other instruments places special demands on the magnetometer.
The magnetic fields investigation has an array of scientific objectives which drive its measurement objectives together with a service role in support of the in situ investigations on the POLAR spacecraft. The magnetometer is designed to be able to measure small amplitude field oscillations in both low background fields and large ?? the shears in the magnetic fields associated with field-aligned currents both in the ionosphere and in the distant magnetosphere, and to measure the currents associated with pressure gradients throughout the magnetosphere and its boundary plasmas.
Magnetospheric current systems were first proposed by Balfour Stewart  as the cause of the short term variations seen in ground based magnetic records, and have been the subject of intense interest by auroral and magnetospheric physicists ever since. Many of these currents flow at relatively low altitudes in the ionosphere. These strictly ionospheric currents include the equatorial electrojet, the Sq system driven by atmospheric tides and the auroral electrojet driven by magnetospheric processes. Birkeland  was the first to propose that currents could extend along field lines into space but his suggestion lay untested for over half a century until Zmuda et al.  found magnetic disturbances in the auroral ionosphere with a magnetometer on satellite 1963-38C at a height of 1100 km. Shortly afterward, Cummings and Dessler  provided the correct interpretation of these disturbances, that they were Birkeland's field-aligned currents. The measurements of 1963-38C and those of TRIAD at 800 km enabled the field-aligned currents to be mapped throughout the auroral oval [Iijima and Potemra, 1976]. MAGSAT data in 1979 at 400 km altitude showed that closure of the field aligned currents, the Pedersen current, and the orthogonal or Hall current could be inferred from their weak effects on the magnetic field [Zanetti et al., 1983]. The POLAR spacecraft will provide both measurements of unprecedented precision on the low altitude ionosphere but also enable exploration of these current systems and to the boundaries of the magnetosphere.
High latitudes play a special role in magnetospheric physics. At low altitudes, the auroral ionosphere is the region into which the bulk of the energy entering the magnetosphere from the solar wind ultimately is dissipated in the form of Joule heating of the upper atmosphere. These low altitude currents in turn are linked to the stresses applied by the solar wind at high altitudes by field aligned currents. These field aligned currents are the critical element in the coupling of the magnetosphere and the ionosphere because currents are subject to resistive instabilities among others which could decouple the magnetosphere from the ionosphere and allow one to effectively slip with respect to the other. If flux tubes break in this way, the plasma on a flux tube is no longer frozen to the field line and field line potential drops appear that can accelerate particles along the magnetic field. Some are accelerated downward to excite the upper atmosphere; others are accelerated upward to populate the magnetosphere. If we are to understand the physics of the magnetosphere at high latitudes, we must uncover the full range of physical processes occurring in these field aligned currents.
The Dungey model of the reconnecting magnetosphere for southward interplanetary magnetic fields (top) and northward interplanetary magnetic fields (bottom).|
The high altitudes processes acting in the high latitude magnetosphere may be no less important than those acting at low altitudes. Figure 2 shows the Dungey reconnection model of the magnetosphere for both southward (top) and northward (bottom) orientations of the interplanetary magnetic field (IMF) [Dungey, 1961]. When the IMF is southward reconnection of the magnetospheric and magnetosheath magnetic field is believed to take place near the equatorial plane at low latitudes. The magnetic flux is eroded from the dayside magnetosphere [Aubry et al., 1971] and transported to the magnetotail. In steady state the magnetic flux cannot build up indefinitely and a second reconnection region develops in the magnetotail which allows magnetic flux and plasma to complete a circuit to the dayside magnetopause. If the direction of the interplanetary magnetic field is variable, then the transport to the tail is time variable and the reconnection rate in the tail episodically increases and decreases possibly sensibly decoupled from the fluctuating IMF. These variations have been used in a popular model of the substorm process [Russell and McPherron, 1973].
Plasma flows, as expected from reconnection near the subsolar point, have been seen frequently [Paselmann et al., 1979; Sonneup et al., 1981]. However, it is not clear that reconnection always is a subsolar phenomenon even for southward IMF. The phenomenon, known as the flux transfer event (FTE) which was postulated to be the sign of pately, transient reconnection [Russell and Elphic, 1979] could also be caused by high latitude reconnection [Podgorny et al., 1980].
The bottom panel of Figure 2 shows the expected configuration for magnetic reconnection when the IMF is northward. In this case the magnetic field from the solar wind is draped over the magnetopause and finds itself along an antiparallel magnetospheric magnetic field behind the polar cusp in both the northern and southern hemospheres. If reconnection can take place at these times and evidence for accelerated flows for northward IMF have been seen [Gosling et al., 1991], then magnetosheath plasma can mix with magnetospheric plasma in newly closed field lines and the low latitude boundary layer, which appears to be such a mixture, has a ready explanation [e.g. Song and Russell, 1992].
Figure 3.Magnetic fields and standard deviations observed by ISEE-1 on a pass through the magnetosphere.|
The magnetospheric environment is characterized as much by its time varying fields as by its magnetic topology. Figure 3 shows the magnetic field and standard deviations of the field on a typical pass of ISEE-2 through the magnetosphere. Initially on the far left of the figure, ISEE-2 is in the solar wind and the standard deviations and field components are small. When the bow shock is crossed at 1615 UT the field and fluctuations rise. These values are roughly steady until the magnetopause is crossed at 1930 ( vertical line). Inside the magnetosphere, the field strength rises and the wave level drops. Waves are still important but they become a small perturbations on a much stronger background field. Around periapsis, where the field is strongest, the ISEE spacecraft digitized the magnetic field less sensitively than at high altitudes, so that the level shown from 0000 to 0400 represents instrumental limitations. The POLAR magnetometer's resolution has been significantly improved to minimize such digitization problems.
These objectives serve to introduce some of the required characteristics of the magnetometer on the POLAR spacecraft. The magnetometer needs to be sensitive enough to measure the magnetic perturbations associated with the various current sheets and wave processes in the high latitude magnetosphere and it needs a large dynamic range to measure these deep in the magnetosphere. Moreover, since it is critical to the success of the entire POLAR mission, the instrument needs to be redundant so that no single point failure inhibits data return. With these requirements in mind we designed the instrument whose properties follow.
The instrument consists of two triads of orthogonal fluxgate magnetometer sensors (Figure 4) mounted on a 6 Meter boom with associated analog and data processing circuits mounted inside the spacecraft (Figure 5). The key elements of the instrument characteristics are provided in Table 1.
Three views of the flipper sensor assembly: view down the boom (left); exposed sideview (middle) and baseplate (right).|
|Fig. 5. Electronics unit with board placement illustrated.|
The sensor assemblies are fabricated with ring core sensors (Gordon and Brown ) built at UCLA. Each sensor contains a drive winding and a sense/feedback coil surrounding a magnetically permeable core. Carefully timed currents in the drive coil force the core into saturation twice during each drive cycle. External magnetic fields upset the symmetry of these saturations resulting in the generation of even harmonics signals whose amplitude is proportional to the field. These signals are detected by the sense coil and used in a feedback system to maintain the core near zero field
The individual sensors are mounted in an orthogonal orientation in a small sub-assembly. This fixes the interactions between the sensors due to their close positioning. This sub-assembly will be mounted in the sensor assembly flipping mechanism such that it will be rotated about one of the axes. This minimizes the changes in sensor interactions that were observed in previous designs (ISEE and GALILEO), where only two sensors were flipped.
The flipping mechanism used is similar in design to those flown on Explorer 35, Pioneer 9, ISEE 1 and 2, UK-AMPTE, and Galileo. This mechanism uses electrically heated bimetallic springs to operate a series of levers and rotate the sensor shaft by 90 and back again.
The sensor assembly has been modified to accommodate the UCLA sensors and to rotate, about one axis, the orthogonal assembly of three sensors (Figure 4). The completed sensor assembly is mounted to the boom on brackets designed to correct for the 3 droop in the boom. One of the sensors that flip is aligned with the spacecraft spin axis. Thermal protection and EMI/EMC control are provided by MIL blankets.
The electronics package for the MFE is illustrated in Figure 5. This box is mounted inside the spacecraft and contains the fluxgate analog circuits, analog to digital converters, digital circuits, spacecraft interface circuits and power conditioning circuits for the MFE. Most of the 4.5W of power is dissipated in this unit with less than 100 mW going to the sensors. The power and weight distribution are fairly uniform. The power boards contain slightly more weight and power dissipation than the other boards.
The electronics are constructed of radiation hard devices obtained from the common buy or purchased from the Harris rad hard line of parts. All passive parts are purchased from Mil-975 or have approved NSPAR's. The printed circuit boards used are multilayer boards designed at UCLA but manufactured to NASA quality requirements by a local company.
Analog Fluxgate Circuits
The drive and sense circuits of the basic magnetometer.|
The analog circuits for the fluxgate sensors are based on a flight proven design. Changes in the design have been made to accommodate the radiation hard active components and to reduce the weight, power, and noise levels of these circuits. The block diagram for the basic analog circuit is given in Figure 6. This circuit uses a drive frequency of 9KHz and detects the 18KHz second harmonic generated by the sensors when an external magnetic field is present. In order to improve the linearity and stability of the magnetometer, a feedback system is utilized which maintains the sensor core at near zero field. The linearity of this circuit has been tested to better than 1 part in 104, the limits of the test equipment. The fluxgate output voltage is filtered to protect against aliasing. This output can be sampled directly when high range data is required. A 8X amplifier provides the low range data. The frequencies used in the magnetometer are developed and synchronized to the digital circuits in order to minimize/control internal noise distribution. Figure 7 presents typical noise levels measured for the analog system (no digital system operating).
Power spectrum of the noise level measured for the analog system.|
The MFE magnetometer drive circuits operate at a 9KHz frequency derived from the MFE crystal oscillator. The drive circuits are tuned to provide minimum power dissipation and minimum second harmonic distortion. Deriving all MFE timing from the crystal oscillator allows the MFE to control the position of the ADC samples to minimize noise pickup. In addition metal partitions are provided in the MFE electronics to provide structural elements and noise shielding.
In order to maintain the fidelity and accuracy of the basic fluxgate circuits, we have utilized the 16-bit Crystal analog to digital converter (ADC). The circuits for the ADC were developed utilizing designs and layouts suggested by Crystal Semiconductor Corporation. The ADC board for the MFE provides for power regulation, signal selection, and houses the magnetometer antialiasing filters and gain stages for one triad. Two identical boards provide redundancy and the ability to operate both sensors simultaneously (this mode requires an additional 1.2W of power). During normal operations, one of the boards will be powered down. Either ADC board can select which magnetometer to filter/amplify. Once the Mag selection is made, the ADC can collect samples from the axes or the housekeeping information provided at rates up to 1000 Vectors/Second. In order to synchronize the data to the drive frequency (9000Hz) for minimum noise pickup, ADC conversions can occur only on 110 msec boundaries. A metal partition (see Figure 5) separates the analog circuits(fluxgate and ADC's) from the rest of the circuits to maintain the quality of these signals.
Additional enhancements to the quality of the Magnetometer data can be accomplished using digital data processing techniques. By oversampling and averaging the analog data, we can reduce the apparent noise level of the system and also gain stable, identical aliasing filters for the main MFE data. In addition, some corrections can be incorporated that make the data immediately useful when it is received on the ground (i.e. offsets, gains, and pointing directions). We will also use the available processing power to compress the data. This will enable us to use our allocated telemetry bandwidth to support the snapshot capabilities of the instrument. Two 80C86 m-processors provide the computing power required to process normal data, capture snapshots of high rate data, compress and provide the data to the spacecraft for transmission to earth. These processors control the operation of the complete instrument, including command detection and execution, data sampling, filtering, storage, and compression. Each processor board has crystal controlled clocks, a watchdog timer, local memory (RAM and ROM) and the 80C86 processor. Each processor board also has a dedicated power system, spacecraft communications circuits, mass memory, and ADC board. The power circuits are adequate to power the processor boards plus one of the sensor triads. Redundant control of the fluxgates, flipper control relays, and power crosstrap relays are provided. The system block diagram is given in Figure 8.
System block diagram.|
The watchdog timer was included to protect the system from single event upset errors which could cause errors in program execution. If the timer is not reset periodically (7 milliseconds), a full system reset is initiated.
The spacecraft interface boards contain the standard Polar interface circuits for the following functions:
A. Command/Time Receipt
B. Telemetry Data, Clocks, Frame Pulses
C. Hi-Rate MFE data (>100 V/S)
D. Spin Phase Clock, Reference
All inputs from the spacecraft are received on both boards for redundancy. Redundant outputs to the spacecraft are provided on each board. Only one processor (via software control) will be providing active data at any given time.
Two additional circuit elements were placed on the interface boards, a Processor/Processor interface, which provides a high speed path that the two processors use to communicate and pass data, and the 512K Byte Mass memory module.
Block diagram of the power system.|
The power system designed for the MFE operates at 50.4KHz and provides redundant power by using two separate circuits with minimal cross strapping (Figure 9). Each system is dedicated to one µP, ADC and provides power for one of the Magnetometers. The actual Magnetometer that can be powered is controlled by the power crossover relays. An additional command by the µP that turns on the Magnetometer is required. The use of the crossover relay insures that either power source can supply either Mag. This provides additional redundant capability with minimum impact on system design. The power circuits are separated from the rest of the instrument by a metal partition for EMI control as shown in Figure 5.
Magnetometer Commands and Modes
Commands can be sent to the instrument to control the operation, data processing, and data formats used. All commands are received by both processors and operated on if required. The MFE only uses minor mode commands for instrument control even though some commands have major impacts on the instrument operations. The minor mode command allows the use of the parameter fields used in the MFE commands. The command codes used employ Hamming codes for error detection/correction purposes. The interpretation and execution of all commands is under processor control. The defined commands can be used to set the operating modes and parameters for the MFE. Time is sent to the instrument as a command approximately once a minute. The estimated rate for all other commands is one command per hour.
The MFE has three spacecraft power commands, one for each side of the magnetometer and one for flipper power. When the magnetometer is powered up, additional commands are provided to control the instrument. Except for memory loads and a reset command, all commands are buffered for execution at the beginning of the next major frame.
MFE instrument commands provide control of the power to the magnetometers, analog to digital converters, and flipper heaters. Additional commands provide for time tagging, memory loads, memory dumps, control over spacecraft communications, data selection and data analysis techniques, and data compression control.
The memory load capability will be used to load and update rotation matrices, offsets and gains so that the data returned is in spacecraft co-ordinates with instrument offsets and gains corrected.
On-board Data Processing
Block diagram of the on-board processing flow.|
A flow diagram of the onboard data processing is presented in Figure 10. The high rate vector sampling is handled by the interrupt routine in order to obtain consistent sampling rates. The interrupt rate is controlled by a selectable count rate which is set at 1000 interrupts/sec. Approximately 50% of the processing power is utilized by the interrupt routine to sample, convert, and process the vector data. Since the instrument design provides for only one interrupt, all housekeeping and synchronization must be compatible with this timing. This data will be decimated to 100 Vectors/Sec for the snapshot and lower rate data processing. High rate MFE data (4 vectors/minor frame) which is distributed onboard the spacecraft will also be derived from this data and distributed.
During normal operations, the 100 vectors/second data is sent to the second processor for Snapshot collection and data compression. Less than 20% of the processing time of this processor will be required to service its interrupt.
There will be operating modes where a single processor can collect, format and distribute data in case of a failure. During these times, the normal data corrections and highest speed sample rates may have to be modified in order to obtain useable data.
The high rate data stream provides data that is in nominal spacecraft co-ordinates with the offsets removed. The accuracy of this nominal spacecraft system will depend on the accuracy of the alignment of the spacecraft axes and the sensor array. Corrections for gain factors will not be applied to this data. Alignment accuracy and gain factors are expected to be accurate to within 1%. The normal data will process the high rate data to correct for the gains and rotate to actual spacecraft co-ordinates. In addition, this data will be filtered, decimated to 10 vectors per second rate, and compressed, thus providing room in the telemetry for snapshot data. There is a mode of operations without data compression that provides g vector/second.
Gordon, D. I., and Brown, R. E.: 1972, IEEE Transactions on Magnetics, MAG-8, 76-82
Kivelson, M. G., et al:.....(Galileo Mag.)
Russell, C. T.:1978, IEEE Transactions on Magnetics, MAG-16(3)
Snare, R. C., and Means, J.D.:1977, IEEE Transactions on Magnetics, MAG-13(5)
Hedgecock, P. C., Magnetometer experiments in the ESRO HEOS spacecraft, Space Sci Inst I, 61-82, 1975
Van Allen, J. A. Observed currents on the Earth's high-latitude magnetosphere, J. Geophys Res, 97, 6381-6395, 1992