A History of Vector Magnetometry in Space

Robert C. Snare

The first vector magnetic measurements in space were made with fluxgate sensors. In order to achieve higher accuracy a variety of vector magnetometers using alkali vapor cells were developed. These were abandoned as the ring core sensor was developed with improved offset and low noise performance. The improved ring core fluxgate magnetometer is used almost exclusively for vector measurements, with the exception of the vector Helium magnetometer which is used on some deep space missions. Over the years a variety of innovative sensor mounting configurations were designed to overcome the lack of spacecraft resources such as mass, power and telemetry bandwidth. The advent of modern integrated circuits such as the low power amplifier, the analog-to-digital converter and the microprocessor has enhanced data recovery from these instrument systems.

INTRODUCTION

To make the precise vector magnetic measurements in space required by scientific investigations one needs a magnetometer system, each of whose elements pushes the state of the art in linearity, low noise, stability and accuracy. These elements include the vector sensor and its associated circuits, an analog-to-digital converter (ADC) and a high fidelity data system. The sensor should be highly linear and have scale factors that are stable with temperature. The sensor should truly make vector measurements as the angular response should be equal to B cos f where f is the angle between the field vector and the sensor axis. The noise level should be low. It should maintain this low noise while providing a wide dynamic range with low, stable zero levels or offsets. The electronics that operates the sensor should be simple and reliable with no limited life components. The data system should have proper anti-aliasing filters for the data sampling rates used. The ADC should be monotonic, linear, and have high resolution. The data processing system should have the speed to process the sampled data quickly with a variety of algorithms. Although we will not discuss magnetic cleanliness in detail herein it is assumed that a magnetically clean spacecraft is available with a boom of suitable length for sensor mounting. The spacecraft should provide time and attitude data in the telemetry that will support high time resolution magnetic field measurement data reduction.

From the earliest spacecraft to those of today there has been constant improvement in the quality of magnetic measurements made in space. This improvement has been the result of the availability of greater spacecraft resources such as power, mass and telemetry bandwidth. An even greater impact is the result of advances in the electronics industry such as radiation hardened, low power analog and digital integrated circuits and the advent of microprocessor systems. A variety of innovative techniques have been employed with sensor and data systems to overcome the lack of abundant spacecraft resources. These factors affected not only magnetometers but all instruments and spacecraft systems.

Vector Sensors

The majority of vector measurements in space have been made with fluxgate sensors. The first fluxgate sensors were developed by Aschenbrenner and Goubau [1936]. This was followed by rapid development during 1940's and 1950's. Fluxgate magnetometer systems were installed in aircraft to search for submarines, to aid in making navigation charts and for geophysical exploration, Rumbaugh and Alldredge [1949].

The alternating drive current in the fluxgate sensor drive winding drives the permeable core material alternately deep into saturation. (See Figure 1). Because of the non-linear coupling due to core saturation the induced voltage in the sense winding is rich in harmonics. The amplitude of the even harmonics is proportional to that component of the ambient magnetic field aligned with the sense winding. Normally the second harmonic is filtered and synchronously detected to produce a voltage proportional to the field. The output voltage is fed back through a scaling resistor to the feedback winding on the sensor. With proper design the result is a highly linear, stable vector measuring instrument.

Figure 1. Block diagram of fluxgate magnetometer [Snare and Benjamin, 1966].

There are a number of magnetic core configurations that can be used with the fluxgate magnetometer. The fluxgate sensor that was used in the early years of space exploration was the parallel core design both in the Vacquier and the Förster configurations [Primdahl 1979]. These configurations are shown in Figure 2. The parallel core design requires high drive power. They can be troubled by high offsets that have a tendency to vary with time.

Figure 2.Parallel core magnetometers can use two configurations of windings, the Vacquier configuration shown on the left and Foster configuration on the right [Dolingov et al., 1960].

Cylindrical core magnetometers have been constructed by Schonstedt [1959] and Primdahl [1970]. The Schonstedt sensor consists of a ceramic cylinder with thin permalloy tape helically wound around the cylinder, thus the name HELIFLUX sensor. The drive winding is applied toroidally through the tube with a solenoid feedback coil around the tube. The HELIFLUX sensor exhibited low noise and stable offsets. The Schonstedt HELIFLUX sensor has been used on a number of space flights. In his cylindrical sensor Primdahl used a tube of ferrite material for the core with toroidal and solenoidal windings for the drive and feedback as shown in Figure 3. These sensors were used for sounding rocket flights and ground observatories.

Figure 3.Tabular ferrile core fluxgate sensor from [Primdahl 1970].

Aschenbrenner and Goubau [1936] constructed a magnetometer core in 1928 using a circular bundle of soft iron florist wire. This configuration was in effect the first ring core sensor [Primdahl, 1979]. The ring core design was revived by Geyger [1962]. Refinements were made at Naval Ordnance Laboratory, White Oak. The modern ring core uses several layers of thin (0.0254 mm) permalloy wound around the edge of a non-magnetic stainless steel ring. The excitation winding is toroidally wound on the ring. Sense and feedback windings are wound on a bobbin on the outside of the ring and define the sense axis of the sensor as shown in Figure 4. Both noise and offset performance were enhanced with the development of 6-81.3 Mo-Permalloy (6% Mo, 81.3 Ni, 12.7% Fe) low-magnetostriction tape material by Gordon et al. [1968].

Figure 4.Simplified view of the ring core sensor developed jointly by the Naval Ordnance Laboratory and NASA Ames Research Center, [Dyal and Gordon, 1973].

The first ring core fluxgate magnetometers used in space were in the Lunar Surface Magnetometer package left on the surface of the moon by the astronauts of Apollo 16 in April 1972 [Dyal and Gordon 1973]. There was a rapid transition in the use of the ring core during the 1970's. Today nearly all scientific vector measurements in space are made with ring core magnetometers. This is because of the low mass and simplicity of the circuitry.

Acuna [1981] reported nonlinear fluxgate response when MAGSAT sensors were exposed to large, >5,000 nT, uncompensated transverse fields. Primdahl et al. [1992] further investigated this phenomena and found that a fluxgate on a spinning vehicle produced signals with magnitude ~10-4 of the applied field at the third harmonic of the spin frequency as a result of this nonlinearity. The large transverse field must be in the plane of the ring core to see the nonlinearity. The parallel core designs such as Vacquier and Foster do not exhibit this problem.

To compensate for this problem Primdahl and Jensen [1982] constructed a three axis fluxgate sensor with the three feedback windings on the surface of a sphere. The magnetic cores with their drive windings resided inside the sphere. Thus the ring cores were not exposed to uncompensated transverse fields. Other investigators have used cubical configurations to achieve similar results.

A variety of optically pumped alkali vapor scalar magnetometers have been used in space. Casium 133, Rubidium 85 and Rubidium 87 has been employed in sensors [Ness, 1970]. The alkali vapor magnetometers utilize transitions between the Zeeman sublevels in the ground state of the atom. The energy separation for the Zeeman splitting in proportional to the magnetic field. In the Rubidium self oscillating magnetometer a light from a Rubidium lamp is passed through a filter that only passes the 7948 Angstrom Rubidium light. The light passes through a circular polarizer and then through a cell with Rubidium vapor. The light is collected with a photoelectric cell. It is then amplified, phase shifted and then applied to a feedback coil around the Rubidium vapor cell see Figure 5. The circuit will oscillate at the Larmor frequency. The frequency for Rb 87 is approximately 6.99 Hz nT-1 which for a field of 50,000 nT would be ~350 KHz. The oscillation depends upon the presence of a magnetic field and ceases when the field vector is within "12o of parallel or "7o of normal to the optical axis of the sensor. There are offsets in the frequency as the field vector is reversed. This is because the circuit oscillates with a broad unsymmetrical resonance that shifts with field direction [Ruddock, 1961]. A design that minimizes this problem is one that has a single Rubidium lamp illuminating twin absorption cells, as shown in Figure 6. The Rubidium magnetometer does not qualify to be classified as an absolute instrument with sensitivities derived directly from atomic constants. They have variable offsets that require careful calibration [Allen, 1968].

Figure 5. Block diagram of self oscillating single cell rubidium vapor magnetometer [Ruddock 1961].

If one places a scalar magnetometer in a Helmholtz coil and collects data with first no current applied then with current applied and again with the current reversed the angle of the ambient field with respect to the coil axis can be calculated see Shapiro et al. [1960]. By using two sets of deflection coils one can measure the total field and two angles, thus defining the field vector. On a spinning spacecraft only one set of deflection coils is required if the field is steady [Heppner, et al., 1963].

Figure 6.Block diagram of twin cell rubidium vapor magnetometer [Ruddock 1961].

As previously mentioned both the single cell and the twin cell Rubidium magnetometers have null zones. Explorer 10 used two twin cell magnetometers with optical axes aligned 55o to one another as shown in Figure 7. This arrangement reduced the null area to 500 square degrees solid angle.

Figure 7.Prototype of EGO rubidium vapor magnetometer showing crossed twin cells with spherical thermal cover removed. For vector measurements bias coils were mounted on the inside surface of the spherical thermal cover [Heppner 1963].

The Rubidium vector magnetometer does not provide a continuous readout of the vector. It takes a finite time to apply the deflection fields. Therefore, the data are time aliased if there are temporal variation. Another problem is that in practice the Rubidium lamp operates only from 30o to 50o C with the optimum performance within a few degrees of 40o C. Vector alkali vapor magnetometers were used on several spacecraft in the 1960's and early 1970's. The circuitry of the alkali vapor vector magnetometer is more complicated than that of the fluxgate magnetometer. After the fluxgate magnetometer's performance was enhanced in the 1970's with the introduction of the ring core sensor the vector alkali magnetometer was no longer seen on spacecraft.

Another optically pumped sensor is the Helium magnetometer, a Helium cell can be made to resonate at the Larmor frequency in a manner similar to the alkali vapor magnetometer to form a scalar sensor. The details of the Zeeman splitting is somewhat different from Rubidium. For details see Keyser et al., [1961].

The low field Helium vector magnetometer does not resonate the Helium cell at the Larmor frequency but uses the cell as a null detector. As shown in Figure 8 circularly polarized light with 1.08 mm wavelength is passed through the Helium cell. The ambient magnetic field affects the pumping efficiency of the metastable Helium population. The Helmholtz coils around the cell are driven to create a rotating magnetic field alternately in two planes. The light received by the detector is the vector sum of the rotating sweep field and the ambient field. The detector signal is then amplified and phase detected using the sweep field as a reference. This signal is then applied to feedback coils in each axis. The feedback keeps the Helium cell in zero field. The current required to null the system is then proportional to the ambient field in each axis. The maximum dynamic range of the Helium vector magnetometer is limited to a few hundred nT. The sensitivity, stability and offsets are the result of careful design and must be accurately calibrated for verification of the performance of the instrument [Slocum and Reilly, 1963 and Smith et al., 1975]. The vector Helium magnetometer has been modified to be a combination vector-scalar magnetometer. This will be first flown on the Cassini spacecraft. The scalar mode will be used at close approach to Saturn where the scalar Helium magnetometer will make absolute measurements accurate to 1 nT [Kellock et al., 1996].

Figure 8.Functional Sketch of vector helium magnetometer sensor. Absorption of circularly polarized light from the He lamp by the He cell is modulated by vector sum of ambient field H and rotating field generated by coils; resulting variations are monitored by infrared detector. For clarity, only two of the three coils are depicted [Smith et al., 1975].

Search coils have occasionally been used on spacecraft as the primary magnetometers. On spinning spacecraft a coil transverse to the spin axis can measure two components of the static field in the spin plane. The search coil magnetometer uses a long thin core of highly permeable alloy containing Nickel and Iron. Both crystaline and amorphous metals have been used. A winding of many turns of fine wire is added to the center of the magnetic core. The sensitivity of the search coil is a direct function of the number of turns of wire. The low frequency noise of the coil is the thermal noise of the resistance of the wire. Therefore, there are tradeoffs between the number of turns, wire size, the desired resonance and the mass of the coil. The amplifier connected to the coil must have low noise referred to the input. The standard design practice is to match the coil thermal noise to the amplifier input noise.

Magnetometer Designs

During the late 1940's and 1950's several scalar magnetometers were flown by the United States on V-2 and Aerobee sounding rockets at White Sands Proving Ground, New Mexico, [Maple et al., 1950, Heppner et al., 1958]. These flights achieved maximum altitudes greater than 150 km.

The first magnetometer was carried into earth orbit May 1958 aboard the Soviet artificial earth satellite Sputnik 3. The three axis fluxgate sensor was mounted in an assembly which could be rotated in two axes by servo motors, [Figure 9]. The objective was to align the primary sensor with the ambient field. This was achieved by driving the mechanism until the transverse magnetometer sensors outputs were nulled. The primary sensor field magnitude and shaft angles were telemetered to earth. The spacecraft was not stabilized and had no attitude reference. It had a rotation period of approximately 136 sec. Therefore, the magnetometer angles defined the spacecraft attitude and provided no angle information for the magnetic field, [Dolginov et al. 1960]. Thus, the instrument as designed functioned as a scalar magnetometer. The magnetometer with electromechanical orientation unit is similar to those used in aircraft as reported by Rumbaugh and Alldredge [1949]. It appears to be rather heavy but the Soviet spacecraft did not have the same mass restrictions as those faced by other countries.

Figure 9.Diagram of magnetometer orientation unit for Sputnik 3. The motors and drive gears orient the main sensor 8 parallel to the earth's magnetic field [Dolginov, 1960].

Pioneer 1 and 5 and Explorer 6 spacecraft each carried a single search coil magnetometer [Judge et al., 1960]. The search coil was chosen because of its low mass and simplicity. These were mounted normal to the spacecraft spin axis and measured the static field perpendicular to the spin axis. Additionally Pioneer 5 and Explorer 6 had sun aspect detectors that provided a spin reference for the spacecraft. The magnetic vector direction in the spin plane could then be determined. Pioneer 5 also employed a fluxgate sensor mounted parallel to the spin axis but the fluxgate did not provide data during the flight.

The first fully vector magnetic measurements in space were made aboard Lunar 1 and 2. These spacecraft used 3 separate single axis fluxgate magnetometers. The sensors were mounted orthogonally. Each sensor had its separate electronics each with unique drive and second harmonic frequencies [Dolginov et al., 1961].

The Pioneer 6, 7 and 8 were small spacecraft with a launch mass of 66 Kg. An unusual approach was used to achieve vector measurements [Ness et al., 1966]. In order to conserve instrument mass a single fluxgate sensor was mounted 2m from the spacecraft spin axis. The sense axis of the fluxgate was tilted away from the spin axis by the angle 54o 45' as shown in Figure 10. The magnetometer was digitally sampled each 120o of spacecraft rotation. Thus, any three samples form an orthogonal measurement of the field. The spacecraft spin period was one second with a magnetometer bandwidth of 5 Hz which is 3 times the Nyquist sampling frequency of 1.5 Hz. This technique works in a quiet magnetic field, but not with an active field [Fredericks et al., 1962]. Many of the most scientifically interesting magnetic fields are the most active.

Figure 10.Diagram of tilted sensor on a spinning spacecraft as used by Pioneer 6, 7 and 8.

The first Applications Technology Satellite (ATS-1) had a two axis magnetometer. Each sensor was mounted 45o to the spacecraft spin axis, (see Figure 11). The output of the two magnetometers were passed through sum and difference amplifiers to yield field components parallel and normal to the spacecraft spin axis [Barry and Snare, 1966]. The design provided redundancy as data could still be retrieved if one of the magnetometers failed.

Figure 11.The two sensors for ATS-1 were each mounted 45o to the spin axis [Barry and Snare, 1966].

The astronauts of Apollos 12, 15 and 16 placed magnetometers on the surface of the moon. The sensor configuration was rather unique. A single axis sensor was mounted at the end of an orthogonal set of arms 1 meter in length as shown in Figure 12. The three sensors measured the vector magnetic field on the Lunar surface. An articulating mechanism could reverse the pointing of each sensor by 180o. This feature enabled one to calculate the sensor zero levels. The mechanism could also place each sensor parallel to each X, Y, Z coordinate. This feature provided data for the calculation of gradients between the sensors in all three directions [Dyal et al., 1970]. The Lunar Surface Magnetometer for Apollo 16 used the new ring core sensors developed by Naval Ordnance Laboratory, White Oak [Dyal and Gordon, 1973].

Figure 12. Lunar surface magnetometer with a single axis fluxgate at the end of each 1 meter arm [Dyal and Parkin, 1971].

The Pioneer Venus spacecraft had three sensors. One sensor was parallel to the spacecraft spin axis one was normal to the spin and the third was at a point 2/3 of the length of the 5m boom and tilted at a 45o angle in the radial direction. At low frequencies in quiet fields, calculation of the vector field from the sine wave amplitude and phase and the steady component parallel to the spin axis could be provided by either the inboard tilted sensor on the two outboard sensors. The difference between these two yielded a measure of the spacecraft field. At low spacecraft telemetry rates only the two outboard sensors were sampled and at lowest data rates when the sampling rate was near or below the spacecraft spin period, the data were despun on board using a Walsh transform. At high data rates the measurements from all these sensors were combined to provide instantaneous vector measurements of the field and were despun on the ground. This configuration also provided some redundancy if a sensor should fail [Russell et al., 1980].

The most accurate vector measurements made in space to date were those of the MAGSAT spacecraft [Acuna et al., 1978]. The project provided data for improved modeling of the time varying magnetic field generated within the core of the earth and to map variations in the strength and vector characteristics of the crustal magnetization [Langel et al., 1982]. The fluxgate sensor was designed and calibrated with great care in order to provide accurate vector data. A Cesium vapor scalar magnetometer provided absolute magnetic field data for calibrating the fluxgate in space. An elaborate system of attitude measurement devices provided spacecraft and magnetic sensor attitude data.

Another project with objectives similar to that of MAGSAT is the Danish spacecraft OERSTED to be launched soon. It will carry the compact, spherical three axis fluxgate magnetometer which was previously described [ Primdahl and Jensen, 1982]. The fluxgate magnetometer is mounted on an optical bench that also carries a nonmagnetic star imager camera. The camera provides data for determining the magnetometer attitude in inertial space. For absolute field measurements the spacecraft has an Overhauser, continuous wave, proton magnetometer [Duret et al., 1995 and Primdahl, 1997].

Offset Determinations

Accurate measurement of zero levels or offsets, as seen by vector magnetometers has long been a problem. The sensors and sensor electronics can have finite offsets and variations in these offsets with time and temperature. The spacecraft can also have fields that add to the offsets as measured by the magnetometer. Depending on their source these spacecraft fields may be static or also vary with time.

On spinning spacecraft the sensors mounted normal to the spacecraft spin axis are easy to evaluate by averaging the data over several spin periods. This yields the sum of the sensor offset and the rotating spacecraft field. This technique could be extended to the third axis by mechanisms that rotate the sensors such that the non spinning sensor is placed in the spin plane for offset calibration. This technique, however, does not help with determining fields that do not rotate with the sensor. The flipping mechanisms were initially driven by thermal devices such as wax pellet actuators and later by bi-metallic springs. Another technique called electronic flipping is used. Switches reverse the polarity of the sensor in such a manner that the offsets can be measured [Behannon et al., 1977].

Ness et al. [1971] proposed using two triaxial magnetometers on one boom to determine the spacecraft field by measuring its gradient. This was followed by Neubauer [1975] studying the use of as many as four magnetometers on one boon and determining the capabilities and errors of such a system. Spacecraft fields are complex, rarely exhibiting a simple dipole structure that most of the calculations rely on. Moreover, the zero levels of sensors are important contributors to the differences between sensors. Therefore, defining the static spacecraft magnetic field by two or more magnetometer is seldom feasible. However, a second magnetometer closer to the spacecraft is useful in identifying and calibrating spacecraft dynamic magnetic fields and provides redundancy. Today nearly all spacecraft of any size carry two magnetometers. Often the outboard magnetometer is for low field measurements and the inboard magnetometer for higher fields. On some deep space missions the magnetometers have been mixed types such as the fluxgates and Helium sensors of Ulysses and Cassini [Balogh et al., 1992; Southwood et al., 1992].

Scales and Ranging

The problem of resolving small changes in the presence of large fields is a common problem for magnetometers. However, the dynamic range of the telemetry system often constrained early measurements. The first spacecraft used frequency modulation (FM) telemetry. The voltage to be measured was passed to a voltage controlled oscillator (VCO). This circuit transformed voltage to frequency. Multiple VCOs each with its unique frequency band were mixed together. This composite spectrum was then sent to earth by a high frequency radio carrier. At the earth the frequencies were separated by bandpass filters and discriminated to reconstruct the original signals. The accuracy and resolution of such a system approached 1% if all elements of the system were calibrated and functioning properly. In practice FM telemetry systems often degraded to 3% measurements.

The introduction of the analog-to-digital converter (ADC) and pulse code modulation somewhat improved telemetry resolution. The first ADC's had only 6 to 8 bits with resolution not much better than the FM telemetry. Presently 16 bit ADC's are routine with advanced ADC's of 20 to 24 bit ADC's in the development laboratory.

Before advent of modern high precision analog to digital converters, field offset systems were developed to extend the dynamic range of magnetometers and maintain their resolution. Such systems are needed around magnetized planets where the inverse cube dependence of the magnetic field on radial distance causes the field strength to vary by several orders of magnitude. In many senses this technique is simply creating an instrument with ADC. One problem with this technique is that the offset fields add noise and must be very well calibrated as they add step function noise when they switch. The system that was used on Sputnik 3 is shown in Figure 13. The magnetometer had a dynamic range of " 2400 nT. This was telemetered to the ground by two separate VCO's, one for the positive voltage and another for the negative voltage [Dolginov et al., 1960]. The offset system passed current from a relay controlled resistor network through a winding on the sensor. The offset incremented in steps of 3000 nT and had a total range of 64,000 nT.

Figure 13.Diagram of field offsetting circuits for Sputnik 3 magnetometer [Dolginov, 1959].

A similar system using solid state components was used on a number of spacecraft including ATS-1, ATS-6 and OGO-5. The system is shown in Figure 14. When the voltage from the basic magnetometer exceeds its dynamic range of "10 volts the level detector causes the counter to increment up or down and apply steps of offset current to a winding on the sensor. This system is actually a digital-to-analog circuit built with discrete parts.

Figure 14.Solid state field offset systems used for ATS-1, OGO-5 and ATS-6 [McPherron et al., 1975].

Another technique that has been widely used is to automatically switch the magnetometer dynamic range when the measured field exceeds the present range. This technique varies the absolute resolution of the magnetometer but maintains the relative or percent resolution. This technique has been used on a great number of spacecraft and is presently planned for the Cassini magnetometers. The switching ranges for Voyager 1 and 2 are shown in Figure 15. The magnetometer automatically selects the proper range to keep the signal in the center of the magnetometer dynamic range. If the average field remains in the area of the guard bands for several seconds the magnetometer will switch up or down as required. The range switching can be overridden by ground command [Behannon et al., 1977].

Figure 15.Illustration of a portion of the total magnetometer range switching strategy for the Voyager 1 and 2 magnetometers [Behannon et al., 1977].

Figure 16 depicts the dynamic range and the resolution, i.e. least-significant-bit (LSB), of the telemetered data for several spacecraft. Voyager 1 and 2 had magnetometers with 12 bit ADCs. The main low field magnetometer ranges and resolution are depicted in the figure for each of the 8 gains available for the magnetometer.

Figure 16.Dynamic Range and Resolution for Several Spacecraft.

Both the ISEE 1 and 2 and the Galileo magnetometer used 12 bit ADCs that were specially manufactured with each step trimmed to an accuracy of better than 3 the value of the LSB. The data was sampled at a high rate and averaged to produce output data that had an accuracy of 15 bits. The magnetometer for GGS-Polar used an ADC with full 16 bit resolution. With the use of 16 bit ADC's and modern techniques the need for multiple range switching is reduced. One can see the effect that 15 and 16 bit ADC's has on extending the dynamic range without sacrificing the high resolution of these magnetometers.

Data Processing

The telemetry bandwidth from spacecraft is generally determined by the power available on the spacecraft, because scientists can make good use of every available bit and tend to request the greatest possible bandwidth. The available telemetry is divided between various scientific instruments and spacecraft engineering systems. The telemetry allotted to magnetometers is often insufficient to transmit all of the information desired. This has resulted in a variety of data processing techniques to partially analyze the data on board the spacecraft and telemeter the results. This is particularly important on spinning spacecraft whenever the telemetry rate is insufficient to resolve the spin tone in the spin plane associated with the rotation of the spacecraft in the ambient magnetic field. We recall that the Nyquist criterion states that at least two points must be measured on any sine wave per wave period in order to determine its amplitude and phase. If this is not achieved, then the spin tone will appear to be at an incorrect frequency. This effect is called aliasing because a signal appears to occur at a frequency other than its true frequency, i.e. it assumes an alias. Of course all frequencies not just the D.C. field are affected.

The first such systems were executed with analog circuitry and were simple sum and difference amplifiers as previously mentioned on ATS-1. The Apollo 15 Subsatellite Magnetometer had one fluxgate sensor parallel to the spin axis and one transverse to the spin axis. The parallel magnetometer was sampled directly. The transverse signal was rectified and filtered before sampling and the result represented the transverse field magnitude. The time between the zero crossing of the transverse signal and a sun pulse provided phase information relative to the spacecraft sun line [Coleman et al., 1972].

Explorer 33 was the first spacecraft to carry a magnetometer with a spin demodulator. The two transverse sensor outputs were multiplied by sine and cosine signals derived from the spacecraft sun pulse [Sonett 1966; Sonett et al. 1968]. These functions were implemented with analog circuits.

The Pioneer Venus Orbiter had a 12 sec spin rate and because of the varying distance from earth the spacecraft telemetry bit rate varied from 4096 to 8 bits per second. At the lowest bit rate the magnetometer could only send one vector each 21 or 64 spins depending on which telemetry format was being used. Thus it was necessary to despin the measurements before transmittal. Power and mass for the magnetometer were not enough to fully implement a sine-cosine demodulator. The approach used was to multiply the data by a Walsh transform. The Walsh transform is basically the multiplication of the spinning data by two square waves one in phase and one in quadrature to the spacecraft spin.

The spinning data were digitally sampled and sent to averaging registers. The data were clocked in during the first half spacecraft rotation. During the second half spacecraft rotation the data was inverted. This process is repeated in parallel but with inverted signals from the one-quarter to three-quarter mark of the rotation. This functions as a full wave sun synchronous demodulator and a low pass filter. All of the spin demodulator functions were constructed using hard wired CMOS digital circuits [Russell et al., 1980]. The Pioneer Venus magnetometer also created "floating point" words rather than "fixed point" to make maximum use of the telemetry system.

The ADC and digital processing allowed one to recover greater accuracy than that achievable with analog circuits. The advent of the microprocessor gave the instrument designer a menu of computational functions. Often the data are sampled at a high rate and averaged to create a low pass filter. The digital averaging produces a precise filter algorithm. Recursive filters have been used when computational power is inadequate to handle a large averaging chore. However, recursive filters are asymmetric in time and hence introduce phase shifts.

Another technique is to monitor the data for rapid changes. When a particular shock or similar phenomena is detected, high rate data can be stored in a large solid state memory. The stored high rate data is then returned to Earth slowly at a lower data rate. More sophisticated functions such as the Fast Fourier Transform (FFT) have been programmed to enable the study of the physics of the fine-scale structure of shock waves, directional discontinuities and boundary structures [Lepping et al., 1995; Reidler et al., 1986]. An example of the data processing flow using two 80C86 Processors is that for the GGS-POLAR magnetometer shown in Figure 17. The prime data is sampled at 500 samples per second, filtered with a recursive filter and decimated to 100 samples per second. The data is again filtered and decimated to 10 vectors per second which is the output rate to the spacecraft telemetry. At 10 vectors per second and a spin period of 6 seconds the transverse data did not require despinning [Russell et al., 1995]. However, there is adequate processing capability to despin the data on board and provide for easy quick look interpretation of the data.

Figure 17.Data Processing flow diagram for GGS-Polar magnetometer [Russell et al., 1959].

CONCLUSION

The continued development of booster rockets with the addition of orbit injection stages has enhanced the capability to launch larger heavier scientific spacecraft. Thus the power available for instruments has increased and the enhanced telemetry transmitter power has increased the data recovery rate several orders of magnitude over that of early satellites.

Most modern spacecraft have the capability of deploying long booms such that instruments sensitive to spacecraft emissions can achieve a quieter environment. Magnetometers and plasma wave antennae are usually found on such booms.

The rapid development of radiation hardened, miniature, semiconductors and other electronic devices has enabled the instrument designer to build more sophisticated electronics and data systems. The result of these developments is a richer return of scientific data from spacecraft. Much of this progress has been driven by military spending. With the present reduction in both military and scientific funding one might question the direction that scientific instrumentation will take in the future.

Acknowledgments. I would like to thank Paul Coleman, Margaret Kivelson, Robert McPherron and Christopher Russell for giving me the opportunity to make a career of magnetometery. The constructive comments of the reviewers have contributed by pointing out oversights and misstatements made by the writer. This work was supported by the National Aeronautics and Space Administration contract NAS 5-30373 and Jet Propulsion Laboratory contract 959483.

REFERENCES

Acuna, M. H., C. S. Scearce, J. B. Seek and J. Scheifele, The MAGSAT Magnetometer - A Precision Fluxgate for the Measurement of the Magnetic Field, NASA Technical Memorandum 79656, September 1978.

Acuna, M. H., MAGSAT - Vector Magnetometer Absolute Sensor Alignment Determination, NASA technical Memorandum 79648, September 1981, NASA Goddard Space Flight Center, Greenbelt, Maryland 20771.

Acuna, M. H. and N. F. Ness, The Pioneer XI Hiah Field Fluxgate Magnetometer, Space Sci. Instrum., pp177-188, 1975.

Allen, J. H., Long Term Stability of Self-Oscillating Rubidium Magnetometers, J. Geomag. Geoelect., 20, pp197-204, 1968.

Aschenbrenner, H. and G. Goubau, Eine Anordnung Registrieung rascher magnetischer Storungen, Hochfreg Tech. Elektroakust. 47, pp178-181, 1936.

Balogh, A., T. S. Beek, R. S. Forsyth, P. C. Hedgecock, R. S. Marquedant, E. J. Smith, D. J. Southwood and B. T. Tsurtani, The Magnetic Field Investigation on the Ulysses Mission: Instrumentation and Preliminary Results, Astronomy and Astrophys., 92, pp221-236, 1992.

Barry, J. D. and R. C. Snare, A Fluxgate Magnetometer for the Application Technology Satellite, IEEE Trans. Nucl. Sci., NS-13, No.6, pp326-331, 1966.

Behannon, K. W., M. H. Acuna, L. F. Burlaga, R. P. Lepping, N. F. Ness and F. M. Neubauer, Magnetic Field Experiment for Voyagers 1 and 2, Space Sci. Rev., 21, pp235-257, 1977.

Coleman, Jr., P. S., G. Schubert, C. T. Russell and L. R. Sharp, The Particles and Fields Subsatellite Magnetometer Experiment, Apollo 15 Preliminary Science Report, pp22-1 - 22-9, NASA-SP289, 1972.

Dolginov, S. Sh., L. N. Zhuzgov and V. A.. Selyutin, Magneto-meters in the Third Soviet Earth Satellite, Artificial Earth Satellites, 4, pp358-396, 1960.

Dolginov, S. Sh., E. G. Yeroshenko, L. N. Zhugov and N.V. Pushkov, Investigation of the Magnetic Field of the Moon, Geomagnetism and Aeronomy, 1, pp21-29, 1960.

Duret, D., J. Bonzom, M. Brochier, M. Frances, J. M. Leger, R. Odru, C. Salvi, and T. Thomas, Overhauser Magnetometer for the Danish Oersted Satellite, IEEE Trans. Magnetics, Vol. 31, No.6, pp3197-3199, November 1995.

Dyal, P., C. W. Parkin and C. P. Sonett, Lunar Surface Magnetometer, IEEE Trans. Geosci. Electronics, GE-8, pp203-215, 1970.

Dyal, P. and C. W. Parkin, The Magnetism of the Moon, Scientific American, 224, No. 2, pp62-73, Aug. 1971.

Dyal, P. and D. I. Gordon, Lunar Surface Magnetometers, IEEE Trans. Magn., MAG-9, pp226-231, 1973.

Fredericks, R. W., E. W. Greenstedt and C. P. Sonett, Magneto-dynamically Induced Ambiguity in the Data from Tilted Spinning Fluxgate Magnetometers: Possible Application to IMP-1, J. Geophys. Res., 72, pp367-382, 1967.

Geyger, W. A., The Ring Core Magnetometer - A New Type of Second-Harmonic Flux-Gate, AIEE Trans. Communications and Electronics, 81, pp65-73, 1962.

Gordon, D. I., R. H. Lundsten, R. A. Chiarado and H. H. Helms Jr., A Fluxgate Sensor of High Stability for Low Field Magnetometry, IEEE Trans. Magnetics, MAG-4, pp397-401, 1968.

Heppner, J. P., J. D. Stolarik and L. H. Meredith, The Earth's Magnetic Field above WSPG, New Mexico, From Rockef Measurements, J. Geophys. Res., 63, pp277-289, 1958.

Heppner, J. P., N. F. Ness, C. S. Scearce and T. L. Skillman, Explorer 10 Magnetic Field Measurements, J. Geophys. Res., 68, pp1-46, 1963.

Heppner, J. P., The World Magnetic Survey, Space Sci. Rev., 2, pp315-354, 1963.

Judge, D., L. McLoud and A. R. Sims, The Pioneer 1, Explorer VI and Pioneer V High-Sensivity Transistorized Search Coil Magnetometer, IEEE Trans. on Space Electronics and Telemetry, SET6, pp114-121, 1960.

Kellock, S., P. Austin, A. Balough, B. Gerlach, R. Marquedant, G. Musmann, E. Smith, D. Southwood and S. Szalai, Cassini Dual Magnetometer Instrument, Proceedings: The International Society for Optical Engineering, Vol. 2803, pp146-152, August 1996.

Keyser, A. R., J. A. Rue and L. D. Schearer, A Metastable Helium Magnetometer for Observing Small Geomagnetic Fluctuations, J. Geophys. Res., 66, pp4163-4169, 1961.

Langel, R., G. Ousley, J. Berbert, J. Murphy and M. Settle, The MAGSAT Mission, Geophys. Res. Ltr., Vol. 9, No.l4, pp243-245, April 1982.

Lepping, R. P., M. H. Acuna, L. F. Burlaga, W. M. Farrell, J. A. Slavin, K. H. Schatten, F. Mariani, N. F. Ness, F. M. Nebauer, Y. C. Whang, J. B. Byrnes, R. S. Kennon, P. V. Panetta, J. S. Scheifele and E. M. Warley, The Wind Magnetic Field Investigation, Space Sci. Rev., 71, pp207-229, 1995.

Maple, E., W. A. Bowen and S. F. Singer, Measurement of the Earths Magnetic Field at High Altitudes at Wite Sands, New Mexico, J. Geophys. Res., 55, pp115-126, 1950.

McPherron, R. L., P. J. Coleman, Jr. and R. C. Snare, ATS-6 UCLA Fluxgate Magnetometer, IEEE Trans. Aerospace and Electronic Sys., AES-11, pp1110-1116, 1975.

Ness, N. F., C. S. Scearce and S. Cantarano, Preliminary Results from the Pioneer 6 Magnetic Fields Experiment, J. Geophys. Res., 71, pp3305-3313, 1966.

Ness, N. F., Magnetometers for Space Research, Space Sci. Rev., 11, pp459-554, 1970.

Ness, N. F., K. W. Behannon, R. P. Lepping and K. H. Schatten, Use of two Magnetometers for Magnetic Field Measurements on a Spacecraft, J. Geophys. Res., 76, pp3565-3573, 1971.

Neubauer, F. M., Optimization of Multimagnetometer Systems on a Spacecraft, J. Geophys. Rev., 80, pp3235-3240, 1975.

Primdahl, F., A Ferrite Core Fluxgate Magnetometers, Danish Meterological Institute Geophysical Papers, R-12, pp1-66, 1970.

Primdahl, F., The Fluxgate Magnetometer, J. of Phys. E: Scientific Instruments, 12, pp241-253, 1979.

Primdahl, F. and P. Anker Jensen, Compact Spherical Coil for the Fluxgate Magnetometer, J. Phys. E: Sci. Instrum., Vol. 15, pp221-226, 1982.

Primdahl, F., H. Luhr and E. K. Lauridsen, The Effect of Large Uncompensated Transverse Fields on the Fluxgate Magnetic Sensor Output, Danish Space Research Institute Report 1-92, 1992.

Primdahl, F., Scalar Magnetometers for Space Applications, This Monograph, 1997.

Riedler, W., K. Schwingenschuh, Y. G. Yeroshenko, V. A. Stayshkin and C. T. Russell, Magnetic Field Observations in Comet Halleys Coma, Nature, 321, pp288-289, 1986.

Ruddock, K. A., Optically Pumped Rubidium Vapor Magneto-meters for Space Research, Space Res., II, pp692-700, 1961.

Rumbaugh, L. H. and L. R. Alldredge, Airborne Equipment for Geomagnetic Measurements, Trans. Amer. Geophys. Union, 30, pp836-848, 1949.

Russell, C. T., R. C. Snare, J. D. Means, D. Pierce, D. Dearborn, M. Larson, G. Barr and G. Le, The GGS/POLAR Magnetic Fields Investigation, Space Sci. Rev., 71, pp.563-582, 1995.

Russell, C. T., R. C. Snare, J. D. Means and R. C. Elphic, Pioneer Venus Orbiter Fluxgate Magnetometer, IEEE Trans. on Geosci. and Remote Sensing, GE18, pp32-35, 1980.

Schonstedt, E. O. and H. R. Irons, Airborne Magnetometers for Determining all Magnetic Components, Trans. AGU. 34, pp363-378, 1953.

Shapiro, I. R., J. D. Stalarik and J. P. Heppner, The Vector Field Proton Magnetometer for IGY Satellite Ground Stations, J. Geophys. Res., 65, pp913-920, 1960.

Slocum, R. E. and F. N. Reilly, Low Field Helium Magnetometer for Space Applications, IEEE Trans. on Nuclear Sci., NS-10, pp165-171, 1963.

Smith, E. J., B. V. Connov and G. J. Foster, Jr., Measuring the Magnetic Fields of Jupiter, IEEE Trans. Magnetics, MAG-11, pp962-980, 1975.

Snare, R. C. and C. R. Benjamin, A Magnetic Field Instrument for the OGO-E spacecraft, IEEE Trans. Nucl. Sci., NS-13, No.6, pp333-340, 1966.

Sonett, C. P., Modulation and Sampling of Hydrodynamic Radiation, Space Research, 6, pp280-322, 1966.

Sonett, C. P., P. S. Colburn, R. G. Currie and J. D. Mihalov, The Geomagnetic Tail: Topology, Reconnection and Interaction with the Moon, Physics. of the Magnetosphere, D. Reidel Publishing Co., Dordrecht-Holland, pp461-484, 1968.

Southwood P. J., A Balogh and E. J. Smith, Dual Technique Magnetometer Experiment for the Cassini Orbiter Spacecraft, J. British Interplanetary Soc., 45, pp371-374, 1992.

______________________

Robert C. Snare, Institute of Geophysics and Planetary Physics, Unisversity of California, Los Angeles, CA 90095-1567.