Dual Magnetometer Space

REVIEW OF SCIENTIFIC INSTRUMENTS VOLUME 73, NUMBER 11 NOVEMBER 2002
REVIEW ARTICLE
Space-based magnetometers
Mario H. Acuñaa)
Laboratory for Extraterrestrial Physics, NASA Goddard Space Flight Center, Code 695, Greenbelt,
Maryland 20771
共Received 16 October 2001; accepted 3 June 2002兲
The general characteristics and system level concepts for space-based magnetometers are presented
to illustrate the instruments, principles, and tools involved in making accurate magnetic field
measurements in space. Special consideration is given to the most important practical problems that
need to be solved to ensure the accuracy of the measurements and their overall impact on system
design and mission costs. Several types of instruments used to measure magnetic fields aboard
spacecraft and their capabilities and limitations are described according to whether they measure
scalar or vector fields. The very large dynamic range associated with magnetic fields of natural
origin generally dictates the use of optimized designs for each particular space mission although
some wide-range, multimission magnetometers have been developed and used. Earth-field magnetic
mapping missions are the most demanding in terms of absolute accuracy and resolution,
approaching ⬍1 part in 100 000 in magnitude and a few arcsec in direction. The difficulties of
performing sensitive measurements aboard spacecraft, which may not be magnetically clean,
represent a fundamental problem which must be addressed immediately at the planning stages of
any space mission that includes these measurements. The use of long, deployable booms to separate
the sensors from the sources of magnetic contamination, and their impact on system design are
discussed. The dual magnetometer technique, which allows the separation of fields of external and
spacecraft origin, represents an important space magnetometry tool which can result in significant
savings in complex contemporary spacecraft built with minimum magnetic constraints. Techniques
for in-flight estimation of magnetometer biases and sensor alignment are discussed briefly, and
highlight some basic considerations within the scope and complexity of magnetic field data
processing and reduction. The emerging field of space weather is also discussed, including the
essential role that space-based magnetic field measurements play in this complex science, which is
just in its infancy. Finally, some considerations for the future of space-based magnetometers are
presented. Miniature, mass produced sensors based on magnetoresistance effects and
micromachined structures have made significant advances in sensitivity but have yet to reach the
performance level required for accurate space measurements. The miniaturization of spacecraft and
instruments to reduce launch costs usually results in significantly increased magnetic contamination
problems and degraded instrument performance parameters, a challenge that has yet to be solved
satisfactorily for ‘‘world-class’’ science missions. The rapidly disappearing manufacturing
capabilities for high-grade, low noise, soft magnetic materials of the Permalloy family is a cause of
concern for the development of high performance fluxgate magnetometers for future space missions.
关DOI: 10.1063/1.1510570兴
I. INTRODUCTION Marco Polo is credited with taking this information to Eu-

rope in the 13th century. Some 300 years later, Sir William
Magnetometry is the science of accurate measurements Gilbert, in ‘‘De Magnete,’’ was the first to describe the Earth
of the strength and direction of magnetic fields and space as a ‘‘giant magnet’’ and Gauss developed the mathematical
magnetometry represents the specialized science of making tools for its analytical representation. We now know that
magnetic field measurements in space, aboard vehicles such magnetic fields are omnipresent in nature and just about ev-
as rockets, balloons, and spacecraft. The Chinese are re- ery major body in the solar system has magnetic fields asso-
ported to have been the first to use the Earth’s magnetic field ciated with it, either of interior origin, induced, or remanent.
for navigation purposes. More than 4500 years ago they used The interplanetary magnetic field is the extension into space
splinters of magnetite as compasses for direction finding, and of the Sun’s magnetic field carried outwards by the super-
sonic flow of the solar wind. Sources of magnetic fields in-
a兲
Electronic mail: mario.acuna.1@gsfc.nasa.gov clude electrical currents circulating in fluids, conductors, and
0034-6748/2002/73(11)/3717/20/$19.00 3717
Downloaded 20 Oct 2005 to 128.183.22.32. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
3718 Rev. Sci. Instrum., Vol. 73, No. 11, November 2002 Mario H. Acuña
ionized gases, ‘‘permanent’’ magnets like those that occur the martian field is organized in linear structures and is
naturally 共magnetite or lodestone兲 and those made industri- closely correlated with the age of the crustal terrain. The
ally of materials like iron, iron oxides, cobalt, nickel alloys, Lunar Prospector 共LP兲 mission23–25 provided a new and de-
and rare earth compounds. tailed view of the lunar, crustal magnetic fields discovered by
The beginnings of space magnetometry trace back to the Apollo missions.26 –31 These are correlated in a puzzling
measurements of the Earth’s magnetic field with instruments way to the antipodes of the great impacts that created the
flown on balloons and rockets in the early 1950s. The mea- lunar basins. The MGS and LP magnetic field observations
surement techniques used in these experiments were largely are revolutionizing our thinking about the early history and
adapted from instruments developed around the time of thermal evolution of Mars and the Moon. They are certain to
World War II, when magnetometers were widely used in geo- result in significant revisions of our views of Earth’s mag-
physical applications, navigation, salvage, and mine and an- netic anomalies and their origin. The role played by giant
tisubmarine warfare. The early space probes made significant impacts during the formation and early evolution of the solar
discoveries: the Earth’s magnetosphere, the comet-tail-like system has now achieved a new level of relevance in our
geometry of the antisunward Earth’s magnetic field, the in- limited understanding of the processes that took place more
terplanetary magnetic field 共IMF兲, its sector structure, and than four billion years ago.
the many boundaries associated with the interaction of the Planetary space missions such as Pioneer 10 and 11, Pio-
Earth’s magnetic field with the solar wind. Dolginov neer Venus, the Mariner series, Voyager, Helios, Ulysses,
et al.,1–3 Sonnet,4 Heppner et al.,5 Cahill,6 and Ness7,8 were Giotto, Mars Global Surveyor, Galileo, and NEAR have car-
among the first investigators to equip rockets and satellites ried out magnetic field measurements in the immediate vi-
with magnetometers and carry out measurements in the cinity of most of the planets in the solar system, as well as
Earth’s ionosphere and magnetosphere in the interplanetary around comets and asteroids.32– 41 The NEAR–Shoemaker
medium and around the Moon. spacecraft was placed in orbit around the asteroid 433 Eros
Why is magnetometry so important in space explora- on 14 February 2000 and spent one year making magnetic
tion? Magnetic field measurements are essential to organize field measurements from 35 and 50 km orbits. On 12 Febru-
and understand energetic charged particle and plasma mea- ary 2001 the spacecraft landed on 433 Eros and continued to
surements and to derive fundamental information about the make magnetic field observations from the surface for sev-
environment surrounding different bodies in the solar sys- eral days.33,42– 44 Contrary to expectations based on measure-
tem. Charged particles move easily along magnetic field ments and analyses carried out at asteorids Braille and
lines because their transverse motion is opposed by the Lo- Gaspra by Deep Space-1 and Galileo,36,38 no magnetic field
renz force. Knowing the geometry of the field is equivalent of asteroidal origin was detected, making 433 Eros a remark-
to having a map of the pathways followed by the particles’ ably nonmagnetic, undifferentiated primitive object.42 The
guiding center. Magnetic field measurements also represent weak IMF has been measured by spacecraft such as the In-
one of the few remote sensing tools used by spacecraft 共grav- terplanetary Monitoring Platform 共IMP兲 series, Pioneers,
ity being another兲 that provide information about the deep Voyager, Mariner, WIND, ACE, GEOTAIL, and many
interior of a planet rather than just its surface and/or atmo- others.8,40,45,46 A recent review of the history of vector mag-
sphere which are customarily studied with multispectral im- netometry in space was given by Snare.47 Magnetic field
aging instruments. Planetary magnetic fields like those of the measurements are used as well in many Earth-orbiting space-
Earth, Jupiter, and Saturn are generated by currents circulat- craft for engineering applications. These include attitude de-
ing in their liquid metallic cores or perhaps the core–mantle termination and control, spacecraft momentum management,
interface. The primary energy source powering these dyna- and scientific instruments pointing.48,49 The Earth’s magnetic
mos is thermal convection of electrically conducting fluids field provides a convenient ‘‘natural’’ frame of reference,
driven by the internal heat flux and organized by planetary which can be modeled with good accuracy, and modern sys-
rotation. The outermost giant planets Uranus and Neptune tems and spacecraft utilize it to establish their orientation and
are not assumed to have formed metallic cores and their activate control systems. Where absolute angular measure-
magnetic fields are generated closer to the surface, where ment accuracy of the order of 1°–2° degrees is acceptable,
electrical currents can flow in high-conductivity crustal magnetic systems provide significant cost, simplicity, and re-
‘‘oceans.’’ 9–12 In the case of the terrestrial planets, Venus liability advantages over inertial sensor-based systems. A
does not possess an intrinsic magnetic field13 while Mercury common space-based application of magnetometers aboard
is believed to be currently magnetized by the remains of an spacecraft in low and geostationary Earth orbit is the control
ancient dynamo which is decaying over time.14 The study of of electromagnets which, when energized, apply torque to
the Hermean field is one of the primary objectives of the the spacecraft by interacting with the geomagnetic field.
MESSENGER and Beppi Colombo missions that will be These ‘‘magnetic torquers’’ are generally used to ‘‘de-
launched in the 2003–2005 time frame.15,16 The existence of saturate’’ momentum wheels 共reduce the accumulated angu-
an intrinsic magnetic field at Mars was a topic of consider- lar momentum resulting from the operation of reaction
able discussion.17–19 Recent measurements by the Mars Glo- wheels aboard a spacecraft兲 or to orient the spacecraft along
bal Surveyor 共MGS兲 Mission not only demonstrated that a desired direction.48 –50
Mars does not currently possess an internal field, but made Numerous discoveries have been made studying the
the remarkable discovery of the existence of strong crustal sources, geometry, and temporal behavior of solar system
magnetic fields of paleomagnetic origin.20–22 In many areas magnetic fields but there is still much to be learned. In recent
Rev. Sci. Instrum., Vol. 73, No. 11, November 2002 Space-based magnetometers 3719
years the new field of ‘‘space weather’’ has evolved from the individual sensing technologies. The reason for this approach
extended, multipoint observations of the Sun–Earth environ- is that the latter are reasonably well developed and have been
ment that have been carried by international collaborative extensively reviewed in the literature.8,49,69 On the other
programs such as the International Solar-Terrestrial Physics hand, system level trade-offs, which determine the ultimate
共ISTP兲 Program since 1992 that utilize multiple spacecraft success or failure of magnetic field measurements on a given
and ground-based systems.51,52 Solar transient events such as space mission, have not received equal attention. Spacecraft,
coronal mass ejections 共CMEs兲, solar flares, and magnetic components, and instrument design philosophies have
clouds can affect operation and even cause failure in Earth- changed significantly in the last 20 years, particularly with
orbiting spacecraft and power distribution systems on the respect to the use of ‘‘heritage’’ or ‘‘off-the-shelf’’ sub-
ground such as the North American power grid.53–58 The systems to reduce development costs. To ensure the accuracy
reliable operation of contemporary telecommunications, glo- of magnetic field measurements, space vehicles have to be
bal positioning, and strategic satellites today depends to a designed to minimize the generation of undesirable static and
significant extent on near-real time knowledge of space dynamic fields, regardless of the type of sensor used, and
weather conditions. The prevention of power grid failures these requirements are seldom met by low cost, readily avail-
and associated large area blackouts in North America and able components. The magnetic signature reduction process
northern Europe which are caused by ground currents in affects all aspects of spacecraft, instruments, and mission
transmission lines induced by the time-varying geomagnetic design and can have a dramatic impact on the cost and
field and can destroy distribution transformers, is also depen- schedule and even the viability of sensitive magnetic field
dent upon timely access to space weather information.59– 61 measurements, if not addressed early and correctly in the
Space magnetometry is now an established and mature sci- development cycle.
ence which plays a critical role in the activities described
above and is expected to remain so for the foreseeable future.
In the US, the National Space Weather Program, NASA’s II. MEASURING MAGNETIC FORCES
Living with a Star 共LWS兲 program and the Sun–Earth Con- The magnetic force between two magnetic poles is given
nection research efforts, including NASA’s Solar Terrestrial by an expression identical to that of the gravitational force
Probes, all rely on space magnetometry as one of the key between two masses,
measurements to be carried out. Similar programs in Europe,
Russia, and Japan reflect the essential importance of mag- F⫽ 共 m 1 m 2 / ␮ r 2 兲 r, 共1兲
netic field measurements in space.62– 66 where F is the force in dynes 关centimeter-gram-second 共cgs兲
The magnetic field measurement techniques and instru- system兴 between the poles that are separated by r cm 共r is the
ments described in this review date from the late 1930s until unit vector directed between m 1 and m 2 ). The permeability
the present time. Important sensor developments took place of the medium is denoted by ␮. Since the lines of force are
in the 1936 –1955 time frame such as the invention of the closed, magnetic poles cannot exist by themselves and al-
fluxgate67 and the application of nuclear magnetic resonance ways exist in pairs (ⵜ•B⫽0), while the force between them
to magnetometry. The traditional definition of space-based can be repulsive or attractive depending on the polarity of
‘‘magnetometry’’ does not include the sensing of time- the poles involved.
variable magnetic fields by induction sensors such as search The magnetic quantities derived are the magnetic field
coils, which will not be discussed in this review. Contempo- strength H, which is defined as the force F per unit pole and
rary advances in wide-dynamic-range magnetic field sensing is measured in oersteds 共cgs兲, and the magnetic field induc-
technology have focused on improvements in sensor noise tion, which is the field B induced by the excitation H in a
performance, analog and digital signal processing tech- medium of permeability ␮,
niques, as well as instrument miniaturization. Modern, high
B⫽ ␮ H. 共2兲
volume magnetic field sensing technologies, e.g., anisotropic
magnetoresistance 共AMR兲, giant magnetoresistance 共GMR兲, In the cgs system of units, B is measured in gauss. In air
and micromachined devices 共MEMS兲, have yet to compete or vacuum, the permeability ␮ is unity so for all practical
effectively with established sensors in space applications. applications in space oersted and gauss can be used inter-
Devices based on magnetoresistance effects 共AMR and changeably. In the Systeme International 共SI兲 of units, the
GMR兲, although they offer adequate sensitivity for Earth- magnetic induction is measured in tesla (104 G), while the
field sensing, suffer from problems like hysteresis and poor magnetic field strength is measured in ampere/turns/cm. The
stability that are introduced by the use of ‘‘flux concentra- permeability of air 共vacuum兲 in the SI system is ␮ ⫽4 ␲
tors’’ or biasing arrangements to enhance the output signal or ⫻10⫺7 . To quantify weak magnetic fields the nanotesla
to linearize the transfer function.49,68 Magnetometers using (1 nT⫽10⫺9 T) and the older, equivalent cgs unit the
superconducting quantum interference devices 共SQUIDs兲 gamma (1 ␥ ⫽10⫺5 G⫽1 nT) are frequently used inter-
can exhibit significantly increased sensitivity over those de- changeably. The Earth’s magnetic field at the surface and
scribed here, but the complex logistics of cryogens and as- near the equator has an approximate strength of 31 000 nT or
sociated systems have prevented their application to space- 0.31 G, whereas Jupiter’s north magnetic pole field strength
based measurements.49 is more than 14 G or 1 400 000 nT. Solar magnetic fields
The primary emphasis of this review is on considerations associated with coronal loops and prominences can reach
at the space systems level rather than detailed discussion of values as high as several thousand Gauss. At the opposite
end of the dynamic range, the IMF at the Earth’s orbit 关1 where B is the magnitude of the external field and the pro-
astronomical unit 共AU兲兴 is typically of the order of 5–10 nT, portionality constant ␶ p is the gyromagnetic ratio of the pro-
while at the orbit of Uranus and Neptune it may be as low as ton. This is the ratio of the proton’s magnetic moment to its
0.05 nT or, equivalently, 5⫻10⫺11 G. This very large dy- spin angular momentum. The value of (2 ␲ / ␶ p ) is known
namic range of magnetic field intensities presents unique very accurately from quantum mechanical principles,
challenges to the design of magnetic field instrumentation for 23.4874 共nT/Hz兲, and the proton precession magnetometer is
space missions to the outer planets. the primary standard used in the calibration of other magne-
tometers and coil systems.
The basic proton precession magnetometer is therefore
III. MAGNETIC FIELD MEASURING INSTRUMENTS composed of a liquid sample rich in protons 共water, naphta,
kerosene, etc.兲, a polarizing/sensing coil that surrounds the
As already introduced in the above paragraphs, the in- sample, an ac amplifier, and a duty cycled current source,
struments used to measure the strength 共and direction兲 of the which are alternatively connected to the coil. After being
magnetic field are called magnetometers. Since the magnetic used to polarize the sample the coil is connected to the am-
field is a vector quantity that has both magnitude and direc- plifier to sense the Larmor signal, which lasts only a few
tion, we usually differentiate between two generic classes of seconds and whose frequency is then measured with an or-
instruments: dinary computing counter. The polarize/count cycle of con-
ventional proton precession magnetometer designs is typi-
共a兲 scalar magnetometers, which measure only the total
cally 1 s or more. The liquid sample volume is relatively
strength or magnitude of the ambient magnetic field
large and massive, particularly when the polarizing coil mass
regardless of its orientation, and
is considered, and liquids that can operate over a wide tem-
共b兲 vector magnetometers, which produce an output pro-
perature range are required. The power required to generate
portional to the strength and direction of the magnetic
the 100 G or more polarizing field is appreciable, and useful
field, referenced to a principal axis in the sensing ele-
signals can only be obtained for ambient fields larger than
ment. The polarity or sign of the output in general de-
approximately 20 000 nT and only if the local spatial field
pends on the direction of the ambient field with respect
gradient is small. These limitations have restricted the use of
to the magnetometer sensing axes.
traditional proton precession instruments for space measure-
Both classes of instruments have been used for space ments to a few, very specialized applications such as sound-
measurements but vector magnetometers are far more coming rockets and balloons 共see Ref. 49, and references
mon due to their capability of providing directional informa- therein兲.
tion. This is essential for understanding of the physical phe- Recent improvements to this basic sensing technique,
nomena being studied or for the intended application in like the Overhauser effect proton precession magnetometer,
attitude determination and control. We will discuss below use an indirect technique to ‘‘polarize’’ the sample much
some of the principal characteristics of these instruments, more efficiently and can generate a continuous Larmor pre-
their advantages and limitations, and the challenging prob- cession signal.65,66,70 The proton-rich liquid sample is doped
lem of performing sensitive magnetic field measurements with a free radical as a source of electrons which are pumped
aboard space platforms such as spacecraft, planetary probes, with rf energy at ⬃20– 60 MHz. The electrons are effectively
rockets, and balloons. For a comprehensive review of early coupled to the protons in the sample, which are then polar-
space research magnetometers and techniques, see the review ized dynamically by the rf excitation. This method of polar-
by Ness,8 and references therein. ization is much more efficient than the dc method 共a few
watts versus tens of watts兲 and yields Larmor signal ampli-
A. Scalar magnetometers tudes which are 100 times larger, allowing continuous mea-
The most common scalar magnetic field measuring in- surements to be made with sensitivity of ⬃ 0.02 nT.
strument in general use is the proton precession magnetome- Optically pumped magnetometers are another class of
ter, which is based on the phenomenon of nuclear magnetic magnetic field measuring instruments which have found ap-
resonance. At the end of World War II it was discovered that plication in space measurements, both as scalar as well as
many atoms possess a net magnetic moment and behave as vector instruments.41,71–74 In the scalar configuration, they
small magnets. A sample of a liquid rich in protons 共hydro- are capable of measuring magnetic fields over a wider range
gen nuclei兲 surrounded by a coil is magnetically polarized to than the proton precession instruments and with much higher
align all of its magnetic moments in a given direction. The time resolution. These instruments use the energy required to
sample is then allowed to ‘‘relax’’ in the presence of an ex- transfer atomic electrons from one energy level to another as
ternal magnetic field. The protons that have been aligned the mechanism for magnetic field detection. A cell containing
parallel to the polarizing field will precess like spinning tops a suitable gas is irradiated with light from a discharge lamp
around the ambient magnetic field and induce an ac signal in at the proper frequency to excite the atoms to a quantum
the polarizing coil whose frequency is proportional to the level, which becomes overpopulated as a consequence of for-
magnitude of the field. This frequency is called the Larmor bidden transitions in the system. Under these conditions, the
frequency and is given by gas cell becomes transparent to the irradiating beam 共the pro-
cess used to achieve this is called optical pumping兲. If the
f 共 Hz兲 ⫽ 共 ␶ p B/2␲ 兲 , 共3兲 cell is then subjected to a rf signal with energy appropriate to
FIG. 1. Schematic for an optically pumped magnetometer 共after Ref. 8兲: 共a兲
basic configuration and 共b兲 self-oscillating loop.
cause depopulation of this energy level, it will become

opaque again and, block the transmission of light at the dis-
charge lamp frequency. The most commonly used elements
for optically pumped magnetometers are helium and alkali
metals like cesium, rubidium, and sodium. Helium and its
isotopes in particular are in wide use in high accuracy, high
time resolution scalar magnetic field measurements for mili-
tary applications. Modified versions of these ‘‘metastable he-
lium’’ magnetometers were developed to perform wide dy-
namic range scalar and vector measurements aboard
spacecraft and are described briefly below in Sec.
III B.41,64,71,75 FIG. 2. Two-cell optically pumped scalar magnetometer sensor 共after Ref.
A generic optically pumped magnetometer is illustrated 8兲: 共a兲 schematic of the arrangement and 共b兲 physical arrangement of the
sensors required to avoid ‘‘null’’ zones.
in Fig. 1. It consists of a discharge lamp, which irradiates one
end of a gas cell containing the element chosen in gaseous
form through a system of filters and polarizers. At the oppo- detector.8 A typical two-cell sensor arrangement is shown in
site end, a solid state photodetector measures the intensity of Figs. 2共a兲 and 2共b兲. Another important issue associated with
the incident light. The electrons in the gas cell will precess these magnetometers is the use of high-level rf signals for
about the axis of the external magnetic field at the Larmor optical pumping. These need to be carefully controlled and
frequency of the chosen element, and modulate the intensity shielded to avoid interference with vector magnetometers
of the light incident upon the photodetector at the same rate. and sensitive electronics, communications, and other space-
Thus, the output from the photodetector is an ac signal at the craft systems. The high power dissipation and associated
Larmor frequency given by thermal flux and to the use of nonmagnetic metallic shields
around the sensor will often give rise to significant measure-
f ⫽ ␶ e B/2␲ , 共4兲
ment errors associated with thermoelectric currents due to
where ␶ e is the electron gyromagnetic ratio for the chosen Thompson and Seebeck effects.
element. For helium ( ␶ e /2␲ ) corresponds to 28 共Hz/nT兲, Scalar magnetometers of the types described above have
considerably larger than the value of 0.042 576 02 共Hz/nT兲 a typical dynamic range of 20 000–70 000 nT and cannot be
obtained from proton precession instruments, making pos- used to measure weak fields such as those associated with the
sible higher time resolution measurements. The correspond- interplanetary medium, the Moon, Mars, Venus, or Mercury.
ing figures for cesium and rubidium are 7 and 4.67 共Hz/nT兲, Their principal use is in low Earth orbit and in conjunction
respectively. with vector magnetometers to acquire very accurate mag-
The output of the photodetector is amplified and fed netic field mapping data.65 The Cassini mission to Saturn
back to a coil wound on the cell, causing the system to os- included a scalar helium magnetometer to map this planet’s
cillate continuously at the Larmor frequency. Because of the remarkable axisymmetric field.64 A scalar magnetometer is
high frequencies involved, it is possible to measure very often used as an absolute calibration tool in conjunction with
small magnetic field variations 共⭓0.01 nT兲 superimposed on a vector magnetometer to allow real time determination of
large background fields. correction factors. Laboratory-based proton precession mag-
If the external magnetic field is aligned with the axis of netometers are routinely used as primary calibration stan-
the gas cell the amplitude of the Larmor signal decreases to a dards.
minimum, giving rise to the appearance of ‘‘null zones.’’ For The mass and power consumption associated with scalar
this reason most optically pumped scalar magnetometers magnetometers can be significant. Classical proton preces-
used for space measurements utilize two cells oriented at sion magnetometers may draw several tens of watts during
different angles to fill in the null zones associated with each the polarizing cycle while the mass of the sensor alone may
exceed 1 kg and that of the electronics 2–3 kg. Helium,

rubidium, and cesium magnetometers require significant rf
power for excitation, with 10–15 W being typical figures.
The high levels of rf present in these instruments also require
close attention to shielding and the prevention of generation
of local magnetic fields by rectification or thermoelectric ef-
fects. The Overhauser and helium scalar magnetometers
mass and power requirements can be reduced to a few watts
and 1–2 kg with modern components and specialized design
techniques.65,66 The typical noise performance of these in-
struments is in the range of 0.01–0.03 nT root mean square
共rms兲 over a bandwidth of 0–1 Hz, and the practical upper
limit to their dynamic range is ⬃1.5– 2⫻105 nT, an impor- FIG. 3. Block diagram of a generic, single axis fluxgate magnetometer. A
tant consideration for missions to planets with strong mag- tuned single axis ring core sensor is shown.
netic fields like Jupiter.
also widely used in countless industrial, military, and scien-
tific applications. The fluxgate sensor, as its name implies, is
B. Vector magnetometers
a device which is used to ‘‘gate’’ the ambient magnetic flux
Vector magnetometers are, by a very large margin, the threading a sensing coil, converting it from a time stationary
most widely used type of instrument for magnetic field mea- field into a time varying field. The latter gives rise to induced
surements in space. In addition to providing information voltage in a sensing coil proportional to the strength and
about the strength of the ambient field they also measure its direction of the field. Gating of the ambient field is accom-
direction and sense. Triaxial orthogonal arrangements of plished by driving cyclically to saturation a high permeabil-
single axis sensors are used to measure the three components ity, nonlinear magnetic core with a large amplitude excitation
of the ambient field in a coordinate system aligned with the signal. A differential coil wound around the core and the
sensor magnetic axes. In contrast to proton precession and excitation coil is used to sense the output signal. When the
optically pumped scalar magnetometers whose accuracy is core material is in the saturated state its permeability is very
determined by quantum mechanical constants, vector magne- low and the total magnetic flux threaded by the sensing coil
tometers must be calibrated using accurately generated mag- is not very different from that which would exist in free
netic fields, both in strength and direction, at specialized fa- space. However, when the core is in its high permeability
cilities. The magnetometer output for the zero applied field, state, the flux threaded through it and the pick-up coil is
scale factor, and stability over temperature and time depends greatly increased depending on the effective permeability of
on electrical component values which may drift as the instru- the core 共and core geometrical shape兲. Thus, by cyclically
ments age or are exposed to effects of the space switching the core in and out of saturation at frequency f, the
environment.8,76 – 80 The alignment of the sensors may change ambient magnetic flux will be modulated at twice this fre-
as thermal or vacuum induced stresses deform the supporting quency and a corresponding ac voltage will appear at the
structures or reference coordinate systems.77,81,82 In spite of terminals of the sensing coil. The amplitude of this signal is
these challenges, extremely reliable, low power 共⬍1 W兲 high proportional to the magnitude and direction of the ambient
performance vector instruments capable of measuring mag- field, and its phase with respect to the excitation signal 共0° or
netic fields over a very large dynamic range, ⬍5⫻10⫺3 to 180°兲 will depend on the sense of the field. In general, bal-
over 2⫻106 nT, were developed in the early 1970s for outer anced core arrangements are used to prevent the excitation
planets exploration.35,41,83 They are capable of operating over signal from appearing at the pick-up coil and to avoid gen-
a wide temperature range and have proven to be extremely erating a large signal at the excitation frequency. Many flux-
radiation tolerant. Ultraprecise vector instruments with arc- gate sensor geometries have been used for space-based in-
sec resolution have also been used to map the Earth’s mag- struments, from commercial sensors using proprietary helical
netic field from orbit with unprecedented accuracy, both in high permeability cores to sensors using ultralow noise ring
magnitude as well as in direction.77,79,84,85 The CLUSTER II cores developed by government laboratories. The very high
mission launched in August of 200086 included a comple- performance of the latter has made them the sensors of
ment of four spacecraft flying in formation, each incorporat- choice for advanced space missions.8,49,69,76,88 –94
ing dual, high performance fluxgate magnetometers designed A block diagram of a typical fluxgate magnetometer is
to make three-dimensional measurements of the ambient shown in Fig. 3. A reference signal at frequency 2 f 0 is de-
magnetic field.62 One of the primary objectives of that mis- rived from a stable oscillator and applied to a divider, from
sion’s magnetic field instrumentation was to obtain, for the which the excitation signal at frequency f 0 is used to drive
first time, measurements of local currents through the rela- the sensor toward saturation 共in this case a ring core sensor
tion J⫽“ÃB. with a toroidal excitation winding兲. A differential sensing
The fluxgate magnetometer, which was invented by coil is wound around the outside of the ring core and, in most
Aschenbrenner and Goubau67 and developed during World applications, tuned to the second harmonic of the drive fre-
War II as a submarine detector,69,87 is the most common type quency, 2 f 0 . Thus, if both core halves are identical, no sig-
of instrument used in space platforms. This type of sensor is nal at the excitation frequency will appear at the sense wind-
ing terminals. The presence of an external magnetic field will

cause the appearance of a signal at frequency 2 f 0 and other
even harmonics of the excitation frequency at the terminals
of the sensing coil. This is due to the fact that one half of the
ring core will not be balanced with respect to the other, simi-
lar to operation of a balanced modulator. This signal is am-
plified by the tuned ac preamplifier and applied to the syn-
chronous detector or ‘‘lock-in’’ system. The magnitude of the
output of this detector is proportional to the amplitude of the
signal present at the output of the preamplifier, whereas the
polarity of the output depends on the phase of the input
signal with respect to the reference frequency 2 f 0 . The out-
put of the synchronous detector is then applied to a high gain
integrating dc amplifier whose output is used to generate a
current, which is fed back to the sense winding in the flux-
gate sensor. The feedback current flowing through the sens-
ing coil produces a magnetic field that opposes 共and cancels
almost completely兲 the original external field and the flux-
gate sensor is used essentially as a null detector. This feed-
back arrangement yields an instrument with excellent linear-
ity whose dc response is given by
e 0 共 B, ␪ 兲 ⫽kR f B cos共 ␪ 兲 ⫹V z , 共5兲
where e 0 (B, ␪ ) is the output voltage, k is a constant related to
the physical characteristics of the sensing coil with dimen-
sions given in 关ampere/Gauss兴, R f is the value of the feed-
back resistor, B is the scalar magnitude of the field applied,
and ␪ is the angle between the magnetic axis of the fluxgate
sensor and the direction of the external field. V z is a small dc
offset voltage, which is present when the external field is
zero and is produced by asymmetries, offsets, and imbal-
ances in the system. The magnetometer scale factor can be
easily modified by changing the value of R f and this is done
in instruments that must cover a large dynamic range like
those used in the Voyager, MGS, ACE, WIND, and LP
missions.32,45,46,83 For a feedback system with a single inte-
grator the transfer function is given by FIG. 4. 共Color兲 Typical vector magnetometer instruments. 共a兲 The sensor
seen on top of the electronics box is mounted remote from the spacecraft
H 共 s 兲 ⫽1/关共 s 2 / ␻ 2n 兲 ⫹2s 共 k/ ␻ n 兲 ⫹1 兴 , 共6兲 bus, typically on a deployable boom. 共b兲 Block diagram of a high perfor-
mance, space-based magnetometer system. Note the extensive use of redun-
where ␻ n and k are the characteristic natural frequency and dant systems to ensure reliability in missions of long duration.
damping factor of a second order system, respectively. The
latter is directly proportional to the value of R f and inversely average power at a sample rate of 1–2 samples/s. Figure 4共a兲
proportional to the ac gain in the system so simultaneous shows a typical high performance magnetic field instrument
adjustments need to be made to the open loop gain and time comprising a triaxial sensor assembly and an electronics box,
constants to preserve the stability of the feedback loop for all which includes signal, data processing, and spacecraft inter-
scale factors selected. face circuits. The sensor shown on the top of the electronics
Mass and power requirements for fluxgate magnetome- box in Fig. 4共a兲 is mounted far from the spacecraft bus to
ters are significantly reduced with respect to those associated minimize magnetic interference from spacecraft systems and
with scalar instruments. Typical general purpose, fluxgate other instruments. A block diagram, which illustrates the
magnetometers for attitude sensing and control require ⬍0.5 typical elements of a fully redundant instrument used for
kg and ⬍0.5 W, respectively, with miniature units widely planetary exploration, is shown in Fig. 4共b兲.
available. The latter integrate the sensor with the electronics A second type of vector instrument that has been used in
in a single package, which is not necessarily an optimal ar- a few planetary spacecraft is the Vector helium
rangement for all applications as will be discussed below. magnetometer.41,71,75 Its operating principles are the same as
Space research-grade instruments that incorporate wide dy- those of the optically pumped magnetometers described ear-
namic range capabilities and advanced digital processing lier in Sec. III A, except that a Helmholtz coil system around
techniques have been built and they consume as little as 1 W the gas cell has been added. This coil system is used to
and weigh 1 kg per unit.34,37,45,62,83 Duty-cycled instruments generate synchronous sweep fields in two orthogonal planes
have also been developed that draw just a few microwatts that intersect along the optical axis by means of an auxiliary
electronic system. These sweep fields produce synchronous where in this review, this advantage cannot be totally real-
modulation in the photodetector light output, which can be ized in practical spacecraft applications due to external ef-
used to derive vector information about the external mag- fects. A state-of-the-art fluxgate magnetometer will exhibit
netic field. Typically, a frequency of 200 Hz is used for the zero level drift of ⬍0.2 nT/yr and ⬍0.5 nT over the tempera-
sweep field and the resulting signal at the photodetector is ture range of ⫺55 °C⬍T⬍75 °C. The noise performance of
amplified, synchronously rectified, and the current fed back fluxgates is strongly affected by mechanical and thermal
to the sensor assembly to cancel out the field applied. The stresses, which vary over time and with exposure to extreme
operation is very similar to that implemented in a fluxgate environments. These are usually at their maximum during
magnetometer. Some differences are that the excitation and instrument development and testing and it is common for
signal frequencies are identical and vector information is de- noise performance to improve over time after the launch,
rived not from individual sensors but from sweeping fields particularly if exposure to the environment is constant or
applied to a common sensor along two orthogonal planes. limited. The Voyager mission magnetometers improved their
The directions of the sensing ‘‘axes’’ are thus derived elec- noise performance from 0.005 nT rms immediately after the
tronically and are not uniquely related to the mechanical launch in 1977 to 0.001 nT rms 共0–1 Hz bandwidth兲 during
alignment of individual sensors, as in fluxgates. The transfer the Jupiter fly-by 2 years later.
function of the vector helium magnetometer 共VHM兲 is iden- The traditional fluxgate magnetometer design shown in
tical to that of the fluxgate magnetometer given by Eq. 共5兲. Fig. 3 has also evolved to include ‘‘short-circuited’’
The mass and power requirements of VHMs are still higher fluxgates,95,96 ‘‘digital’’ fluxgates,97–99 fast Fourier transform
than comparable performance fluxgate magnetometers with 共FFT兲-based fluxgates,97 ‘‘current output’’ fluxgates,100 and
2–3 kg and 4 – 6 W being representative figures. The power many others. The basic principles remain the same, with the
required to generate the sweep fields over the volume of the differences being mostly in how the error signal is processed
cell and increasing transfer function nonlinearity limit the to generate the feedback field.101 It is difficult to assess the
upper limit measurement capability to ⬃1.5⫻105 nT. As is advantages or disadvantages of each implementation given
the case for all optically pumped magnetometers, the lifetime the multiplicity of requirements and constraints associated
of the gas cell and the stability of the light source are most with each space mission. However, low cost, accuracy,
important considerations for missions of long duration low power consumption, small mass and volume, large
共⬎3–5 yr兲. In 1978 –1979 NASA funded Scintrex, Canada, dynamic range, low noise, and good zero level stability
to study the aging process of cesium vapor lamps for the remain the premier characteristics sought in space-based
MAGSAT mission. It was concluded that lifetimes of the magnetometers.8,49,76,95,102
order of ‘‘several years’’ could be achieved under a stable
thermal regime which maintained the lamp stem at 50 °C and IV. ELECTRON REFLECTION MAGNETOMETRY
the lamp envelope some 150–160 °C hotter. New techniques
in striking the lamps and bringing them to their stable oper- The Mars Global Surveyor and Lunar Prospector mis-
ating point were also necessary to extend their operational sions included magnetic field investigations capable of mak-
life. In the case of helium magnetometers, like those on the ing in situ measurements with high performance fluxgate
Pioneer 10 and 11 missions, the cell and lamp were con- magnetometers like those described above, as well as remote
structed using special glass and seals to minimize gas leak- sensing instruments capable of deriving the intensity of the
age. magnetic field at a significant distance from the spacecraft.
The stability of the zero level or bias over time and The latter use the technique of electron reflectometry, which
temperature and the noise performance of a vector magneto- is based on the fact that ambient electrons, like those in the
meter are some of the most important performance param- solar wind and in planetary environments, can travel along
eters to be considered for weak field measurements such as magnetic field lines and be reflected when they encounter an
those of the IMF, the distant terrestrial magnetosphere, and increase in flux density associated with a source of
around unmagnetized bodies. The data reduction and analy- magnetism.26,29,32 Thus, an electron reflectometer is an in-
sis effort associated with weak field magnetic field investi- strument that measures the flux of electrons moving along
gations are traditionally dominated by ‘‘zero level’’ determi- the local magnetic field lines and the ‘‘pitch angle distribu-
nation issues. The scale factors and intrinsic sensor tion’’ for both directions. The presence of a magnetic source
alignment are stable and once calibrated remain largely con- in the path of the electrons will cause them to be reflected,
stant for the life of the mission. The noise power spectral altering the pitch angle distribution. This information can be
density of a high performance fluxgate magnetometer is typi- cast in the form of a reflection coefficient, which is directly
cal of ‘‘shot noise’’ devices and characterized by a 1/f spec- proportional to the intensity of the field at the point of reflec-
trum with a typical value of ⬃10⫺5 (nT2 /Hz) at 1 Hz for tion. This sensitive technique is particularly useful in the
space research-grade instruments. The source of the noise is case of the Moon where no atmosphere exists to absorb the
attributed to Barkhausen-like mechanisms that affect the mo- electrons and thus surface magnetic fields can be measured
tion of domains in ferromagnetic material in the sensor from an orbiting spacecraft.26 –28 In the case of Mars, the
cores.49 The absence of ferromagnetic materials in optically presence of an atmosphere limits the measurement altitude to
pumped magnetometers results in a flat noise spectrum for the top of the atmosphere or ⬃100 km although the existence
these devices and, in principle, improved long term stability of strong crustal fields limits the usefulness of this
over fluxgates. However, and for reasons discussed else- technique.32 An important advantage of electron reflection
FIG. 5. 共Color兲 Artist’s illustration of the Voyager spacecraft and mission to the outer planets. The 14 m long magnetometer boom can be seen in the lower
half of the picture and dominates the dimensions of the spacecraft when deployed. Four triaxial sensors are mounted on this boom, one at the tip, a second
two thirds of the distance from the tip to the base, and two on the support canister.
magnetometry is the improved spatial resolution obtainable agers, engineers, and scientists. Since it is practically impos-
since the ‘‘footprint’’ of the electrons at the point of reflec- sible to reduce the stray spacecraft magnetic field to the
tion is their Larmor radius 共⬍ 2 km for lunar fields兲. The smallest levels required for sensitive measurements, place-
spatial resolution obtainable from in situ measurements with ment of the magnetic sensors away from the main body of
conventional magnetometers aboard an orbiting spacecraft is the spacecraft is commonplace. This technique exploits the
at best equivalent to the orbit altitude above the surface. fact that the magnetic fields produced by finite sources de-
crease rapidly with distance, proportionally to 1/r 3 as a mini-
V. ROCKETS AND SPACECRAFT AS MAGNETIC mum 共higher order multipoles decrease even more rapidly兲,
FIELD MEASUREMENT PLATFORMS where r is the distance to the source. The above is true when
The measurement platforms for the instruments de- the dimensions of the source are smaller than the distance to
scribed above are spacecraft, rockets, and balloons, and they the measurement point. Many missions use long, deployable
include complex systems of mechanical, electrical, and elec- ‘‘booms’’ or masts for this reason and they must be rigid and
tronic components. They all have the potential of producing preserve the alignment required between the magnetic sen-
magnetic fields of their own and careful control techniques sors and the attitude determination sensors mounted on the
must be used to minimize errors introduced by these un- main spacecraft body, which imposes limits on the practical
wanted sources. Batteries, solar arrays, motors, wiring, ma- length of booms. The magnetometer boom of the twin Voy-
terials, etc. must be especially designed and/or selected to ager spacecraft launched in 1977 to explore the outer planets
minimize the generation of ‘‘stray’’ magnetic fields that will of the solar system had a length of 14 m 共42 ft兲 when fully
affect measurements of the ambient field. The design and deployed. An artist’s illustration of the Voyager spacecraft
implementation of a magnetically ‘‘clean’’ spacecraft that shown in Fig. 5 gives a sense of the relative dimensions
meets the stringent requirements of a high accuracy Earth- involved. For a given spacecraft design, the trade-off be-
orbiting or interplanetary mission is an extremely demanding tween boom length and level of magnetic cleanliness re-
task that has tested the fiber of many seasoned project man- quired in the main body and major subsystems is a major
FIG. 6. 共Color兲 Illustration of the Mariner 10 spacecraft showing an example of a rigid magnetometer boom and a dual magnetometer arrangement. The boom
is 6 m long and composed of two rigid segments which deploy sequentially.
decision that must take into account many conflicting re- by a center lanyard or restraining mechanism. This lanyard is
quirements, including whether the spacecraft is spin or three also used during deployment to control the rate at which the
axis stabilized. Details on the magnetic interference control, structure unwinds out of the canister. Booms as long as
spacecraft testing, and magnetic cleanliness programs used 45–50 m have been manufactured and mass-per-unit-length
on early spacecraft were given in a review by Ness.8 The ratios of the order of 150 g/m are achievable. Careful control
early boom designs were mostly of rigid tubular type, folded of expansion coefficients and the effect of solar illumination
against the top or side of the spacecraft in the launch con- on the elements yield booms, which exhibit torsional and
figuration. To achieve longer lengths several rigid segments bending stability of the order of a fraction of a degree for
could be deployed sequentially, but the complexity of the lengths up to 10 m or more. An important consideration in
mechanisms required to insure correct deployment limit the computing the total mass associated with a long magnetome-
maximum number of segments to no more than three. Other ter boom system is the weight of the cables connecting the
boom types included ‘‘scissor’’ booms, helically wound sensor at the tip of the boom, since the electronics are located
booms, etc. Figure 6 shows the rigid boom used in the Mari- in the main body of the spacecraft. This mass can easily
ner 10 spacecraft, while Fig. 7 illustrates the scissorstyle exceed that of the boom itself by a significant factor. The
boom used in the MAGSAT spacecraft. rigidity of the sensor cables in cold temperatures and its im-
In the early 1970s a new type of truss-like, deployable pact on boom deployment safety margins are also critical
boom was developed, generically called an ‘‘astromast.’’ issues, particularly in missions where successful boom de-
This very interesting structure is shown deployed in Fig. 8 ployment is a mission critical event.
and is characterized by an extremely favorable length-to- The high cost of deployable booms, their potential ef-
mass ratio. Three, lightweight main longerons, typically fects on spacecraft dynamics, mission success, and magneto-
made of fiberglass, are stabilized by a series of battens and meter alignment, and the complexity of the associated me-
diagonal elements to form a rigid truss. In the undeployed chanical systems limit the affordable boom length that can be
state 共Fig. 9兲, the longerons are coiled inside a canister realized for a given spacecraft design. The dual magnetome-
mounted on the spacecraft and are prevented from deploying ter technique was introduced in 1971 by Ness and co-
FIG. 7. 共Color兲 MAGSAT spacecraft and its 6 m scissors boom. Optical mirrors are mounted on the magnetometer sensor platform to ‘‘transfer’’ its
orientation to the main body of the spacecraft using infrared beams.
FIG. 8. 共Color兲 Deployed astromast-style boom used in NASA’s WIND and POLAR spacecraft 共12 and 6 m in length, respectively兲 to separate the sensors
from the body of the spacecraft.
FIG. 9. 共Color兲 Undeployed astromast-style boom showing the center lanyard and deployment mechanism.
workers to ease the problem of making sensitive magnetic The dual magnetometer method is illustrated schemati-
field measurements in the presence of a significant spacecraft cally in Fig. 10. The spacecraft-generated magnetic field de-
field.103,104 This method is based on the experimental obser- creases with distance due to the finite size of the sources,
vation that beyond a certain distance most spacecraft- which are assumed located at the center of the spacecraft,
generated magnetic fields decrease as expected for a simple and leads to the existence of a spatial gradient between the
dipole source located near the center of the spacecraft two magnetometer sensors. The ambient field being mea-
(⬃1/r 3 ). Thus it can be shown that if two magnetometer sured is identical at both sensors because its spatial gradient
sensors are used, mounted along a radial boom and located at is insignificant over the dimensions of the spacecraft and
distances r 1 and r 2 , respectively, it is possible to uniquely boom. Thus, each sensor measures a different mixture of
separate the spacecraft-generated magnetic field from the ex- spacecraft and ambient field. For the special case of a dipolar
ternal field being measured. If we denote by B1 and B2 the spacecraft field, the fields due to the spacecraft at the two
vector fields measured at radial locations 1 and 2 with r 2 sensor locations are related by a simple proportionality con-
⬎r 1 , the ambient field and the spacecraft field at location r 1 stant. Thus, the ambient and spacecraft field can be separated
are given by
Bamb⫽ 共 B2 ⫺ ␣ B1 兲 / 共 1⫺ ␣ 兲 , 共7兲
Bs/c1 ⫽ 共 B1 ⫺B2 兲 / 共 1⫺ ␣ 兲 , 共8兲

where
B1 ⫽Bamb⫹Bs/c1 , 共9兲
B2 ⫽Bamb⫹Bs/c2 , 共10兲
Bs/c2 ⫽ ␣ Bs/c1 , 共11兲
␣ ⫽ 共 r 1 /r 2 兲 3 . 共12兲 FIG. 10. Illustration of the dual magnetometer technique and associated
formulas. The spacecraft-generated magnetic field (Bs/c) is assumed to be
Note that Eqs. 共11兲 and 共12兲 assume that the spacecraft dipolar and centered at the body of the spacecraft while the external field
field can be represented accurately by that due to a dipole (Bamb) is gradient-free over the separation of the sensors. Bobs is the field
quasicentered on the main body. measured which includes a constant offset Bz0 .
analytically as shown above. An immediate benefit of the time and resources, the magnetic fields associated with ex-
dual magnetometer technique is that shorter booms can be isting or heritage components and systems. Current philoso-
used to achieve a level of accuracy equivalent to that af- phy concentrates on first minimizing time-variable space-
forded by a much longer boom. The small increase in mass craft fields associated with soft magnetic materials and
represented by the second magnetometer is more than offset circulating currents in power systems, loads, motors, actua-
by the reduction in boom length and associated cabling. It is tors, solar arrays, etc. since these give rise to unknown and
customary to see this arrangement in spacecraft used for variable errors which are difficult to measure and remove
outer solar system exploration missions where the from the data. Truly static fields associated with hard mag-
spacecraft-generated magnetic field at the sensors may be netic materials can be measured prelaunch and subtracted
many times larger than the weak ambient field. A particular from the measurements assuming that they are time and tem-
advantage of the dual magnetometer method is that it allows perature stable. The techniques used for static and dynamic
unambiguous real time identification and monitoring of the field compensation and minimization include permanent
time variation of the spacecraft-generated field. In a single magnets, single point grounding systems for power distribu-
magnetometer system a change in the field measured can be tion, twisted pair wiring for all significant loads, and the use
attributed to either a change in the spacecraft-generated field, of dummy current loops to create equal but opposite signa-
or a change in the ambient field which leads to a unresolv- tures from circulating currents. Advanced computer con-
able ambiguity in interpretation. Last and most important, trolled laboratory magnetometers that use synchronous sam-
the use of two sensors provides full measurement redun- pling techniques to reject magnetic signatures at 50/60 Hz
dancy, which has proven essential for success in missions of and higher power line harmonics have been developed to
long duration such as those to the outer planets.83 carry out sensitive measurements 共⬃⬍0.2 nT sensitivity兲 in
Some miniature magnetometers used for attitude deter- the Earth’s field and in the normal environment of laboratory
mination and control integrate the fluxgate sensors and elec- and spacecraft development and integration facilities. These
tronics into a single package. This approach is adequate if instruments have largely eliminated the high cost and logistic
accuracy of the order of 10%–20% is acceptable, such as in complexities associated with time-consuming measurements
momentum wheel desaturation applications. However, the at specialized distant magnetic test facilities. They typically
proximity of the sensor to the electronics introduces addi- incorporate real time visualization of the magnetic field data
tional biases and hysteresis effects and limits the environ- being measured, and have literally transformed spacecraft
ment to which the package can be exposed. Fluxgate sensors, magnetic control techniques by eliminating the ‘‘black
by themselves, are capable of withstanding penetrating radia- magic’’ aspects of the field and significantly reducing costs
tion doses of tens of megarads and temperature extremes of and complexity. It is important to emphasize that the useful-
⫺100 to ⫹200 °C without major difficulty, a fact that is not ness of these measurements made in the presence of the
true for the signal processing electronics. Thus, integrated Earth’s field is limited to those applications where induction
sensor-electronics instruments must be mounted within the effects are small. When large structures made of high perme-
main body of the spacecraft where temperature extremes are ability materials, particularly those having a large length-to-
within operational limits but unfortunately the close proxim- diameter ratio 共e.g., magnetic torquers兲, are present, induc-
ity to many sources of stray magnetic fields, substantially tion effects will dominate, thereby introducing great
limits their accuracy. difficulties in interpretation of the data.
The complexity and cost of the early spacecraft magnetic Another important magnetic control tool has been added
control programs were truly staggering by today’s standards. by today’s widespread availability of mathematical modeling
Special nonmagnetic materials were selected for custom- software in personal computers, which has allowed rapid nu-
built passive and active electronic components and very of- merical and/or analytical computation of the magnetic fields
ten were incompatible with reliability considerations. Every generated by permeable structures and different circulating
spacecraft component was magnetically characterized in de- current geometries. These are particularly useful in the case
tail at magnetic test facilities and either compensated for or of solar arrays where many linear current segments and cur-
its design modified to use nonmagnetic parts. The Pioneer 10 rent sheets are involved and a minimal magnetic signature
and 11 spacecraft launched in 1973 to explore Jupiter and due to the variable circulating currents is desired. In some
Saturn105 were able to achieve a residual static field of less cases, like in the Mars Global Surveyor mission, the large
than 0.01 nT 共basically, the limit in practical measurement size of the solar arrays 共⬃15 m2兲 made it possible to use
capability at magnetic test facilities兲 at the end of a 3 m them as booms to place the magnetometer sensors some 5 m
boom where the principal magnetometer sensor was located. from the main body of the spacecraft, at the outer edge of
Similar results were obtained for the Interplanetary Monitor- each panel, obviating the need for a classical deployable
ing Platforms 共IMP兲 launched to explore the Earth’s mag- boom 共see Fig. 11兲. However, that approach required that the
netosphere and the interplanetary medium in the late current paths in the array and solar cells be modeled, calcu-
1960–1970s.8 lated, and manufactured 共to precision of 1.5 mm兲 to cancel
The evolution of technology markets and spacecraft cost out the magnetic field at each sensor location. The final con-
reduction efforts has made it impossible to implement the tamination level achieved in Mars’ orbit for the solar array
kind of magnetic control programs described above in con- contribution was ⬃0.25 nT. This impressive performance
temporary space missions. Most efforts are directed towards was unfortunately compromised by the presence of large ar-
the goal of minimizing or compensating, within available rays of uncompensated magnets in the traveling-wave-tube
FIG. 11. Mars Global Surveyor spacecraft, solar panels, locations of mag-
netometer sensors 共OB and IB兲, and high gain antenna which includes the
traveling-wave-tube amplifiers and their magnet assemblies.
amplifiers 共TWTAs兲, which were mounted on the back side

of the articulated high gain antenna 共HGA兲共see Fig. 11兲,
thus creating a source of time varying fields as the spacecraft
orbited Mars. The magnitude of the error signal created by
the TWTA magnets is illustrated in Fig. 12 where data ac-
quired when the solar panels were being articulated is shown
for each of the triaxial magnetometer sensors mounted at the
edge of the solar panels 共Bx, By, and Bz兲. Under ideal con-
ditions no variations in any of the components should have
been observed as a function of panel position and angle.
However, the presence of the TWTA tube magnets leads to
the time varying signals shown in Fig. 12, and a measure-
ment error as large as ⬃9 nT, almost three times the average
value of the interplanetary magnetic field of Mars. These
data were acquired during special maneuvers and solar array FIG. 12. Mars Global Surveyor spacecraft-generated magnetic fields intro-
duced by the magnet assemblies in the TWTAs. Articulation of the solar
motion designed to estimate parameters for analytical models panels and antenna introduces time variable signals in an otherwise rela-
of the spacecraft-generated field to be subtracted from the tively constant background. The primary and secondary definitions apply to
raw data. the OB and IB sensors, respectively, shown in Fig. 11. The X,Y,Z coordinates
From the foregoing, it is clear why, in a cost and sched- refer to the spacecraft’s principal axes.
ule constrained environment, many space mission managers
ring magnetic fluctuations falls off with increasing frequency
prefer spacecraft without sensitive magnetometers! Experi-
roughly as (1/f n ) with n varying approximately between 1.5
mental success depends critically on outstanding systems en-
and 3 depending on the phenomena being observed. The pro-
gineering, close collaboration among experimental scientists,
cessing and digitization of the magnetometer analog output
and spacecraft designers, and real management leadership.
signals for eventual transmission to Earth must therefore take
into account these spectral characteristics and the desired
VI. ANALOG SIGNAL AND ONBOARD DATA
time resolution. Analog-to-digital 共A/D兲 conversion resolu-
PROCESSING FOR SPACE MAGNETOMETERS
tions of 12 to ⬎16 bits are commonly used to recover as
The study of the frequency spectrum of dynamic pertur- much of the spectrum as possible without having to resort to
bations of the ambient magnetic field is a powerful tool that prewhitening filters or similar spectrum-shaping techniques
is used to identify the type and characteristics of waves and which increase power consumption. The availability of low
other time variable phenomena detected by spacecraft mag- power, very-high-resolution A/D converters, such as those
netometers. Although most of the applications and research using sigma–delta techniques, has allowed the simplification
activities are concentrated in the 0 to ⬃10 Hz frequency of earlier instrument designs that resorted to complex step-
range, higher frequency and time resolution data are ex- biasing techniques to achieve higher resolution than that ob-
tremely useful for identifying interaction boundaries, plasma tainable with 8 –12 bit A/D converters.8,32,45,46,79,90,106
waves, magnetohydrodynamic shocks, and other fast phe- To prevent aliasing errors the magnetometer output must
nomena that take place in planetary magnetospheres and in be sampled at least twice as fast as the highest frequency of
the solar wind. Telemetry and power resources, particularly interest 共the Nyquist rate兲, a fact that coupled to the high
those on spacecraft traveling to the outer reaches of the solar digital resolution required results in the generation of volu-
system, are extremely limited and some form of onboard minous amounts of raw digital data onboard. These data can-
data processing is required to reduce the rate of the data not be transmitted directly to Earth because of telemetry and
transmitted to Earth. The power spectrum of naturally occur- communication system limitations and intelligent data pro-
cessing techniques and algorithms are used to implement usually aligned with the spin axis, the magnetic field compo-
data compression and extraction of information from the on- nents in the spin plane will exhibit large amplitude modula-
board raw data. Many forms of data compression, both loss- tion at the spin frequency. This requires that additional band-
less and lossy, have been used to increase the information width be provided to the magnetometer to resolve the spin
content per magnetometer bit transmitted to the ground. Av- frequency accurately and avoid aliasing, or that spin modu-
eraging and/or data differencing between adjacent samples lation be removed on board prior to data compression and
with periodic baseline recovery is the least complex method transmission to the ground. The same applies if onboard
and is easily implemented with simple logic algorithms. For FFT-based spectral analysis is used: the large amplitude spin
a typical magnetic field power spectrum it achieves compres- induced signals must be removed from the data prior to FFT
sion factors in the range of 2–5 without major penalties or processing to avoid spectral leakage and dynamic range
storage requirements and hence it is often used in data rate problems in digital signal processor 共DSP兲 algorithms. Mod-
limited missions where accurate inversion of the compressed ern space-based magnetometers like those on the WIND and
signal with minimal spectral distortion is an important objec- ACE spacecraft incorporate data de-spinning capabilities in
tive. At the other extreme, complex adaptive lossless algo- their digital processing units45 as part of the onboard FFT
rithms exist which can achieve compression factors in the processor. The de-spinning algorithm utilizes onboard infor-
range of 10–100 but require significant onboard storage and mation on the spin rate and spin phase to transform the mag-
processing capabilities and extremely careful time synchro- netic field data to a coordinate system aligned with a refer-
nization management. It is important to emphasize that the ence direction, usually the Sun.
greater the compression efficiency 共or entropy per bit兲 the The attractive advanced onboard computational capabili-
greater the sensitivity of the system to bit errors introduced ties provided by today’s microprocessors, DSPs, and large
by communication channel performance which may or may memories have to be traded off against the costs and com-
not be correctable. Where significant light-trip times are in- plexities associated with the space environment, the avail-
volved, retransmission of data packets 共as in ground data ability of high reliability components, fault-tolerant software
networks兲 is not an option since space magnetometers gen- development, and testing costs. The high radiation environ-
erally require the acquisition of continuous data streams ment associated with orbits where spacecraft spend a signifi-
without significant gaps and fixed data rate allocations are cant amount of time in the radiation belts or missions to
commonplace. Typical effective data rates for spacecraft Jupiter like Voyager and Galileo severely limits the types of
magnetic field experiments may range from a few bits per devices that can be considered for flight instruments.
second to 2 kbits/s or more while for rockets and balloons Latchup and single-event upsets in memories, processors,
rates of up to several tens of kbits/s are common for direct and complex logic devices must also be considered in de-
links. signing magnetic field instruments for space applications.
Onboard spectrum analyzers based on FFT algorithms The radiation protection afforded by the spacecraft bus, fuel
have been used in advanced magnetic field experiments like tanks, and subsystems can be substantial and is usually taken
the one developed for the WIND and ACE spacecraft.45,46 into account in trade-off studies. Magnetometers with sepa-
The availability, starting in the early 1980s, of high reliabil- rate sensors and electronics can take full advantage of this
ity, low power digital signal processor chips and radiation protection since the sensors are typically much more radia-
hardened memories made possible the efficient implementation tolerant than the electronics and are placed on booms far
tion of FFT engines to process high sample rate magnetome- from the main body of the spacecraft for magnetic contami-
ter data on board spacecraft. Since the magnetometer data nation reasons.
consist of three simultaneous time series that correspond to
the three orthogonal components of the field, each one must VII. PRE-LAUNCH AND IN-FLIGHT CALIBRATION AND
be stored and processed independently. Phase information is ALIGNMENT TECHNIQUES
important, hence both real and imaginary components of the
Accurate calibration of space-based magnetometers with
power spectral matrix are preserved to compute auto- and
a dynamic range of a few nanoteslas to hundreds of thou-
cross-spectral power densities as a function of frequency. In
sands of nanoteslas full scale presents unique problems
general, logarithmic frequency scales are desired and these
which require special techniques and facilities. Among the
must be synthesized from the resulting linear FFT transfor-
many issues involved, the following are typical.
mation. A number of real time and/or frequency domain data
weighting windows are also implemented to attenuate the 共1兲 Shielding or cancellation of the Earth’s magnetic field
spectral contamination effects of discrete length data sets. and the generation of accurately known magnetic fields
Spinning spacecraft are usually preferred for simple at- 共magnitude and direction兲 including very weak fields,
titude stabilization and studies of the plasma and charged ‘‘true zero field,’’ and the minimization of thermoelectri-
particle space environment if full solid angle coverage by the cally induced magnetic fields caused by temperature gra-
instruments is desired. The spinning motion of the spacecraft dients. The typical frequency range considered is 0 to
sweeps the limited field of view of particle instruments 25–50 Hz.
through solid angle space to achieve a much larger spatial 共2兲 The extreme environmental range over which the mag-
sampling of the charged particle population 共the distribution netic field sensors and their electronics must operate and
function兲. Since the magnetometer is fixed to the body of the be characterized 共temperature, pressure, radiation, vibra-
spacecraft or to the boom and one of its sensing axes is tion, shock兲. The parameters of interest are, at a mini-
FIG. 13. 共Color兲 Voyager magnetometer boom, supports, and magnetic shields during spacecraft integration and tests. The shields are required to attenuate the
Earth’s field 共⬃31 000 nT兲 so test measurements in the most sensitive range 共⫾8 nT兲 can be made.
mum, noise, zero level, scale factor, and dynamic re- 100 000 and a few arcsec, respectively. Figure 13 illustrates
sponse for each magnetometer dynamic range. some of the complexities associated with integration activi-
Depending on the complexity of the associated onboard ties that involve long magnetometer booms, in this case that
analog and digital processing units, many additional of the Voyager spacecraft. The large structures are not de-
functions and parameters are typically monitored and signed to be self-supporting in the Earth’s gravitational field
characterized. and relatively complex mechanical support equipment must
共3兲 Determination of the angular alignment of the sensors be used to deploy them.
with respect to an external reference coordinate system It is a well-known fact that the launch environment and
over the environment defined in 共2兲 and in controlled the transition to the mission phase environment always result
magnetic fields defined at the beginning of Sec. II. in minor shifts in magnetometer calibration and performance
共4兲 Determination, calibration or elimination of induced parameters except where absolute instruments are involved.
magnetization effects caused by the presence of compo- It therefore becomes necessary to apply in-flight calibration
nents with ferromagnetic materials of high relative per- techniques that have been developed to recover accurate es-
meability 共e.g., electromagnet cores兲. Although the ma-
timates of the true performance parameters for a given space-
terial may not contribute significantly to the stray field
based instrument. To this end, spacecraft attitude maneuvers
its permeability distorts the geometry of the ambient
constitute a powerful calibration and verification tool. They
field sensed by the magnetometer.
can be used effectively to establish effective zero levels and
The cost and complexity of the calibration and space- sensor alignment with respect to the reference axes.43,107
flight qualification activities for space magnetometers can Spinning spacecraft are particularly good for the estimation
match and exceed those of their development and manufac- of zero levels for the two magnetic field components in the
turing phase. For this reason ‘‘heritage’’ instruments that spin plane and the angular alignment of the component par-
have already demonstrated their capabilities and limitations allel to the spin axis. In many three-axis-stabilized spinning
over many space missions are usually preferred since many spacecraft, mechanical sensor ‘‘flippers’’ are used to rotate
of these tests can be simplified or eliminated. The most de- the fluxgate sensors through 90° or 180° and thus establish
manding calibration programs are those associated with ul- the effective zero level of the sensors 共see Fig. 14兲. This
trahigh accuracy instruments designed to map the Earth’s mechanical rotation is equivalent to inverting the leads in a
magnetic field77,79,81 where the absolute magnitude and an- bipolar voltmeter to determine its zero offset. Typical devices
gular accuracy required are of the order of 1–2 parts in used to develop the mechanical force needed to rotate the
ments are needed for the three-dimensional case. The non-

square inverse matrix 关 ⌬Bi 兴 ⫺1 is usually computed by
pseudoinverse techniques.
The least-squares estimation problem described above is
representative of a whole family of similar techniques that
have been developed to estimate parameters associated with
magnetic field measurements ‘‘in flight.’’ If highly accurate
models of the ambient field are available, as in the case of
the Earth, effective scale factors, angular alignment, and zero
biases are estimated routinely for many space missions.
Where interplanetary and magnetospheric missions are con-
cerned the estimation of effective zero levels usually consti-
tutes a major continuing calibration effort due to spacecraft
FIG. 14. Mariner Jupiter–Saturn 77 共Voyager兲 sensor mechanical flipper
effects and their variability. Scale factors are monitored
assembly used initially to establish zero offsets. Heated wax pellet actuators,
similar to automobile radiator thermostats, develop the force that causes the through built-in instrument calibration capabilities using step
sensor assembly to rotate 共flip兲 without generating a significant magnetic functions from precision current sources and absolute accu-
field. Radioisotope heater units 共RHUs兲 are used to maintain the sensor racy better than 0.1%–0.25% is seldom required. In some
temperature within its operating range.
instances, it is difficult to simulate zero-G conditions during
testing on the ground and boom alignment angles are not
assembly without generating a magnetic field include wax known to the desired accuracy 共Fig. 12兲. Lightweight, large
pellets or bimetal strips. For Earth-orbiting magnetometers diameter coils, usually wound around a parabolic antenna
many calibration parameters can be derived by comparing structure, have been used83 to generate calibration magnetic
the actual measurements to available models of the geomag- fields whose direction is known precisely and allow the de-
netic field49 and deriving least-squares fitted correction algo- termination of in-orbit alignment angles. A single coil allows
rithms. Magnetometers on interplanetary spacecraft do not the estimation of only two orientation angles and two coils
benefit from the availability of accurate models of the inter- mounted orthogonal to each other are required for estimation
planetary magnetic field which is extremely time varying. of the full set of alignment angles. However most long
However, many years of observation have shown that most booms exhibit preferential distortion about a single axis 共tor-
of this variability is in the angular direction and not in the sion or bending兲 and a single coil is usually satisfactory.
field magnitude. It is possible then to formulate a statistical
least-squares estimation algorithm for magnetometer zero
VIII. COORDINATE SYSTEMS
level errors. If we denote the true ambient magnetic field at
two instants of time as Ba1 and Ba2 and the field measured Generally, to carry out the research intended the mea-
by the magnetometer at those same times as Bm1 and Bm2 , surements must be expressed in a physical coordinate sys-
respectively, we can write tem, which is different from that of magnetic field
sensors.52,108 Therefore, coordinate transformations are re-
⌬B⫽Bm2 ⫺Bm1 , 共13兲
quired which must take into account not only the orientation
and, because of the assumptions discussed above, of the spacecraft in inertial space, but also the internal rota-
tions associated with booms and other instruments and sen-
0.5共 Ba1 ⫹Ba2 兲 "⌬B⫽0. 共14兲
sors aboard. A simplified general expression for the desired
The fields measured are given by magnetic field quantities described in a reference coordinate
system and starting with the magnetometer output voltages
Bm1 ⫽Ba1 ⫹Z and Bm2 ⫽Bm2 ⫹Z, 共15兲 in three channels 关V兴 can be given as
where Z denotes the ‘‘offset’’ vector due to bias in the mag- B⫽ 关 k兴关 V⫺Z兴关 Msensor兴关 Mboom兴关 Ms/c inert兴 , 共18兲
netometer electronics and stray fields generated by the space-
craft. We call this the ‘‘effective zero’’ of the magnetometer. where B is the measured magnetic field expressed in the
We can estimate the value of Z in a least-squares sense from coordinate system desired, Z is the effective zero level of the
a series of vector measurements (i⫽1,2,3,....,n) where the magnetometer including the A/D converter and spacecraft
ambient magnetic field has primarily changed direction and contributions, k is a vector whose components are the mag-
perhaps magnitude ‘‘adiabatically’’ as follows: netometer scale factors and A/D converter factors for each
axis, Msensor is a nonorthogonal matrix that transforms the
0.5共 Bmi ⫹Bm 共 i⫹1 兲兲 "⌬Bi ⫽Z"⌬Bi , 共16兲 measurements obtained in the physical sensor system to an
orthogonal system defined in the sensor frame, Mboom is an
Z⫽ 关 0.5共 Bmi ⫹Bm 共 i⫹1 兲兲 ⌬Bi 兴关 ⌬Bi 兴 ⫺1 , i⫽1,2,3,...,n, orthogonal matrix that transforms the measurements from the
共17兲
sensor frame to the frame of the body of the spacecraft where
where the matrices shown have been computed from the time inertial references are located, and Ms/c inert is an orthogonal
series of data. A minimum of three data points is needed to matrix that relates the spacecraft inertial reference frame to
estimate Z in two dimensions, such as the offsets in the spin the reference frame desired. It is clear that the ultimate ac-
plane for a spinning spacecraft, while four or more measure- curacy achievable in magnetic field measurements depends
strongly on precise knowledge of these parameters, many of sarily designed with sensitive magnetics performance re-
which are closely related to system level performance char- quirements in mind. The current stock of high performance
acteristics. Scalar measurements are obviously insensitive to grade, fluxgate sensor materials will be exhausted in the near
coordinate transformation issues and are concerned only with future, just when spacecraft constellation programs involving
spacecraft magnetic contamination levels. perhaps hundreds of spacecraft 共Magnetospheric Multi-
Scale, SWARM, ANTS, etc.兲 are supposed to be coming on
line. Unless a significant investment is made to recover pro-
IX. THE FUTURE duction capabilities for high-grade Permalloys, the perfor-
The preceding discussion has illustrated the significance mance of magnetometers aboard these missions will fall far
of space-based magnetic field measurements and their rel- short of the requirements for ‘‘world-class’’ science and en-
evance, not only for space physics research and planetary gineering.
exploration, but also for the emerging field of space weather
prediction and quantitative assessment of the geoeffective-
NOMENCLATURE
ness of solar-terrestrial events.53 Engineering applications in
low and geostationary orbit will continue to take advantage The following are space mission acronyms used in the
of the reliability, simplicity, and low cost of magnetometers text.
as attitude sensors and a control element for magnetic torqu- ACE—Advanced Composition Explorer, NASA Explorer
ing systems. Fluxgate magnetometers are expected to con- mission launched in August of 1997 to measure and compare
tinue to dominate the field although recent improvements in the elemental and isotopic composition of the solar corona
magnetoresistance technology will certainly have an impact and to monitor the interplanetary medium.
on high volume applications such as spacecraft constellations ANTS—Autonomous Nano-Technology Swarm, a fleet of
used for communications from Earth’s orbit where absolute miniature spacecraft that would cruise independently to the
accuracy may not be a primary requirement. In the area of Asteroid Belt. Each probe would hoist its own solar sail to
space research applications, a significant shortage of low capture the minute pressure of the Sun’s rays.
noise, high permeability materials, such as those in the Per- BEPPI COLOMBO—A joint mission by the European Space
malloy family, is a source of concern. The widespread use of Agency 共ESA兲 and Japan’s Institute for Space and Astronau-
ferrites as magnetic core materials in industrial and commer- tical Science 共ISAS兲 to explore the planet Mercury. It will be
cial applications, with their significant cost and performance launched in the near future.
advantages, has caused the almost total disappearance of CASSINI mission—This spacecraft was launched by NASA
U.S. manufacturers of tape-wound cores and specialty in 1997 to explore the Saturn system beginning in 2004.
nickel–iron–molybdenum alloys used in high performance CLUSTER II—An ESA mission consisting of four identical
fluxgate sensors. The increasing use of amorphous metallic spacecraft launched in the summer of 2000 to explore near-
glasses to replace Permalloys49,109,110 is a compromise since Earth space.
their lower Curie temperature 共⬍300 vs ⬎600 °C兲 implies DEEP SPACE 1—A NASA New Millenium Mission
poorer temperature stability. Specialized temperature anneal- launched in October 1998 to demonstrate new space tech-
ing of these alloys to improve noise performance does not nologies in propulsion and guidance.
always result in better temperature stability. The market for GALILEO—The GALILEO spacecraft was launched by
research grade instruments is extremely limited and it is not NASA in October of 1989 to carry out a 2 year detailed
expected that recovery of the high-grade Permalloy produc- exploration of Jupiter, its moons, and environment.
tion capabilities of the 1970s will take place any time soon. GEOTAIL—The GEOTAIL mission is a collaborative
Limited high-grade Permalloy production capabilities exist project by Japan’s ISAS and the National Aeronautics and
in Europe and Japan but their future is probably just as un- Space Administration 共NASA兲. Its primary objective is to
certain as it is in the US. study the dynamics of the Earth’s magnetotail over a wide
The current trend of spacecraft miniaturization consti- range of distances. It is also part of the International Solar
tutes another significant problem for sensitive, research- Terrestrial Physics Program.
grade, magnetic field measurements. In addition to the sig- GIOTTO—An ESA mission named after the Italian artist
nificant decrease in signal-to-noise-ratio resulting from Giotto and launched in July of 1985 to encounter the comet
reduced sensor dimensions, as the spacecraft size is reduced Halley.
the distance from the sensor to the sources of unwanted HELIOS—A joint NASA–Federal Republic of Germany
fields is also reduced, and their effect amplified as ⬃(r 0 /r) 3 project consisting of two spacecraft 共Helios 1 and Helios 2兲
where r 0 denotes the reference distance and r the reduced launched in 1974 and 1976 to explore interplanetary space
distance to the source. The overall reduction in spacecraft close to the Sun 共⬃0.4 AU兲.
power consumption with miniaturization is not sufficient at INTERPLANETARY MONITORING PLATFORM 共IMP兲—
present to compensate for the deleterious effects of the prox- IMP-8 was launched by NASA on October 26, 1973 to mea-
imity of the sources. In principle, custom spacecraft and sub- sure the magnetic fields, plasmas, and energetic charged par-
system designs that take into account magnetic field mea- ticles of the Earth’s magnetotail and magnetosheath and of
surement objectives could result in acceptable performance the near-Earth solar wind. IMP-8 was the last of 10 IMP
but are ruled out by cost considerations. Low cost, spacecraft launched over 10 years.
commercial-off-the-shelf 共COTS兲 subsystems are not neces- LIVING WITH A STAR 共LWS兲—This is a space-weather
focused, applications-driven NASA research program. Its WIND and POLAR Missions—The WIND and POLAR
goal is to develop scientific understanding of those aspects spacecraft are part of the ISTP to study Solar-Terrestrial
connected with the Sun–Earth system that directly affect life Physics. They were launched in 1994 and 1996, respectively.
and society.
LUNAR PROSPECTOR—A NASA Discovery mission 1
Sh. Sh. Dolginov and N. V. Pushkov, Dokl. Akad. Nauk USSR 129, 77
launched in January of 1998 to orbit the Moon and map its 共1959兲.
2
surface composition and polar ice deposits, and measure Sh. Sh. Dolginov, L. N. Zhuzgov, and N. V. Pushkov, Artificial Earth
magnetic and gravity fields. Satellites 2, 63 共1959兲.
3
Sh. Sh. Dolginov, L. N. Zhuzgov, and V. A. Selyutin, Iskusstvennyye
MAGNETOSPHERIC MULTI-SCALE 共MMS兲—A fleet of Sputniki Zemli 4, 135 共1960兲.
four spacecraft similar to CLUSTER designed to study the 4
C. P. Sonnet, J. Geophys. Res. 68, 1229 共1963兲.
Earth’s magnetosphere. This mission is part of the NASA 5
J. P. Heppner et al., Space Res. 1, 982 共1960兲.
6
L. J. Cahill, Space Physics 共1964兲, pp. 301–349.
Solar Terrestrial Probes program. 7
N. F. Ness, J. Geophys. Res. 70, 2989 共1965兲.
MAGSAT—The Magsat project was a joint NASA/U.S. 8
N. F. Ness, Space Sci. Rev. 11, 459 共1970兲.
Geological Survey 共USGS兲 effort to map near-Earth mag- 9
N. F. Ness et al., Science 233, 85 共1986兲.
netic fields on a global basis. The spacecraft was launched in
10
N. F. Ness et al., Science 246, 1473 共1989兲.
11
J. E. P. Connerney, M. H. Acuña, and N. F. Ness, J. Geophys. Res., 关Solid
October 1979 and re-entered the Earth’s atmosphere six Earth Planets兴 92, 15234 共1987兲.
months later. 12
J. E. P. Connerney, M. H. Acuña, and N. F. Ness, J. Geophys. Res., 关Solid
MARINER missions—A series of NASA missions launched Earth Planets兴 96, 19023 共1991兲.
13
since the early 1960s designed to explore the interplanetary C. T. Russell, R. C. Elphic, and J. A. Slavin, J. Geophys. Res. 85, 8319
共1980兲.
medium and the terrestrial planets. The last spacecraft in the 14
N. F. Ness, Annu. Rev. Earth Planet Sci. 7, 249 共1979兲.
series were the twin Mariner Jupiter–Saturn probes launched 15
S. C. Solomon, New Sci. 165, 2223 共2000兲.
in 1977 later renamed Voyagers 1 and 2 共see below兲. 16
17
D. MacKenzie, New Sci. 168, 2261 共2000兲.
Sh. Sh. Dolginov, L. N. Zhuzgov, and S. I. Shkolnikova, Kosm. Issled. 22,
MARS GLOBAL SURVEYOR—A NASA Mars exploration
792 共1984兲.
mission launched in 1996 to map the red planet and study is 18
J. G. Luhmann, Rev. Geophys. 29, 965 共1991兲.
topography, magnetic field, and gravity. 19
W. Riedler et al., Nature 共London兲 341, 604 共1989兲.
MESSENGER—A NASA Discovery mission to explore the
20
M. H. Acuña et al., Science 279, 1676 共1998兲.
21
M. H. Acuña et al., Science 284, 790 共1999兲.
planet Mercury to be launched in 2003. 22
J. E. P. Connerney et al., Science 284, 794 共1999兲.
NEAR-SHOEMAKER Mission—The Near-Earth Asteroid 23
A. B. Binder, Science 281, 1475 共1998兲.
Rendezvous Mission 共NEAR兲 was launched by NASA in 24
G. S. Hubbard et al., Acta Astronaut. 41, 585 共1997兲.
25
February of 1996 to rendezvous with the Asteroid Eros 433 G. S. Hubbard, A. B. Binder, and W. Feldman, IEEE Trans. Nucl. Sci. 45,
880 共1998兲.
and orbit this object for 1 year to conduct imaging, gravity, 26
R. P. Lin et al., EOS Trans. Am. Geophys. Union 55, 1181 共1974兲.
and magnetic field studies. 27
R. P. Lin et al., Science 281, 1480 共1998兲.
ØERSTED—The main purpose of the ØRSTED satellite is 28
J. E. McCoy et al., EOS Trans. Am. Geophys. Union 56, 1012 共1975兲.
29
K. A. Anderson et al., EOS Trans. Am. Geophys. Union 56, 1012 共1975兲.
to provide precise global mapping of the Earth’s magnetic 30
L. L. Hood et al., Phys. Earth Planet. Inter. 20, 291 共1979兲.
field. This small spacecraft was built by Denmark and 31
L. L. Hood et al., Geophys. Res. Lett. 26, 2327 共1999兲.
launched as a secondary payload in January 1999 aboard a 32
M. H. Acuña et al., J. Geophys. Res., 关Planets兴 97, 7799 共1992兲.
U.S. rocket.
33
M. H. Acuña et al., J. Geophys. Res., 关Planets兴 102, 23751 共1997兲.
34
A. Balogh et al., Astron. Astrophys., Suppl. Ser. 92, 221 共1992兲.
PIONEER missions—A series of NASA low cost missions 35
M. G. Kivelson et al., Space Sci. Rev. 60, 357 共1992兲.
designed to explore the interplanetary medium, our Sun’s 36
M. G. Kivelson et al., Science 261, 331 共1993兲.
environment, and the planets Jupiter, Saturn 共Pioneers 10 and 37
F. M. Neubauer et al., J. Phys. E 20, 714 共1987兲.
11兲, and Venus 共Pioneer Venus兲. The last Pioneer mission
38
I. Richter et al., Geophys. Res. Lett. 28, 1913 共2001兲.
39
C. T. Russell et al., IEEE Trans. Geosci. Remote Sens. 18, 32 共1980兲.
共13兲 was launched in 1978. 40
E. J. Smith, Adv. Space Res. 13, 5 共1993兲.
SOLAR TERRESTRIAL PROBES—A six mission NASA 41
J. E. Smith and C. P. Sonnett, IEEE Trans. Geosci. Electron. GE-14, 154
program to study the Earth’s magnetosphere, the atmosphere/ 共1976兲.
ionosphere system, the Sun, and the interplanetary medium.
42
M. H. Acuña et al., Icarus 155, 220 共2002兲.
43
B. J. Anderson et al., IEEE Trans. Geosci. Remote Sens. 39, 907 共2001兲.
SWARM—SWARM will be placed into orbit around asteroid 44
D. A. Lohr et al., Johns Hopkins APL Tech. Dig. 19, 136 共1998兲.
Ceres. From this vantage point, its instrumentation will be 45
R. P. Lepping et al., Space Sci. Rev. 71, 207 共1995兲.
able to study the asteroid at close range. Its primary mission 46
C. W. Smith et al., Space Sci. Rev. 86, 613 共1998兲.
47
will be to observe Ceres and transmit its observations back to R. C. Snare, in Measurement Techniques in Space Plasmas, edited by R.
Pfaff, Geophysical Monograph No. 103 共American Geophysical Union,
Earth. 1998兲, pp. 101–115.
ULYSSES—A joint NASA–ESA mission named after the 48
J. S. Eterno, in Space Mission Analysis and Design, 3rd ed., edited by J. R.
Greek hero to chart the unexplored regions above and below Wertz and W. L. Larson 共Kluwer Academic–Microcosm Press, 1999兲.
49
P. Ripka, Magnetic Sensors and Magnetometers 共Artech House, Dedham,
the Sun and the ecliptic plane. It was launched in October of
MA, 2001兲.
1990 from the Space Shuttle. 50
M. Shuster, in Fundamentals of Space Systems, edited by V. L. Pisacane
VOYAGER 共Voyagers 1 and 2兲—The twin Voyager space- and R. Moore 共Oxford University Press, Oxford, 1994兲.
craft were launched by NASA in 1977 共as Mariner Jupiter–
51
M. H. Acuña et al., Space Sci. Rev. 71, 5 共1995兲.
52
W. H. Mish et al., Space Sci. Rev. 71, 815 共1995兲.
Saturn probes兲 to explore the planets Jupiter and Saturn. The 53
D. N. Baker, Solar-Terrestrial Relations: Predicting The Effects On The
mission was later expanded to encompass the exploration of Near-Earth Environment 22, 7 共1998兲.
all the outer planets with the exception of Pluto. 54
N. C. Maynard, Rev. Geophys. 33, 547 共1995兲.
55
M. A. Shea and D. F. Smart, Solar-Terrestrial Relations: Predicting The 84
O. V. Nielsen et al., Sens. Actuators A 59, 168 共1997兲.
Effects on The Near-Earth Environment 22, 29 共1998兲. 85
F. Primdahl and P. Jensen, J. Phys. E 15, 221 共1982兲.
56
D. F. Webb, IEEE Trans. Plasma Sci. 28, 1795 共2000兲. 86
C. P. Escoubet, C. T. Russell, and R. Schmidt, Space Sci. Rev. 79 共1997兲.
57
R. Pirjola et al., Phys. Chem. Earth 25, 333 共2000兲. 87
W. A. Geyger, Non-Linear Magnetic Control Devices 共McGraw-Hill, New
58
D. N. Baker, IEEE Trans. Plasma Sci. 28, 2007 共2000兲. York, 1964兲.
59
D. N. Baker and J. G. Kappenman, Science 273, 168 共1996兲. 88
R. L. Mcpherron, P. J. Coleman, and R. C. Snare, IEEE Trans. Aerosp.
60
J. G. Kappenman and V. D. Albertson, IEEE Spectrum 27, 27 共1990兲. Electron. Syst. 11, 1110 共1975兲.
61
V. D. Albertson et al., IEEE Trans. Power Deliv. 8, 1206 共1993兲. 89
T. A. Potemra, L. J. Zanetti, and M. H. Acuña, IEEE Trans. Geosci. Re-
62
A. Balogh et al., Space Sci. Rev. 79, 65 共1997兲. mote Sens. 23, 246 共1985兲.
63
S. Kokubun et al., J. Geomagn. Geoelectr. 46, 7 共1994兲. 90
C. T. Russell, IEEE Trans. Geosci. Electron. E-16, 239 共1978兲.
64
M. W. Dunlop et al., Planet. Space Sci. 47, 1389 共1999兲. 91
C. T. Russell et al., Space Sci. Rev. 71, 563 共1995兲.
65
D. Duret et al., IEEE Trans. Magn. 31, 3197 共1995兲. 92
R. C. Snare and J. D. Means, IEEE Trans. Magn. 13, 1107 共1977兲.
66
D. Duret et al., IEEE Trans. Magn. 32, 4935 共1996兲. 93
D. J. Southwood, W. A. C. Mierjedrzejowicz, and C. T. Russell, IEEE
67
H. Aschenbrenner and G. Goubau, Hochfreq Tech. Electroakust 47, 178
Trans. Geosci. Remote Sens. 23, 301 共1985兲.
共1936兲.
68
94
L. Zanetti et al., Space Sci. Rev. 70, 465 共1994兲.
J. Nickel, Report No. HPL-95-60, 1995. 95
F. Primdahl et al., J. Phys. E 22, 1004 共1989兲.
69
F. Primdahl, J. Phys. E 12, 241 共1979兲. 96
F. Primdahl et al., Meas. Sci. Technol. 2, 1039 共1991兲.
70
N. Kernevez and H. Glenat, IEEE Trans. Magn. 27, 5402 共1991兲. 97
E. B. Pedersen et al., Meas. Sci. Technol. 10, N124 共1999兲.
71
R. E. Slocum and F. N. Reilly, IEEE Trans. Nucl. Sci. NS-10, 165 共1963兲.
72
R. E. Slocum, P. C. Cabiness, and S. L. Blevins, J. Phys. E 42, 763 共1971兲.
98
J. Piilhenriksen et al., Meas. Sci. Technol. 7, 897 共1996兲.
73
R. E. Slocum, Phys. Rev. Lett. 29, 1642 共1972兲.
99
F. Primdahl et al., Meas. Sci. Technol. 5, 359 共1994兲.
74
W. H. Farthing and W. C. Folz, Rev. Sci. Instrum. 38, 1023 共1967兲.
100
P. Ripka and F. Primdahl, Sens. Actuators A 82, 161 共2000兲.
75
A. M. A. Frandsen et al., IEEE Trans. Geosci. Remote Sens. 16, 195
101
P. Ripka et al., Sens. Actuators A 46, 307 共1995兲.
共1978兲.
102
F. Primdahl and P. A. Jensen, J. Phys. E 20, 637 共1987兲.
76
M. H. Acuña, IEEE Trans. Magn. MAG-10, 519 共1974兲.
103
N. F. Ness et al., J. Geophys. Res. 76, 3564 共1971兲.
104
77
M. H. Acuña et al., NASA Rep. No. TM-79656, 1978. R. P. Lepping, K. W. Behannon, and N. F. Ness, EOS Trans. Am. Geo-
78
H. Luhr et al., IEEE Trans. Geosci. Remote Sens. 23, 259 共1985兲. phys. Union 55, 410 共1974兲.
79
O. V. Nielsen et al., Meas. Sci. Technol. 6, 1099 共1995兲.
105
C. F. Hall, Science 188, 445 共1975兲.
80
W. Magnes et al., Meas. Sci. Technol. 9, 1219 共1998兲.
106
M. H. Acuña, Radio Electron. Eng. 52, 431 共1982兲.
81
J. M. G. Merayo et al., Meas. Sci. Technol. 11, 120 共2000兲.
107
W. M. Farrell et al., IEEE Trans. Magn. 31, 966 共1995兲.
82
R. L. McPherron and R. C. Snare, IEEE Trans. Geosci. Remote Sens. 16, 108
C. T. Russell, Cosm. Electrodyn. 2, 184 共1971兲.
134 共1978兲. 109
B. B. Narod et al., Can. J. Phys. 63, 1468 共1985兲.
83
K. W. Behannon et al., Space Sci. Rev. 21, 235 共1977兲. 110
B. B. Narod and J. R. Bennest, Phys. Earth Planet. Inter. 59, 23 共1990兲.

Dual Magnetometer Space

Uploaded by

Copyright:

Available Formats

Dual Magnetometer Space

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dual Magnetometer Space

Uploaded by

Copyright:

Available Formats

REVIEW OF SCIENTIFIC INSTRUMENTS VOLUME 73, NUMBER 11 NOVEMBER 2002

I. INTRODUCTION Marco Polo is credited with taking this information to Eu-

cause depopulation of this energy level, it will become

exceed 1 kg and that of the electronics 2–3 kg. Helium,

ing terminals. The presence of an external magnetic field will

Bs/c1 ⫽ 共 B1 ⫺B2 兲 / 共 1⫺ ␣ 兲 , 共8兲

Bs/c2 ⫽ ␣ Bs/c1 , 共11兲

amplifiers 共TWTAs兲, which were mounted on the back side

ments are needed for the three-dimensional case. The non-

You might also like