Fluxgate effect in twisted magnetic wire

Fluxgate effect in twisted magnetic wire
Journal of Magnetism and Magnetic Materials 00 (2007) 000–000
Fluxgate effect in twisted magnetic wire
M.Buttaa1, P. Ripkaa, S. Atalayb, F. E. Atalayb, X.P. Lic
Dept. of Measurement, Czech Technical University, Technická 2, 166 27 Praha, Czech Republic
Inonu University, Science and Arts Faculty, Physics Department 44069-Malatya, Turkey
National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260
Elsevier use only: Received date here; revised date here; accepted date here
In this paper a novel kind of fluxgate is presented. The sensor is based on helical anisotropy of the ferromagnetic layer, electrodeposited on cupper wire.
The saturating field is provided by current flowing in the wire, and the output voltage is measured directly at the terminations of the wire. Therefore no
coils are necessary, making possible high miniaturization of the sensor. The effect has been tested twisting the wire: the second harmonic is shown to be
strongly dependent on the applied torsion. Consideration on the practical use of the sensors are finally presented.
© 2007 Elsevier B.V. All rights reserved
PACS: Type pacs here, separated by semicolons ;
Keywords: coil-less fluxgate, electrodeposited wires, helical anisotropy, twisting effect.
Fluxgates are the most precise vectorial magnetic field
sensors. They are used when field resolution better than 1
nT is required and/or in systems which require high
precision. Linearity as high as units of ppm is necessary, if
the magnetometer is used to detect small disturbances in
large background DC magnetic field.
The basic principle of fluxgate sensor is unchanged from
1930’s. The excitation field should bring the whole sensor
core volume into deep saturation. Only then the sensor
output is stable without remanence (also called perming
effect) [1]. This makes fluxgates preferable to other sensors
such as GMI, which do not saturate the ferromagnetic core.
The hysteresis curve of fluxgate sensor and the excitation
waveform (usually squarewave voltage) should be very
symmetrical – then in zero measured field the output voltage
contains only odd harmonics. The measured DC field causes
non-symmetry of the magnetization and creates even
harmonics. The amplitude of the 2nd harmonic voltage for
required 100 pT resolution is typically 160 dB below the
voltage at the 1st harmonics. This makes unacceptable
requirements on the dynamic reserve of the electronics.
There are two ways how to suppress the large unwanted
Corresponding author. Tel.: +420-224-352-178; fax: +420-233-339-929.
E-mail address: [email protected], [email protected]
signal geometrically: either to use double-core sensors or
make the excitation perpendicular to the sensing coil.
The orthogonal fluxgate usually works at 2nd harmonics. A
“fundamental-mode orthogonal fluxgate” [2,3] should use
large dc current bias, which makes the sensor unpractical.
For practical applications which require highest
miniaturization the main disadvantage of fluxgate sensors is
the presence of coils. A first step to simplify the structure
was done with orthogonal type fluxgate: the excitation
current can flow through the core, so that no excitation coil
is required. Unfortunately a pick-up coil is still necessary
for the detection of the second harmonic component of the
core flux.
In this paper we present a novel type of fluxgate which
does not require any coil, making its practical application
much easier.
The core used for this coil-less fluxgate is a copper wire
covered by a layer of magnetic alloy, which has been
already used for conventional orthogonal fluxgate [4]. We
found that a second harmonic component of the voltage
between the terminals of the wire depends on the axial
magnetic field. Using this voltage as an output, the fluxgate
needs no coil at all.
Author name / Journal of Magnetism and Magnetic Materials 00 (2007) 000–000
By a series of measurements with two types of twisted wires
we show that this new effect is caused by helical anisotropy.
The conditions necessary for the proper operation of the
sensor and a feasibility of practical applications are
2.Sensor structure
The first sensor is composed by a 50 µm diameter copper
wire covered by a layer of electrodeposited ferromagnetic
material. The wires have been produced by S. Atalay and
F.E. Atalay, and it has been fully characterized in [5]. The
ferromagnetic layer is 10 µm thick polycrystalline
Co18.97Ni49.60Fe31.43 . The total length of the wire is 3.8 cm.
During characterization of the sensor we observed
unsymmetrical circular hysteresis curves in the presence of
DC axial field (Fig. 1). This was indication of a strong offdiagonal component of the permeability tensor [6].
allowed us to apply a twisting angle to the wire. One of the
two terminations of the twisting device was mounted on a
sleight: in this way we could adjust the distance between the
termination to fit the wire length. The wire was suspended
and was in contact only to the termination of the twisting
device, therefore the only mechanical stress applied on the
wire was the torque produced by the twisting device at its
We injected an alternating current Iwire (f = 30 kHz) into
the wire and we measured the voltage Uwire between its
terminations, while applying a twisting torque (Fig. 2).
Fig. 2 – Structure of the sensor.
The current Iwire saturates the ferromagnetic layer in
circular direction. Iwire is provided by a sinewave generator
whose internal resistance was 50 Ω, approximately 100
times bigger than the wire’s impedance. Therefore the
waveform generator acted as a sinewave current source.
Fig. 3 shows the dependence of the 2nd harmonic of Uwire,
measured by the a SR 760 DSP lock-in amplifier, on the
external field Bext applied by the Helmholtz coil.
The dependence shows a wide linear part in the range
between -100 µT to 100 µT, making the sensor suitable for
measurement of magnetic field in that range.
Fig. 1 B-H circumferential loops for axial dc field of +250 µT and -250
µT. Excitation current 24 mA, 10 kHz. Twisting angle 20 degrees.
One of the possible sources of this component is helical
stress-induced anisotropy. In order to examine the effect, we
intentionally twisted the wire in both directions. The wire
has been soldered at the ends to a twisting device which
Fig. 3 – 2nd harmonic of Uwire [V] vs. external magnetic field B ext [µT],
Iwire=55 mA, 30kHz, twisting angle 30º.
We can therefore define a new kind of orthogonal
fluxgate: in this sensor the 2nd harmonic, proportional to
external magnetic field, is not extracted from the voltage
Author name / Journal of Magnetism and Magnetic Materials 00 (2007) 000–000
induced in the pick-up coil, but directly from the voltage
Uwire measured at the terminations of the wire. We will call
this novel kind of sensor coil-less fluxgate.
The sensitivity of the coil-less fluxgate has been
measured for several values of current Iwire, while changing
the twisting angle. The resulting relation between the
sensitivity and the twisting angle is shown in Fig. 4.
Fig. 4 - Sensitivity of the coil-less fluxgate vs. twisting angle, Iwire=55mA,
30 kHz.
It is clear that the 2nd harmonic in Uwire is due to the
twisting torque applied to the wire. An experiment has been
realized, measuring the 2nd harmonic response for both
positive and negative twisting angles, without changing the
reference phase of the lock-in amplifier. It has been clearly
observed that inverting the direction of the twisting angle
makes also the 2nd harmonic response changing sign: for
negative angles we obtained 2nd harmonic dependence on
Bext with negative slope (Fig. 5). This result suggests that the
2nd harmonic in Uwire is caused by helical anisotropy of the
ferromagnetic layer, caused by mechanical stress associated
with twisting torque.
Isat. It has been observed that the twisting torque influences
also Isat. Fig. 6 shows the dependence of Isat on the twisting
angle. We can therefore derive that applying a torsion on the
wire makes the coil-less fluxgate working properly, giving
rise to a 2nd harmonic (which increases for higher twisting
angle), but it has also the drawback of increasing the
saturation current requirement.
Fig. 6 – Dependence of the saturation current Isat on the twisting angle,
3.Wire with thinner layer
We performed similar measurement on the similar wire,
provided by X.P. Li, whose ferromagnetic layer was much
thinner, typically 2 µm. They have shown a similar
behaviour: also in this case the 2nd harmonic increases when
we apply a torsion to the wire terminations. In this case the
sensitivity was ten times smaller, which is reasonable if we
consider the lower thickness of the magnetic layer.
The main advantage of using electrodeposited wires with
thinner ferromagnetic layer is that they require much
smaller saturation current. In this case we could achieve
saturation of the wire at 7-8 mA, 30 kHz, which could be
acceptable for practical applications.
It has been also observed that this kind of wires with
thinner ferromagnetic layer show a non-zero sensitivity even
without any applied torque. In this case we can deduct that
the helical anisotropy was built-in during the production of
the wires. In any case twisting the wire further increases the
4.Considerations on the practical use of the sensor
Fig. 5 – The sensor characteristics for twisting angle of -30, 0, and + 30
The saturation of the core in orthogonal direction is a
necessary condition for the achievement of a linear
characteristic of the coil-less fluxgate. If Iwire is lower than
the saturation current Isat the 2nd harmonic response is
affected by perming effect, making the sensor useless.
Therefore it is crucial to use a current amplitude higher than
It is worth to be highlighted, that the field-dependent 2 nd
harmonic is only one component of the total voltage Uwire. In
order to evaluate the feasibility of 2 nd harmonic extraction
we applied a constant field B = 125 µT to the coil-less
fluxgate and we evaluated the ratio of the 2nd harmonic
amplitude over the total voltage Uwire peak-peak. We chose
the peak-peak amplitude because we are interested in the
maximum gain we can use while amplifying the signal
without saturating the electronics; in this case the peak-peak
value is much more significant than the RMS value,
Author name / Journal of Magnetism and Magnetic Materials 00 (2007) 000–000
especially taking into account that Uwire is not sinusoidal at
all (the ferromagnetic layer is saturated) – Fig. 7.
wire. The exclusion of the connection cables from the
measurement of the voltage reduces the resistive component
of the Uwire. This gives a lower peak-peak Uwire amplitude
while does not affect the 2nd harmonic, which only depends
on the magnetic properties of the alloy layer. It should be
advisable to design the next generation of coil-less fluxgate
using a core whose material has the lowest possible contact
resistance, in order to maximize the performance of the
We must consider that the results shown in this paper
have been achieved from measurement performed twisting
the wires as low as possible; typically we measured all the
parameters for each position, then we changed twisting
angle. After several times we twisted the wire in alternating
directions, it irreversibly changed its parameters probably
due to the plastic deformation and changes in its structure.
Fig. 7 – Waveforms for B = 250 µT: excitation current [A] (upper trace),
output voltage [V] (lower trace). Excitation current 24 mA, 10 kHz.
Twisting angle 20 deg.
In Fig. 8 we can see the dependence of the 2nd harmonic
at 125 µT, in percent of Uwire p-p on the twisting angle
(Iwire=55mA, 30kHz, thicker wire).The 2nd harmonic
increases while increasing the torsion, as we have shown
previously. Indeed the p-p amplitude of Uwire only slightly
decreases, twisting the wire.
We can observe from Fig. 8 how we can easily achieve a
signal whose 2nd harmonic is ≈3% of peak-peak amplitude.
Such a level of 2nd harmonic is not too demanding, since it
can be extracted using usual techniques.
harm / Uwire pk-pk [%]
Angle [degrees]
Fig. 8 – 2nd harmonic of Uwire in % of Uwire peak-peak, vs. twisting angle.
Excitation current 55 mA, 30 kHz. Bext 125 µT.
Similar calculation, performed on the wire with thinner
alloy layer, gave as a result a lower percentage of the 2 nd
harmonic in Uwire. Nevertheless, a 1% 2nd harmonic can still
be achieved even with this thinner wire.
It should be noticed that the best performance can be
obtained using four terminals connection for the sensing
In this paper we have presented a new kind of fluxgate
realized with electrodeposited ferromagnetic wires. We have
proved that the second harmonic of the voltage at wire’s
termination depends on the torsion applied to the wire.
The sensitivity and the linear range achieved with this
sensors are good enough to consider it feasible for a wide
range of practical applications. If we also consider that lack
of coil makes possible high miniaturization, the
development of this sensors seems to be really promising.
Next step will be the development of coil-less fluxgate
with shorter wire (≈ mm) and built-in helical anisotropy.
[1] Ripka, P. : Fluxgate sensors, in: Magnetic sensors, Artech 2001
[2] Goleman K. and I. Sasada: High Sensitive Orthogonal Fluxgate
Magnetometer Using a Metglas Ribbon IEEE Trans. Magn. 42, NO.
10, 2006, pp. 3276-8
[3] Plotkin A., E. Paperno, A. Samohin, and I. Sasada Compensation of the
thermal drift in the sensitivity of fundamental-mode orthogonal
fluxgates, J. Appl. Phys. 99, 08B305 (2006)
[4] Fan J, Li XP, Ripka P: Low power orthogonal fluxgate sensor with
electroplated Ni80Fe20/Cu wire, Journal of Applied Physics 99 (8):
Art. No. 08B311 APR 15 2006
[5] P. Ripka, M. Butta, M. Malatek, S. Atalay, F. E. Atalay:
Characterisation of magnetic wires for fluxgate cores, to appear in
Proc. of the Transducers/Eurosensors conference, 2007
[6] Kraus L., Kane S. N., Vazquez M., Rivero G., Fraga E., Hernando A.:
Tensor component of the magnetization in twisted Fe-rich amorphous
wire, J. Appl. Phys. 75, 6952-6954, 1994
Research supported by the research program No. MSM6840770015
"Research of Methods and Systems for Measurement of Physical Quantities
and Measured Data Processing " of the CTU in Prague sponsored by the
Ministry of Education, Youth and Sports of the Czech Republic.
Patent request pending: application number PUV 2007-18662.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF