Measurement 43 (2010) 46–53

Contents lists available at ScienceDirect

Measurement

journal homepage: www.elsevier .com/ locate /measurement

A 2D micro-fluxgate earth magnetic field measurement systemswith fully automated acquisition setup

Andrea Baschirotto a, Enrico Dallago b, Massimo Ferri b,*, Piero Malcovati b, Andrea Rossini b,Giuseppe Venchi b

a Department of Physics, University of Milano Bicocca, Piazza Scienza, 20126 Milano, Italyb Department of Electrical Engineering, University of Pavia, Via Ferrata 1, 27100 Pavia, Italy

a r t i c l e i n f o

Article history:Received 15 April 2009Received in revised form 11 June 2009Accepted 15 June 2009Available online 21 June 2009

Keywords:Fluxgate sensorElectronic compassEarth magnetic fieldFerromagnetic core

0263-2241/$ - see front matter � 2009 Elsevier Ltddoi:10.1016/j.measurement.2009.06.007

* Corresponding author. Tel.: +39 0 382985256; faE-mail addresses: andrea.baschirotto@unimib

enrico.dallago@unipv.it (E. Dallago), massimo.ferripiero.malcovati@unipv.it (P. Malcovati), andrea.rosssini), giuseppe.venchi@unipv.it (G. Venchi).

a b s t r a c t

In this paper we present an Earth magnetic field measurement system and an automatedacquisition setup to characterize it. The measurement system consists of a fluxgate sensorand an integrated front-end circuit, both realized in CMOS technology. The couple oforthogonal axes of the sensor makes the system suitable for realizing an electronic com-pass device. Indeed, we can measure not only the amplitude of the Earth magnetic field(whose full-scale value is of the order of 60 lT), but also its direction. The complete mea-surement system achieves a maximum angular error of 1.5� in the measurement of theEarth magnetic field direction. Furthermore, an acquisition setup was developed to evalu-ate the measurement system performance. It consists of a precision mechanical plasticstructure, in tower form, a microcontroller-based interface circuit, that provides a digitaloutput through an RS232 serial interface, a PC software suitably developed to post-processthe data from the acquisition system and a couple of Helmholtz coils to evaluate the line-arity of the system. This setup allows us to perform a completely automated and numeri-cally controlled characterization of the measurement system.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Magnetic sensors have been used since hundreds ofyears. At the beginnings they were available only asmechanical devices for navigation and orientation in openspaces. Recently, to detect a magnetic field, it is possible touse both mechanical and electronic sensors [1]. The mainadvantage of the electronic sensors, which have beenrecently developed, is that they can be integrated togetherwith electronic interface circuits in the data processing flow.This improves the embedding development trend, but intro-duces more complexity in the measurement setup design.

. All rights reserved.

x: +39 0 382985677..it (A. Baschirotto),@unipv.it (M. Ferri),ini@unipv.it (A. Ros-

There are several types of magnetic sensors, but, basi-cally, all of them, when detecting a magnetic field, showa small variation of a physical property or parameter ofthe device. The entity of this variation, which is relatedto the sensitivity of the sensor to the magnetic field ap-plied, makes the sensor itself suitable for a specific typeof application [2]. It is thus possible to classify the mag-netic sensors by using their field sensing range. As shownin Fig. 1, three categories can be identified: low field, med-ium field, and high field sensors. Magnetic fields lowerthan 1 lT are very small and well below the Earth mag-netic field. Sensors with field sensing range from 10 lT to300 lT are considered Earth magnetic field sensors, whilesensors with field sensing range above 1 mT are classifiedas bias magnet field sensors. For measuring the Earth mag-netic field with devices that are suitable for portable appli-cations, the magneto-resistance and magneto-inductance(to be used as discrete sensors) are available, as well as

Magnetic Sensor Technology

Classification of Magnetic Sensors

Detectable Field (Tesla)

10−14 −510 10 10 102

Search Coil Magnetometer

Fluxgate Magnetometer

Optically Pumped Magnetometer

Nuclear Precession Magnetometer

SQUID Magnetometer

Hall Effect Sensor

Magnetoresistive magnetometer

Magnetodiode

Magnetotransistor

Fiber Optic Magnetometer

Magneto Optical Sensor

Magnetoimpedence Magnetometer

−10 −2

Fig. 1. Classification of magnetic sensors.

A. Baschirotto et al. / Measurement 43 (2010) 46–53 47

the Fluxgate magnetic sensors. Fluxgate sensors and mag-neto-resistances require the use of a ferromagnetic mate-rial. They, as well as magneto-transistors and Hallsensors, can be integrated by using CMOS technologies[3–5]. The use of a ferromagnetic material as concentratorcan help in increasing their sensitivity. Fluxgate sensors,Hall sensors with magnetic concentrator and magneto-transistors allow the implementation of 2D measurementson-chip. By contrast, conventional Hall devices can be usedonly for 1D measurements. Each sensor has specific fea-tures that make it suitable for a given range of applications.In addition to the sensitivity, it is necessary to consider therange of temperature, the sensor volume, and its on-chipmanufacturability.

Nowadays, there are several different applicationswhere magnetic sensors can be used. Among them elec-tronic compasses, sensors for traffic control, magnet acti-vated switches for cellular phones, notebooks orhandheld devices can be mentioned. Other applicationsare in the automotive field or home appliances: devicesbased on magnetic sensors are used, for example, to con-trol the car engine or in domestic environment.

In this paper we present an Earth magnetic field mea-surement system and an automated acquisition setup tocharacterize it. The sensing element exploits a 2D planarCMOS integrated micro-fluxgate device [6,7] with inte-grated read-out circuit, which provides digital signals pro-portional to the components of the Earth magnetic fieldparallel to the chip plane and allows its magnitude anddirection to be estimated. A microcontroller-based inter-face circuit is used to program the CMOS read-out circuitvariable gain and signal multiplexing and to acquire thedigital word from the on-chip analog-to-digital converter(ADC) [8], making it available to a PC through a RS232 se-rial port.

A significant part of this work was devoted to the opti-mization of the experimental setup. The characterization ofan Earth magnetic field sensor is not a straightforwardtask: on one side, the linearity of each axis has to be eval-uated, while on the other, the angular accuracy given bythe relative orientation of the sensor with respect to themagnetic field has to be measured. This last performanceis somewhat related to the axis linearity, but it cannot bederived from it directly, since it is also dependent on thesignal-to-noise ratio of the read-out channels and on thestatistical fluctuations which affect the simultaneous read-ings of the two channels. To evaluate the angular accuracythe most effective approach is to rotate the sensor with re-spect to the magnetic field, acquire the read-out circuitoutput signal and reconstruct the angle. The accuracy ofsuch a characterization can be affected by several mea-surement error sources, one of the most critical being themechanical positioning of the device in the Earth magneticfield, which can impair the reliability and the repeatabilityof the measurements. To extract the information on themagnetic field from the measurement system, it is neces-sary to develop a complex acquisition system that includeselectronic, mechanical and software parts, interacting witheach other, as shown in Fig. 2.

In the paper, Section 2 presents the details of the inte-grated measurement system, Section 3 describes the com-pletely automated acquisition setup, while Section 4reports the experimental results obtained.

2. Magnetic field measurement system

The Earth magnetic field measurement system con-sists of 2D planar fluxgate magnetic sensor and an inte-grated read-out circuit, for exciting the fluxgate sensorand reading-out the magnetic field magnitude in digitaldomain.

2.1. Fluxgate sensor

When realized with integrated circuit technologies, thethree-dimensional geometry of a fluxgate sensor evolves ina planar structure [9], as shown in Fig. 3. In this case, theexcitation and sensing coils are implemented as spirals,realized with two different metal layers, while the mag-netic core is usually obtained with a post-processing ofthe silicon wafer. In Fig. 3 both magnetic axes are shownbut, for simplicity, only a pair of sensing coils are indicated.This structure is able to detect a magnetic field coplanarwith the structure itself, the output signal being propor-tional to the projections of the field along the directionsof the two cross arms of the magnetic core. The integratedmicro-fluxgate [7] used, whose photograph is shown inFig. 4, has been developed in a 0.5 lm CMOS process andthe ferromagnetic core is realized as a post-processing stepby dc-magnetron sputtering. The obtained core featuresthe good magnetic properties of the amorphous ferromag-netic material used as target (Vitrovac 6025 X), with a verysmall thickness (about 1 lm). The thickness was chosen asa compromise between the sensitivity of the device andthe power consumption (the thicker the core, the higher

Fig. 2. Measurement and acquisition system interaction.

Excitation Coil

Sensing CoilMagnetic Core

Fig. 3. Planar double axis fluxgate magnetic sensor.

48 A. Baschirotto et al. / Measurement 43 (2010) 46–53

is the current required to bring it into saturation). Accord-ing to the fluxgate sensor operating principle, when ex-cited, the device provides at the sensing coils, two signalswhose second harmonic spectral component is propor-tional to the amplitude of the external magnetic field inthe corresponding direction.

2.2. Integrated read-out circuit

The integrated read-out circuit [6] consists of threemain blocks: an excitation block to provide the requiredexcitation current to the fluxgate sensor, a read-out blockto process the sensor output and an A/D converter [8] to

translate the analog output of the read-out chain into thedigital domain. Fig. 5 shows the block diagram of the entirecircuit, while Fig. 6 shows a microphotograph of the fabri-cated chip. The circuit [6,10] is quite flexible and can copewith fluxgate sensors with different specifications, provid-ing the current necessary for their correct operation andreading-out the output voltage. This has two main conse-quences: first the excitation circuit output stage had tobe implemented with a high voltage technology, in orderto supply the required current (in the tens of milliampererange) into a wide range of coil resistances (with a worstcase of 280 X); second the read-out block needs to havea programmable gain to accommodate the various ampli-tudes of the sensor output signals.

The excitation circuit consists of two different blocks,with two different power supplies: the first one is thelow-voltage block, with a supply voltage equal to 3.3 V,while the second, realized with high-voltage transistors,uses a supply voltage up to 25 V. A linear, class-AB outputstage has been used in order to minimize the distortion ofthe excitation current, and allow the interface circuit to ex-cite sensors with different coil impedance. The first blockgenerates a square wave with a frequency equal to100 kHz and programmable output amplitude, which isthen integrated, in order to obtain a triangular waveform

Fig. 4. Fluxgate sensor microphotograph.

High VoltageStage (25 V)

Low VoltageStage (3.3 V)

Excitation

Read-Out

Fig. 5. Block diagram of the integrated read-out circuit.

A. Baschirotto et al. / Measurement 43 (2010) 46–53 49

centered around half of the 3.3 V supply voltage. The exci-tation of the sensor with a triangular current waveformrepresents a trade-off between the low-noise performanceof solutions based on sinusoidal excitation and the simple

implementation of solutions based on pulsed excitation[11]. The second block consists of a high voltage mirroredoperational amplifier with low-impedance output stage,which receives the triangular waveform at the input and,

Fig. 6. Microphotograph of the integrated read-out circuit.

50 A. Baschirotto et al. / Measurement 43 (2010) 46–53

through a resistive feedback, produces a triangular currentat the output. A mirrored amplifier allows us to achieve themaximum swing at the output terminal. The class-AB out-put stage of the amplifier is designed to provide all the cur-rent required by the sensor. A decoupling stage betweenthe low-voltage and the high-voltage blocks is necessaryto level-shift the triangular wave produced by the low-voltage block around half of the high-voltage powersupply.

In order to ensure proper timing for the excitation andread-out blocks, the whole circuit is driven by a clock at400 kHz. This clock is internally divided, by a cascade offlip-flops. The outputs of this timing circuit are two signals:a 100 kHz square wave signal with its complementary out-put, that is used to drive the excitation block, and a200 kHz square wave signal, used to drive the read-outblock and to realize the second harmonic demodulation,needed to measure the sensor output. By using a 400 kHzmaster clock a duty cycle of 50% on both the 100 kHzand the 200 kHz output waveforms can be ensured. A dutycycle different from 50%, indeed, could compromise thedemodulation of the signals produced by the sensing coilsand, therefore, it has to be avoided.

The two-channel sensor read-out circuit, shown inFig. 5, is able to amplify the differential outputs of the sens-ing coils and to process the resulting signal, as illustratedin Fig. 7. Each channel of the read-out circuit consists offour different blocks. The first block is a gain stage thatamplifies each of the two outputs of the sensing coils (V+

and V�) by a factor of ten. In the second block the differ-ence between the two outputs of the first block(Vþ1 and V�1 ) is amplified again by a factor of six ðVDÞand demodulated ðVMÞ, to translate the second and higherorder even harmonics, which contain information of themagnetic field, down to dc. In order to ensure a correctdemodulation of the sensor signal and to avoid problemsdue to the possible asynchronicity between the clock andthe output itself, a quadrature demodulation was imple-mented. Using this technique and adding together the con-tribution of the two orthogonal signals, it is possible to

avoid errors due to timing misalignments between theread-out clock and the output of the sensor. The demodu-lation of the signal is performed with the 200 kHz clockgenerated by the timing circuit. The third block is a secondorder Sallen-Key low-pass filter that removes all the highfrequency components resulting from the demodulationand returns a dc value that is proportional to the magneticfield. The difference between this output voltage and theanalog ground is then amplified with digitally programma-ble gain (from 1 to 100 with digital signals b1 and b2) in thelast block ðVOÞ. For all the blocks we used conventionaltwo-stage operational amplifiers. The dc output of theread-out chain is finally processed by a 13-bit incrementalADC, and delivered in digital form to the output interface.We used a single ADC with a multiplexer, driven by digitalsignal b0, for both the read-out channels.

3. Automated acquisition system

To guarantee repeatability and reliability of the ac-quired data, a fully automated acquisition system has beendeveloped [12]. The acquisition system is the integrationbetween mechanical and electronic subsystems. To makethe measurement process completely automated, a micro-controller-based interface circuit was developed, togetherwith a plastic rotating tower and a dedicated PC software.

3.1. Microcontroller-based interface circuit

The microcontroller (MCU) system controls the gain andthe signal multiplexing in the interface circuit, as well as thesynchronized precision mechanical structure for rotatingthe system. Moreover, it acquires the digital data providedby the ADC implemented on the read-out circuit chip. Final-ly, it implements the interfacing between the acquisitionsystem and a PC application specifically developed.The MCU is a PIC16F877 with a clock frequency of 40 MHz.The required 5 V supply voltage is generated on the interfacecircuit board. To control the gain and the signal multiplexingof the read-out circuit, five dedicated digital pins are used.Moreover, four additional digital pins are used to generatethe four phases required to control the stepper motor ofthe plastic rotating structure. Finally, 13 input digital pinsallow the acquisition in parallel mode of the output of theread-out circuit ADC. The firmware of the MCU has beendeveloped specifically for this application in high-levellanguage, without any performance loss.

3.2. Precision rotating plastic tower

In order to allow the characterization of the angularaccuracy of the measurement system, a precision mechan-ical tower has been realized. The entire structure is madeof plastic components, including the mechanical coupling,in order to avoid any perturbation of the Earth magneticfield. The only metal part is the stepper motor, which istherefore placed at 50 cm distance from the fluxgate mag-netic sensor. Such a mechanical system allows us tocontrol the angular positioning with 0.1� accuracy, thusensuring the repeatability of the measurements.

H+

No External Field External Field HExt,1 External Field HExt,2

H+ H+

tt tHExt,1

HExt,2

VM

tt t

tt t

VO

tt t

VD

VM

VO

VD

VM

VO

VD

tt t

V1+

1

V1+

1

V1+

1

tt t

V+V+

V+

Fig. 7. Operating principle of the integrated read-out circuit.

A. Baschirotto et al. / Measurement 43 (2010) 46–53 51

3.3. Acquisition system optimization

The proposed fully automated acquisition system is theresult of an optimization process, which started from amanual approach. In the very first acquisition setup, thesensor was mounted on a plastic disc, which was manuallyrotated upon a table with 5� reference marks. The outputsignal of the sensing coils was subtracted by means of asimple operational amplifier based circuit (because of cou-pling effects it was not possible to simply connect the sens-

ing coils in antiseries). The difference was furtheramplified with a gain of 100 and read-out with a spectrumanalyzer. In spite of the intrinsic sensitivity of the spec-trum analyzer, this approach lead to a maximum angularerror of about 4.5�. Partially this was caused by the manualrotation of the system and partially by the fact that, be-cause of the time required by the spectrum analyzer tomake a measurement, a single acquisition per positionwas performed. A first improvement was the introductionof the integrated read-out circuit [6,10], which provides a

0 50 100 150 200 250 300 350−2

−1.5

−1

−0.5

0

0.5

1

1.5

Angu

lar E

rror [

Deg

ree]

Angular Position [Degree]

Fig. 9. Angular accuracy achieved with the automated acquisitionsystem.

0.5

1

1.5

plitu

de

52 A. Baschirotto et al. / Measurement 43 (2010) 46–53

dc voltage directly proportional to the measured field: thissignal is available continuously and it is easier to performan average over a number of subsequent measurements.At this stage the internal average of a Keithley 2000 mul-timeter was used.

Finally, the proposed acquisition system was intro-duced, providing a number of benefits:

� the system has a high degree of integration, even in theauxiliary circuitry, helping in improving the signal-to-noise ratio;

� the chance of making errors while reading the data isstrongly reduced;

� the precision of the mechanical rotation is as high as0.1�;

� speed is maximized.

This last characteristic is relevant, since it allows us toincrease the number of averaged acquisitions for a givenposition and for a given total time required for a full rota-tion. Alternatively, the time required for the 360� rotationcan be minimized for a given number of averages, loweringthe probability of local magnetic perturbation during themeasurement (it is worth stressing the fact that the usedapproach measures the actual Earth magnetic field). Ingeneral, this automatic acquisition system allows mea-surements to be performed with up to 720 steps, leadingto a much finer angle discretization than the 5� used forthe manual rotation. The angular accuracy in the recon-structed position is then improved, as shown in Fig. 8. Allthe values of angular accuracy reported in Fig. 8 are calcu-lated by applying fixed calibration coefficients for correct-ing offset and gain differences between the two axes (foreach setup the coefficients are determined from one mea-surement and then used for any further measurements).

4. Experimental results

The entire acquisition system is fully automated. All themechanical, electronic and software components havebeen developed for this particular application. With thisacquisition system, we were able to push the angular accu-racy of the measurement system down to 1.5�, when rotat-ing the system in the Earth magnetic field, as shown inFig. 9. The angular accuracy achieved with the automatedacquisition system is limited only by the fluxgate sensorand the read-out circuit, thus allowing the actual perfor-mance of the devices to be evaluated.

5

4

3

2

1

0Manual

PositioningBench

Instrumentation

ManualPositioning

Integrated CircuitRead-Out

Manual PositioningSemi-Automated

Acquisition(with Averages)

Fully-AutomatedPositioning

andRead-Out

Angu

lar A

ccur

acy

[Deg

ree]

Fig. 8. Angular accuracy as a function of the acquisition system evolution.

Fig. 10 reports the data acquired from the two axes ofthe sensor during the complete rotation. The results re-ported are referred to a 24-point acquisition along 360�.

The linearity of the entire system in the range of ±60 lThas been evaluated acquiring the output of the sensorwhile varying the intensity of the magnetic field withHelmholtz coils. In order to ensure the reliability of the lin-earity measurement, the axis of the sensor under test hasbeen oriented in the direction perpendicular to the Earthmagnetic field, i.e. the sensor has been rotated in order toacquire the maximum and the minimum voltage output,and it has been stopped in the middle position. This guar-antees a negligible contribution of the Earth magnetic fieldto the field impressed with the Helmholtz coils. In thismeasurement, we achieved a maximum linearity error ofabout 3% of the full-scale, as shown in Fig. 11. The sensitiv-ity obtained is 11 LSB/lT that corresponds to 0.45 mV/lT,considering a 300-mV ADC input voltage swing. All thedata collected are in agreement with the performance of

0 50 100 150 200 250 300 350−1.5

−1

−0.5

0

Angular Position [Degree]

Nor

mal

ized

Am

Fig. 10. Data acquired from the sensor over 360� with the automatedacquisition system.

800

600

400

200

0

−200

−20−40−60

−800

−600

−400

−1000−80

Dig

ital O

utpu

t Cod

e

0 20 40 60 80Magnetic Induction [µT]

Fig. 11. Linearity of the complete system.

A. Baschirotto et al. / Measurement 43 (2010) 46–53 53

the sensor stand-alone, previously measured with dedi-cated test equipment [6].

5. Conclusions

In this paper we presented a fully automated acquisi-tion system for 2D micro-fluxgate based Earth magneticfield measurement. The measurement system consists ofa micro-fluxgate sensor with an integrated read-out cir-cuit, both realized in CMOS technology, thus allowing theintegration on the same chip. The measurement system al-lows the detection of the direction of the Earth magneticfield and, hence, the implementation of an electronic com-pass, with an angular accuracy better than 1.5�. In order toobtain reliable and repeatable results, we developed a fullyautomated acquisition system, consisting of a microcon-troller-based interface circuit, a plastic rotating towerand a dedicated PC software. The experimental results con-firm the significant improvement in performance obtainedby introducing the automated acquisition system.

Acknowledgments

This works was partially supported by PRIN Project2005091051. The authors thank Cedrat, Grenoble France,

for allowing us to use of their FEM simulator and Vacuum-schmelze, Hanau Germany, for Vitrovac� samples.

References

[1] W. Gopel, J. Hesse, J.N. Zemel, Magnetic sensors, in: R. Boll, K.J.Overshott (Eds.), Sensors: A Comprehensive Survey, Wiley-VCH,1989.

[2] P. Ripka, P. Navatril, Fluxgate sensors for magnetopneumometry,Sensors and Actuators A: Physical 60 (1997) 76–79.

[3] S. Choi, S. Kawahito, Y. Matsumoto, M. Ispida, Y. Tadokoro, Anintegrated micro fluxgate magnetic sensor, Sensors and Actuators A:Physical 55 (1996) 121–126.

[4] L. Chiesi, P. Kejik, B. Janossy, R. Popovic, CMOS planar 2D micro-fluxgate sensor, Sensors and Actuators A: Physical 82 (2000) 174–180.

[5] P.M. Drljaca, P. Kejik, F. Vincent, D. Piguet, R.S. Popovic, Low-power2D fully integrated CMOS fluxgate magnetometer, IEEE SensorsJournal 5 (2005) 909–915.

[6] A. Baschirotto, E. Dallago, V. Ferragina, M. Ferri, M. Grassi, P.Malcovati, M. Marchesi, E. Melissano, M. Morelli, A. Rossini, S.Ruzza, P. Siciliano, G. Venchi, A CMOS 2D micro-fluxgate earthmagnetic field sensor with digital output, in: IEEE InternationalSolid-State Circuits Conference Digest of Technical Papers, 2007, pp.390–391.

[7] A. Baschirotto, E. Dallago, P. Malcovati, M. Marchesi, E. Melissano, P.Siciliano, G. Venchi, An integrated micro-fluxgate magnetic sensorwith sputtered ferromagnetic core, in: Proceedings of IEEEInstrumentation and Measurement Technology Conference, 2006,pp. 2045–2049.

[8] V. Ferragina, M. Ferri, M. Grassi, A. Rossini, P. Malcovati, A.Baschirotto, A 12.4 ENOB incremental A/D converter for high-linearity sensors read-out applications, in: Proceedings of IEEEInternational Symposium on Circuits and Systems, 2007, pp. 3582–3585.

[9] A. Baschirotto, E. Dallago, P. Malcovati, M. Marchesi, G. Venchi, Afluxgate magnetic sensor: from PCB to micro-integrated technology,IEEE Transactions on Instrumentation and Measurement 56 (2007)25–31.

[10] A. Baschirotto, F. Borghetti, E. Dallago, P. Malcovati, G. Venchi, ACMOS front-end circuit for integrated fluxgate magnetic sensors, in:Proceedings of the IEEE International Symposium on Circuits andSystems, 2006, pp. 4403–4406.

[11] A. Baschirotto, F. Borghetti, E. Dallago, P. Malcovati, M. Marchesi, E.Melissano, P. Siciliano, G. Venchi, Fluxgate magnetic sensor andfront-end circuitry in an integrated microsystem, Sensors andActuators A 132 (2006) 90–97.

[12] A. Baschirotto, E. Dallago, M. Ferri, P. Malcovati, A. Rossini, G. Venchi,A CMOS 2D micro-fluxgate earth magnetic field detecting systemwith RS232 digital output, in: Proceedings of IEEE InternationalConference on Sensors, 2007, pp. 240–243.