1088 IEEE SENSORS JOURNAL, VOL. 11, NO. 4, APRIL 2011

Transcription

1 1088 IEEE SENSORS JOURNAL, VOL. 11, NO. 4, APRIL 2011 A Three-Axial Search Coil Magnetometer Optimized for Small Size, Low Power, and Low Frequencies Asaf Grosz, Eugene Paperno, Shai Amrusi, and Boris Zadov Abstract A compact and sensitive three-axial search coil magnetometer has been designed, built, and tested. The magnetometer sensitivity threshold equals 12 pt/hz 0 5 at 1 Hz, and the magnetometer dimensions are 72 mm 69 mm 69 mm. All the magnetometer coils, the electronics, and batteries are accommodated within a single electrostatic shield and a single housing. A close to 2 aspect ratio (30 mm diameter, 58 mm total length) of the search coils provides a very high ( 70%) volume utilization factor. Such a small aspect ratio is obtained due to employing 30-mm diameter, 4-mm thick flux concentrators. The magnetometer is optimized for 20 mhz to 7 Hz frequencies and for ultra-low (252 W) power consumption. The ultra-low-power consumption enables a seven-year continuous operation from the four 1/2AA lithium batteries. The effect of the integration of three orthogonal search coils on the magnetometer sensitivity and accuracy has been investigated. Index Terms Compact, low frequency, magnetometer, search coil, three-axial, ultra-low power. I. INTRODUCTION I T HAS BEEN shown recently that a single-axial search coil magnetometer can approach both in size and resolution a fluxgate, even at so low frequency as 1 Hz [1], [2]. This is despite the inherent reduction of the coil sensitivity with decreasing the frequency. It has also been shown in [1] that the power consumption of a search coil magnetometer can be by two orders of magnitude lower than that of fluxgates. Thus, the continuous operation of a search coil magnetometer from a couple of AA batteries can exceed ten years. As a result, search coils become very competitive against fluxgates in such applications as magnetic anomaly detection, geophysical prospecting, earthquake prediction, space research, etc. The size reduction is obtained in [1] and [2], without decreasing the magnetometer resolution, thanks to employing thin, disk-shape flux concentrators and optimizing the search coil together with its electronics. In the present work, we integrate three orthogonal search coils, similar to [1], and design a three-axial magnetometer. Our aim is to minimize the total magnetometer volume. The latter Manuscript received July 30, 2010; revised September 03, 2010; accepted September 05, Date of publication September 27, 2010; date of current version February 16, This work was supported in part by Analog Devices, Inc., National Instruments, Inc., and the Ivanier Center for Robotics Research and Production Management. The associate editor coordinating the review of this paper and approving it for publication was Prof. Evgeny Katz. The authors are with the Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel ( grosz-t@zahav.net.il; paperno@ee.bgu.ac.il; amrusis@gmail.com; borisza@gmail.com). Color versions of one or more of the figures in this paper are available online at Digital Object Identifier /JSEN Fig. 1. Integration of orthogonal search coils: (left) a conventional [4] and (right) the new design. While having a similar sensitivity and weight, the new design occupies a 22 times smaller volume. is a limiting factor in various applications, for example, where many sensors are arranged in a network, or where an inhomogeneous magnetic field should be measured with a high accuracy. To reach this aim, we optimize the coil parameters, considering this new constrain. We also investigate the effect of the integration on the magnetometer accuracy, taking into account the distribution of the applied magnetic flux between the magnetometer coils. The designed three-axial magnetometer has a similar to fluxgates size and resolution at 1 Hz and significantly outperforms them in power consumption. Compared to conventional threeaxial search coil magnetometers with a similar resolution (see Fig. 1), the bulk of the suggested magnetometer is by an order of magnitude smaller. II. MAGNETOMETER DESIGN Much more compact design of search coils [1], [2], where they are thick but short, as compared to traditional designs [3] [6], where the coils are thin but very long, allows us to integrate three orthogonal coils in a single assembly with a very high volume utilization factor (see Fig. 1). One can see from Fig. 1 that three orthogonal coils with an aspect ratio of 2 very efficiently fill in the total cubic volume of the magnetometer. The two remaining small cubic volumes (see Fig. 2) are also utilized: one for accommodating the electronic board and the other for the batteries. Considering the aspect ratio of 2, we have optimized the magnetometer coils. The optimization goal was to reach the magnetometer sensitivity threshold comparable to that of fluxgates, namely, about 10 at 1 Hz X/$ IEEE

2 GROSZ et al.: A THREE-AXIAL SEARCH COIL MAGNETOMETER OPTIMIZED FOR SMALL SIZE, LOW POWER, AND LOW FREQUENCIES 1089 TABLE I PARAMETER OF THE MAGNETOMETER COILS Fig. 2. Magnetometer components (the windings are not shown; some other components, including the electrostatic shield and the magnetometer housing, and also the magnetometer assembly are shown in Fig. 11). The total length of the search coils is 60 mm, including the flux concentrators and conical springs to keep them attached to the coil cores. TABLE II COORDINATES OF THE CENTERS OF THE COIL CORE ENDS Fig. 3. Equivalent electrical circuit of a magnetometer channel. The damper R C is used to stabilize the circuit and to tune the coil resonant frequency to about 11 Hz. The channel electronics, which is represented here by block A, is given in Fig. 12. A. Optimization of the Search Coils The search-coil equivalent magnetic noise can be found as follows [3], [4]: where is the power spectral density of the coil thermal noise, is the Boltzmann constant, is the absolute temperature, is the coil resistance (see Fig. 3), and are the voltage and current spectral noise densities of the preamplifier, is the frequency, is the number of turns, and is the diameter of the core. The apparent permeability of the core with the flux concentrators in (1) can be given as follows [3], [6]: where is the relative magnetic permeability of the coil cores and flux concentrators, is the demagnetizing factor of a rod [7] with the diameter equal to that of the flux concentrators,, and with the length equal to the total length of the search coil, including the flux concentrators. The aim of the optimization is to find the minimum possible value of as a function of the diameter of the core and the diameter of the wire for given search-coil size and the type of preamplifier. (1) (2) The optimization has given the parameters listed in Table I for an OPA333 preamplifier and an MnZn ferrite with a 2000 relative magnetic permeability, which has been selected as the material for the flux concentrators an the coil cores. The optimization results are related to a single magnetometer channel and do not take into account the applied flux distribution between the coils in the three-axial assembly. To address this, we have performed the following numerical simulations. B. The Coil Integration Effect on the Magnetometer Sensitivity and Accuracy 1) Sensitivity: To examine the sensitivity of the three-coil assembly (see Table II), we have computed the apparent permeability of the coil cores with the help of commercially available three-dimensional final-element method (FEM) software, Maxwell The results are shown in Figs The cores apparent permeability averaged along their coil lengths (see Fig. 6) equals 212. Compared to a single coil, it is only by 4% lower. 2) Magnetic Crosstalk: We have also found in the above simulations, that the applied magnetic flux is flowing not only in the longitudinal coil but also in the transverse coils (see Figs. 4 and 5). This is because the transverse cores provide a bypass for the applied flux, and a part of it is flowing through them around the longitudinal core. For the same reason, a part of the secondary flux, generated in the longitudinal core by the electric current flowing in its coil (see Fig. 3), is also conducted by the transverse cores (see Fig. 5).

3 1090 IEEE SENSORS JOURNAL, VOL. 11, NO. 4, APRIL 2011 Fig. 4. Crosstalk due to the applied flux: magnetic induction within the cores for a uniform magnetic field applied along Z-core. Note that there is a net magnetic flux within the transverse cores. Relative to the primary flux in the longitudinal coil, the crosstalk =2:2%. Fig. 6. Crosstalk due to the applied and secondary fluxes: normalized magnetic induction along the coil cores axes. A uniform magnetic field is applied along Z-core (the solid curves, the maximum value of B is 219 T), and an electric current is applied to Z-coil (the dashed curves). Note the enlarged scale for the B and B magnitudes. Fig. 7. Magnetic crosstalk as a function of frequency. The solid curve and the squares represent the theoretical and experimental results, respectively. Fig. 5. Crosstalk due to the secondary flux: magnetic induction within the cores for an electric current flowing in Z-coil. (The coils are not shown.) Note that there is a net magnetic flux within the transverse cores. Relative to the secondary flux in Z-coil, the crosstalk =6:3%. We refer to the above effect as a crosstalk between the cores. The crosstalk can be described by the total magnetic flux flowing through the transverse coils relative to the primary flux, caused by the applied field in the longitudinal coil where is the total flux in the transverse coils, is the flux due to the applied field that causes a unit, and is the flux due to a unit secondary flux. The values of the and factors found with the help of the FEM simulations are 2.2% and 6.3%, respectively. To find these factors, we have averaged the magnetic induction inside the coils for the applied and secondary fields, respectively. (3) Considering Fig. 3, the secondary flux can be found as follows: (4) where and are the resistance and capacitance of a damper used to tune the coil self-resonant frequency and to stabilize the circuit. From (3) and (4), the crosstalk can be found as a function of frequency (see Fig. 7). One can see from this figure that the crosstalk nearly equals to below resonance and rapidly increases at resonance. To evaluate the effect of the crosstalk on the magnetometer accuracy, we have calculated the magnetometer outputs for a field vector rotating in such a way that its tip draws in space a spherical spiral. The following approximations has been obtained for the maximum uncertainty in the measured field direction (in degrees) and for the maximum relative uncertainty in the field magnitude (5)

4 GROSZ et al.: A THREE-AXIAL SEARCH COIL MAGNETOMETER OPTIMIZED FOR SMALL SIZE, LOW POWER, AND LOW FREQUENCIES 1091 Fig. 8. Shaping the magnetometer frequency response: the rise of the coils outputs with frequency is compensated by the roll-off of the preamplifier gain. The steep drop off of the magnetometer gain beyond 12 Hz is obtained with two second-order high-pass filtering stages. The magnetometer gain (sensitivity) is normalized to 7.7 V=pT. Fig. 9. Equivalent magnetic noise. Considering (5) and (6), and below resonance are 1.9 and { 2.2%, 4.4%}, respectively. This accuracy meets the requirements of our applications, where maximum uncertainties in the measured field direction and magnitude should be below and 5% for 20 mhz to 7 Hz frequencies. (6) C. SPICE Simulations We have simulated the magnetometer frequency response and noise by using commercially available SPICE software. The coil model corresponds to that given in [4]. 1) Frequency Response: The electronic circuit of the magnetometer is given in the Appendix (see Fig. 11). Each magnetometer channel comprises a damper, preamplifier U1 (OPA333), two second-order high-pass filters U2A and U2B (OPA2369), and an output amplifier U3 (OPA379). The channels outputs are connected to a 16-bit A/D converter (AD7682). The circuit shown in the bottom of Fig. A2 provides artificial ground to operate the magnetometer channels and the A/D converter in bipolar mode. The magnetometer frequency response is shown in Fig. 8. One can see from this figure, that the frequency response is nearly flat in a 20 mhz to 12 Hz frequency range. This is obtained due to the compensation of the rise of the coils outputs with frequency by the roll-off of the preamplifier gain. The upper limit of the frequency range is set at 12 Hz to shift the coils resonance frequency away from the maximum frequency (7 Hz) at which the crosstalk should still be low (see Fig. 7). The steep drop off of the magnetometer gain beyond 12 Hz is obtained with the two second-order high-pass filtering stages. It is needed to suppress the magnetometer sensitivity to interferences at frequencies of 50 and 100 Hz. This allows avoiding saturation and enables more efficient usage of the dynamic range of the 16-bit A/D converter. Even after the filtering, magnetic interferences exceed the signal in our applications by about 80 db. The 16-bit A/D con- Fig. 10. Magnetometer response to a rotating field. The field rotates in the xy-plane (top), xz-plane (middle), and yz-plane (bottom). verter meets this dynamic range. The magnetometer sensitivity (7.7 ) is high enough to bring the minimum noise density at the AD7682 converter input above its quantization noise. 2) Equivalent Magnetic Noise: Fig. 9 shows the magnetometer equivalent noise as a function of frequency. The predicted noise (12 at 1 Hz) is very close to the

5 1092 IEEE SENSORS JOURNAL, VOL. 11, NO. 4, APRIL 2011 Fig. 11. Assembly of the magnetometer (the coil windings and the batteries are not shown). The electrostatic shield (No. 8) is made of a flexible printed circuit board and is connected to the magnetometer ground terminal. The dimensions of the coils housing (No. 1) are 64 mm 2 64 mm 2 64 mm. The dimensions of the magnetometer housing (No. 13) are 72 mm 2 69 mm 2 69 mm. The magnetometer weight is about 600 g. desired value. The minimum noise at about 12 Hz. is obtained D. Power Consumption To minimize the magnetometer power consumption, the OPA333 zero-drift, ultra-low-power operational amplifiers with low white and no flicker noise have been chosen as preamplifiers (see Fig. 12). Their supply current is 17. To keep the power consumption as low as possible, we, instead of employing magnetic feedback, have flatten the magnetometer frequency response by reducing the preamplifier gain with frequency. The OPA379 and OPA2369 operational amplifiers have been chosen due to their very low supply currents (2.6 and 0.7, respectively) for the filtering stages and also for the circuit producing artificial ground. The supply current of the AD7682 A/D converter is also very low, about 1. Thus, the total supply current is only 70. The magnetometer is powered by 3.6 V lithium batteries: four 1/2AA batteries, 1.1 each, connected in parallel. Thus its total power consumption is only 252. Such ultra-low-power consumption enables a seven-year continuous operation. III. EXPERIMENT Experiments with a prototype have confirmed the theoretical evaluation of the magnetometer accuracy, frequency response, and noise (see Figs. 7 9). The magnetometer digital output was acquired by an external microcontroller. To measure the crosstalk, we have used a three-shell magnetic shield and a solenoid to apply a magnetic field. A very good agreement between the measurements and the theoretical prediction has been obtained (see Fig. 7). To measure the magnetometer accuracy, we used three-axial Helmholtz coils (the BH600-3-B type, manufactured by Serviciencia). A rotating magnetic field was applied to the magnetometer, sequentially in -, -, and -planes, and the channel outputs were recorded (see Fig. 10). These measurements have shown a less than 2.5 maximum uncertainty in the field direction and a less than 5% maximum relative uncertainty in the field magnitude. The magnetometer supply current and the power consumption do not exceed the theoretical values. IV. CONCLUSION A compact and ultra-low-power, three axial search coil magnetometer has been designed, built, and tested. Due to a close to 2 aspect ratio of the search coils, a very high volume utilization factor ( 70%) has been obtained. The coils, electronic board, and the batteries are accommodated within a single electromagnetic shield and housing. By contrast to conventional designs, [3], [4], we have shaped the magnetometer frequency response without using magnetic feedback. This simplifies the magnetometer and keeps its power consumption as low as possible. The obtained resolution (12 at 1 Hz) matches the theoretical value. The magnetometer inaccuracy has been shown to be less than 2.5 for the field direction and 5% for the field magnitude. The magnetometer occupies a 350- volume, and its built-in batteries provide a seven-year continuous operation. For a comparison, a state-of-the-art, ultra-

6 GROSZ et al.: A THREE-AXIAL SEARCH COIL MAGNETOMETER OPTIMIZED FOR SMALL SIZE, LOW POWER, AND LOW FREQUENCIES 1093 Fig. 12. Schematic of the electronic board. The coils are connected to the CON1 connector. The coils output signals are amplified, filtered, and then digitized with an AD7682 A/D converter. The magnetometer is powered by four 1/2AA lithium batteries connected in parallel to the CON3 connector. The circuit shown in the bottom provides artificial ground to operate the magnetometer channels and the A/D converter in bipolar mode. low-power fluxgate of the Bartington Mag 648 type, having a similar resolution and powered from batteries providing the same time span of operation, would occupy a 970- volume. The state-of-the-art three-axial search magnetometers [4] [6] occupy still larger volumes because of the relatively long lengths of their search coils: 10 and 17 cm, respectively. It is interesting to note that the volume of a single search coil in our design is three times larger than the volume of a single search coil in [4]. Integrating the coils, however, reverses the situation: the total volume of the new design becomes 22 times smaller than that of [4], having the same resolution and weight. It was not our aim to optimize the magnetometer weight; however, it exceeds that in [4] only by 37%: 600 g against 380 g. APPENDIX The assembly of the magnetometer is shown in Fig. 11 and the schematic of the electronic board is shown in Fig. 12. REFERENCES [1] E. Paperno and A. Grosz, A miniature and ultralow power search coil optimized for a 20 mhz to 2 khz frequency range, J. Appl. Phys., vol. 105, pp. 07E E710-3, 2009.

7 1094 IEEE SENSORS JOURNAL, VOL. 11, NO. 4, APRIL 2011 [2] A. Grosz, E. Paperno, S. Amrusi, and E. Liverts, Integration of the electronics and batteries inside the hollow core of a search coil, J. Appl. Phys., vol. 107, pp. 09E E703-3, [3] C. Coillot, J. Moutoussamy, P. Leroy, G. Chanteur, and A. Roux, Improvements on the design of search coil magnetometer for space experiments, Sensor Letters, vol. 5, pp , [4] H. C. Séran and P. Fergeau, An optimized low-frequency three-axis search coil magnetometer for space research, Review of Scientific Instruments, vol. 76, pp , [5] J. B. Cao et al., First results of low frequency electromagnetic wave detector of TC-2/Double Star program, Annales Geophysicae, vol. 23, pp , [6] A. Roux et al., The search coil magnetometer for THEMIS, Space Science Reviews, vol. 141, pp , [7] M. Kobayashi and Y. Ishikawa, Surface magnetic charge distributions and demagnetizing factors of circular cylinders, IEEE Trans. Magn., vol. 28, pp , Eugene Paperno received the B.Sc. and M.Sc. degrees in electrical engineering from the Minsk Institute of Radio Engineering, Minsk, Republic of Belarus, in 1983, and the Ph.D. degree (summa cum laude) from the Ben-Gurion University of the Negev, Beer-Sheva, Israel, in From 1983 to 1991, he was with the Institute of Electronics, Belorussian Academy of Sciences, Minsk. From 1997 to 1999, he was a Japan Society for the Promotion of Science Postdoctoral Fellow with Kyushu University, Fukoka, Japan. Since 1999, he has been with the Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev. His current interests include magnetic sensors, including atomic magnetometers, magnetic shielding, magnetic tracking, and magnetic and electronic instrumentation. Shai Amrusi received the B.Sc. degree in electrical and computer engineering from the Ben-Gurion University of the Negev, Beer-Sheva, Israel, in He is currently working towards the M.Sc. degree at the Department of Electric and Computer Engineering, Ben-Gurion University of the Negev. His current research interests are magnetic tracking systems and magnetic sensors. Asaf Grosz received the B.Sc. degree in physics and computer science from Tel-Aviv University, Tel-Aviv, Israel, in He is currently working towards the M.Sc. degree at the Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel. His current interests include magnetometry, magnetic sensors, and low-noise, ultra-low-power electronics. Boris Zadov received the B.Sc. degree in electrical and computer engineering from the Sami Shamoon College of Engineering, Beer-Sheva, Israel, in He is currently working towards the M.Sc. degree at the Department of Electrical and Computer Engineering, Ben-Gurion University of the Negev, Beer-Sheva. His current interests include magnetometry, magnetic sensors, and low-noise, ultra-low-power electronics.