Influence Analysis of a Magnetic Field Focusing Device for Long Range Position Detection Measurement

Transcription

1 2016 UKSim-AMSS 18th International Conference on Computer Modelling and Simulation Influence Analysis of a Magnetic Field Focusing Device for Long Range Position Detection Measurement Marcelo Ribeiro Microsystem Technologies CTR Carinthian Tech Research Villach, Austria marcelo.ribeiro@ctr.at Abstract To overcome the issue of the fast decay of the magnetic fields of a magnet over distance, a device constructed of a highly permeable material was developed, aiming to focus the magnetic fields at a sensing point, hence named field focuser. This device allows designers to increase distances between magnet and sensor, chose weaker and cheaper magnets and work with simpler 1D magnetic sensors instead of more expensive 2 or 3D ones. This work focuses on the analysis of the influences and signal distortions that such a focusing device imposes in the system, aiming to highlight its performance. The findings were accomplished making use of FEM simulations and experimental validation, where the magnetic field on the presence and absence of the field focuser was analyzed. This paper aims to extend the knowledge on magnetic systems and help developers to understand limits and advantages of such setups. Keywords magnetic sensing; magnetic simulations; far magnetic field detection; field focusing I. INTRODUCTION While magnetic field sensors and sensing principles have been around for a long time [1], the last decades have witnessed an explosion in the development of magnetic sensing technologies driven by the technological advance and a rapidly growing demand in industrial applications [2]. This development has been made possible by the numerous advantages provided by magnetic sensors over other measurement systems, such as small packaging sizes and miniaturization [3], low production costs, contactless measurement possibilities, high resolution [4], low power requirements and an excellent robustness against vibrations, temperature, moisture and dirt. Today s magnetic sensors have established themselves in every major industrial sector with applications ranging from mechanical orientation, position and motion sensing, magnetic material and state analysis and detection, via navigation and guidance systems to a variety of applications in medicine, archeology, space exploration and countless others [5, 6]. In this study we focus on magnet-magnetic sensor systems. One of the great advantages when dealing with such systems is that the magnet, being the field source, is in principle a passive component that requires no power supply. At the same time, the magnetic field of a permanent magnet is strongly position dependent. With the potential for contact free position detection in combination with the robustness of such systems modern magnets can last easily for several years without showing demagnetization effects when treated accordingly. While there are several possible implementations, nowadays magnetic position measurement is practiced with two general methods: the simpler one is based on a 1D field measurement as one specific field component can be a linear function of the position; the more sophisticated one features a 2D sensor that picks up two components of the field and determines the position via a complex signal processing mechanism [7]. The advantages of the second are improved robustness against sensor displacement and temperature variation while featuring an excellent linear output signal over a wide range of magnet positions. Despite the advantages of 2D systems over their 1D counterparts, in certain cost driven industrial fields 1D systems have prevailed solely because the fabrication of 2D sensors is much more costly than the fabrication of simple 1D probes. In the context of this work, a field focusing device to enhance weak magnetic field measurement using ordinary sensors was developed, and this paper focuses in the analysis of the influences and distortions in the signal when making use of it. This device allows designers to increase distances between magnet and sensor, chose weaker and cheaper magnets and work with simpler 1D magnetic sensors instead of more expensive 2 or 3D ones. The following topics cover the problem description, the proposed solution and a description of the simulation environment and experimental setup used to test the model. The results from simulations and experiments are presented, the system is validated and its performance is analyzed. The conclusions are drawn in the last section, followed by proposals for future improvements. II. PROBLEM DESCRIPTION When dealing with linear distance measurements using magnet-magnetic sensor systems, one of the biggest difficulties is to achieve the magnetic field amplitudes, of the order of few tens of millitesla, required by the sensor manufacturers, e.g. [8]. This is due to the fact that amplitude of the magnetic field B(r) of a permanent magnet typically decays like (1) for distances that are of the order of the size of the magnet or larger [9] /16 $ IEEE DOI /UKSim

2 (1) Here r denotes the distance from the center of the magnet and M denotes the magnetization which is proportional to the remanence field of the magnet, B r = µ 0 M (µ 0 representing vacuum permeability). As the field amplitude scales linearly with the magnetization, in general one cannot compensate the fast decay with the distance by choosing stronger magnets. For example, to achieve a magnetic field amplitude of 20mT, as required by [8], for a cylindrical bar magnet of diameter D = 10mm, length L = 10mm and a very strong remanence field of B r = 1.1T, typical for NdFeB (Neodymium) magnets, the field amplitude drops to 20mT on the axis perpendicular to the symmetry axis of the magnet when only at ~17mm from the magnet surface, and drops down to ~2mT at a distance of 42mm from the magnet surface. For industrial applications this fact is quite troubling as the prices of permanent magnets, especially the strong rare earth magnets, make this system uneconomical when the distances that should be measured exceed few centimeters. Under the above circumstances, position and distance measurements are therefore only feasible in the immediate vicinity of the magnets. IV. AUXILIARY TOOLS In order to test and understand the system, magnetostatic simulations were performed using the FEM environment ANSYS Maxwell 15, while the experimental part was executed with a self developed system composed of a robot manipulator, a magnetic sensor, and hardware, firmware and software that access sensor data and provide output files for further analysis. A. Simulation setup Linearized models for the materials were used with, e.g. a typical relative permeability of iron (µ r = 3000) for the core, and typical values of the remanence field and permeability of NdFeB (B r = 1.05T and µ r = 1.007) for the magnet. Influences on the magnetic field by the sensor itself have been neglected. A snapshot of the 3D simulation model can be seen on Fig. 2. III. PROPOSED SOLUTION To overcome the limitations when dealing with position and distance measurements, a field focusing device that allows one to detect the weak fields of permanent magnets at large distances was developed. The focuser consists of a core made of a soft magnetic material of high permeability, which is divided into two pieces by a small gap with a magnetic field sensor placed inside. The underlying principle is that the highly permeable core attracts and enhances the weak external magnetic field, as outlined in the sketch in Fig. 1. Figure 1. Sketch of the field focusing principle If the gap is small, compared to the thickness of the core, the field propagation in the focuser can be described using magnetic circuit theory [10]. The two major effects are that the magnetic field in the core becomes homogeneous as the gap acts like a resistance, and that it propagates at full amplitude through the gap. Such a setup, thus, enhances the magnetic field that can be picked up by the sensor, while being robust against positioning tolerances of the sensor inside the core s air gap. Figure 2. Snapshot of the 3D simulation model In the model is possible to identify the magnet (blue), the field focuser (dark grey), the sensing lines (light grey) and the sensing point in the middle of the two cores of the focuser (zoomed in pink), where the origins of the coordinates are also located. B. Experimental setup The system consists of a robot arm with its controller (EPSON E2C351S), two sensor readout boxes (Measurement Computing USB1608FS and Xilinx Spartan 3 series XC3S1200E), a magnetic sensor (Melexis MLX90215), a integration software (CTR RAMSRS V1.0), and tools that can provide physical support for magnets and sensors. The MLX90215 was configured to its highest sensitivity, 140mV/mT, and powered with 5V. The detection of positively oriented fields is expressed from 2.47V upwards, while the detection of negatively oriented fields from 2.47V downwards. The conversion of the sensor readout from Volts to Tesla is explained by (2). Further information can be found in [11]. (2)

3 Here x denotes the sensor output in Volt and B the magnetic flux density expressed in mt. V. FOCUSER DEVELOPMENT The design, optimization and development of the field focuser is covered by another publication, please refer to [12] for further information. The core dimensions as well as the material characteristics based on the simulation outcomes were used to manufacture an according prototype, which can be seen in Fig. 3. The red areas represent strong magnetic fields and the blue areas weak magnetic fields. It is possible to see that the areas with higher values are concentrated around the magnet (as it is the magnetic field source) and in the ferrite cores of the field focuser. The sensing lines and sensing point have a larger concentration of vectors due to the finer mesh applied in the objects in order to achieve higher resolution results. A. Signal increase in the focuser s core Measurements were executed placing the field focuser on a fixed point and moving a permanent magnet in reference to this fixed point, performing readouts at the center of the gap between the ferrite cores for different XYZ magnet positions. Initial measurements were performed placing a rectangular shaped magnet parallel to the field focuser and moving it in X direction, as the sketch in Fig. 5 shows. Figure 3. Field focuser prototype The final design consists of two cylindrical pieces of ferrite with a relative permeability of µ r ~ 2000, a length of 43mm and a diameter of 6mm. It includes a plastic sensor holder, which holds the sensor and joins both core halves. Additionally, there is a hole where a screw can be placed to fixate the field focuser as an integrant of another system. All parts, except the core, are made of non-magnetic materials. VI. PERFORMANCE ANALYSIS There were two approaches used to identify the field focuser s performance: one was measuring the increase on signal in the focuser s core (and sensing point), and the other was to check the deformation in the signal in the vicinities of the focuser. While the first is a direct way of checking its performance, the second bring a deeper understanding in what happens with the signal and how the focuser biases the magnetic fields at a desired point. The measurements were analyzed comparing data of the magnetic field amplitudes in the presence and absence of the ferrite cores, both with simulations and experiments. A snapshot of the magnetic fields plotted in the simulation software, which illustrate the operation of the system, can be found on Fig. 4. Figure 5. Sketch of the measurements for detecting the performance increase over distance The measurements were carried out from distances of 10 to 190mm between magnet and cores (limited by the robot arm s operation range) with 10mm step size, which provided a dataset with 19 results for each of the four scenarios: simulations with and without the ferrite cores as well as experiments with and without the ferrite cores. Fig. 6 presents the magnetic field intensity in function of the distance from core for the described scenarios. Figure 6. Results of the performance increase over distance in the presence and absence of the field focuser Figure 4. Magnetic field vectors from the simulation results in presence of the field focuser The continuous lines represent simulation results and the points experimental results, being the red traces from the data without the ferrite cores and the blue traces with the presence of the cores. Analyzing the chart, it is clear that the field focuser improves the measuring range, as the amplitude is enhanced on the whole range in the presence of the cores when compared to their absence, reassuring the validity of

4 the proposed solution. The plateau of the sensor signal at small distances (blue points) is due to the saturation of the sensor, since the field was too strong as a result of the proximity of the magnet to the cores. This does not present a threat as the system is intended for long range measurements. A deeper analysis of the simulation data (chosen due to its lower fluctuations) brings out the field focuser s performance with a comparison between the signals, shown in Fig. 7. Figure 7. Difference between the signal with and without the ferrite cores extracted from simulations The blue trace represents the difference between the signal with and without the ferrite cores, and the red trace the percentage improvement of the signal strength in the presence of the cores. Although, the absolute signal improvement (blue) decays over distance, its percentage representation (red) shows that the performance increases over distance, which hints at a slower decay than would be expected having (1) in mind. Those results are promising, but in the real world applications low amplitudes result in low signal-to-noise ratios. Therefore, a further performance analysis had to be done taking the characteristics of the readout system used in the application into account. The sensor used in the experiments have noise levels ranging from 8mV pp to 60mV pp (50mV pp for the used configuration), limiting the measurement range to signals above the noise threshold. Fig. 8 shows a semi-logarithm plot of the simulated results with a reference line representing the sensor noise threshold. presence of the cores the critical point would move to approximately 140mm, a performance improvement of about 250%. Further analysis in the signal increase in the focuser s core can be found on [12]. B. Signal distortions in the focuser s vicinities Sensing lines were drawn in the vicinities of the field focuser cores in order to identify the changes in the signal caused by the cores presence. The principle was to observe the signal over specific lines on the presence and absence of the focuser and by checking the difference between both results, being then able to see how much was biased by the focuser into the sensing point. The setup was fixed on both cases with the magnet centered at the sensing point (Y = 0mm) and with a distance (airgap) of X = 50mm, while the sensing lines were drawn at 25mm and 50mm to the back of the cores, as well as at 25mm in between cores and magnet, with a length of 200mm each, as can be seen in Fig. 9. Figure 9. Sketch of the measurements for detecting the signal distortions in the focuser s vicinities Fig. 10 shows the signal in the sensing point in the middle of the ferrite cores for the above configuration on the presence and absence of the ferrite cores. Figure 10. (a) Signal with focuser, (b) Signal without focuser for the setup with X = 50mm and Y = 0mm Figure 8. Performance comparison over distance with noise threshold The blue trace represents the signal in the presence of the ferrite cores, the black trace the data from the simulation without the cores, and the red dashed line represents the sensor s noise barrier. Measurements without the ferrite cores would become critical at approximately 40mm of distance between sensing point and magnet, while in the The signal in the sensing point is about 18.5 times higher in the presence of the focuser than without it. Following measurements took place over the sensing lines without moving the sensor, cores or magnet, but just observing the magnetic field along the lines with the static system. Fig. 11 to 13 (a) show the signal on the presence and absence of the field focuser for all three sensing lines, while (b) represents the difference between both

5 Figure 11. (a) Signal with and without focuser, (b) Difference between signals for the sensor line SL1 It is possible to observe in that the signal without the cores has higher amplitude than on its presence, although only in the order of few µt since the sensing line SL1 is the farthest away from the magnet and the signal by such distance is already weak. The signal without cores has a flat part on its maximum, this is due to the vectors at this point are mostly pointing at the same direction while when on the presence of the cores the signal has a deformation due to the gap between both cores. This effect can also be observed over SL2, showed in Fig. 12. Different than SL1 and SL2, SL3 lies in the other side of the ferrite cores and closer to the magnet, so the signal over the line is way stronger and the effect is opposed: the signal with the cores is stronger than without it, as the cores are attracting more field lines to this region. From the information displayed on Fig. 11 to 13 it is possible easier to understand that the performance increase seem on Fig. 10 came from the biasing of the magnetic fields into the ferrite cores. Including a variation in the distance between magnet and the ferrite cores, Fig. 14 shows 3D plots of the signal over the sensing line SL2 on the presence and absence of the focuser for a fixed value of Y = 0mm and X = 10mm to 190mm with 10mm step sizes. Figure 14. 3D plot of (a) Signal with and (b) without focuser for the sensor line SL2 with X being the distance between magnet and cores, Y the signal over the sensing line and Z the magnetic field amplitude Figure 12. (a) Signal with and without focuser, (b) Difference between signals for the sensor line SL2 Similarly to SL1, SL2 showed a weaker signal in the presence of the cores, but had a more distorted shape and higher difference amplitude as the line lies closer to the ferrite cores and therefore its influence is more noticeable as the magnetic fields are mostly biased through the cores. The difference between signals for SL2 is about 4 times larger than SL1 when looking at the maximum points. It is possible to observe a variation in shape and amplitude as the magnet distance changes. The further the magnet moves away from the cores, the weaker is the signal but the broader is its spectrum, as the far fields are wider spread. The biasing effect of the cores can also be seen in the shape of the 3D plots, where the spectrum is broader in the presence of the cores when compared to the plot without it, although the amplitude decreases the closer the magnet gets to the sensing point. Fig. 15 brings another 3D view, this time having a fixed value for distance between magnet and sensing point X = 50mm and having a variable movement of the magnet parallel to the sensing lines Y = ±100mm with 10mm step size. Figure 15. 3D plot of (a) Signal with and (b) without focuser for the sensor line SL2 with X being the magnet movement parallel to the cores, Y the signal over the sensing line and Z the magnetic field amplitude Figure 13. (a) Signal with and without focuser, (b) Difference between signals for the sensor line SL3 These plots are likely the most expressive of the biasing effect of the ferrite cores, as it is possible to identify that

6 when the magnet is outside of the geometric limits of the cores (±46,5mm) the signal tends to be similar to the one in the absence of the cores, but as it approaches its corners, the signal is attracted to flow through a different path and there is shortly a higher signal concentration. As it moves closer to the inner parts of the cores, the signal is strongly biased into the cores, which causes an intense drop in the signal seen from the SL2. This is also the principle how a magnetic shielding would work, although here with the gap between cores one is able to place a sensor and take advantage of the increase in signal and redirection of the vectors. VII. CONCLUSION The fast magnetic field decay is a known characteristic when detecting fields of magnets at larger distances. The proposed concept of a field focuser was expected to decrease the impacts of this issue, and already at the initial stage it was clear that the proposal was successful. The prototype s performance analysis showed a massive improvement on the magnitude of the magnetic field amplitudes when at large distances from the magnets, and revealed that the signal detectability grows significantly with the increase of the distance between magnet and sensing point, as the concept works specially well for the far fields. Nevertheless, the measurement system poses a barrier by inserting noise in the data, thus limiting the measurement range to a point where data can be distinguished from noise. The analysis of the field distortions brought insights on how the magnetic field is affected by the presence of the cores and how the signal get magnified by the presence of the focuser. The disposal of the sensing lines was specifically chosen in order to see the biasing of the magnetic fields. It was possible to observe the redirection of the vectors of the magnetic fields and how the influence of the focuser concentrates the energy homogenously at the sensing point. The focuser s advantages lie not only on its increase in detectable distance, but also on the redirection and homogeneity of the magnetic field vectors, as this allows designers to develop systems which are less influenced by mechanical tolerances on the positioning of the sensor, as well as allows replacing the costly 2D sensors by simpler and cheaper 1D ones, as the vectors at the gap between cores are mostly pointing straight and do not require an additional component to be detected. The proposed system is a cost effective way to enhance the detection of weak magnetic fields with ordinary magnetic field sensors. Such systems can be used for position measurement, taking advantage of the extension of the measurement ranges. VIII. FUTURE WORK PROPOSAL Considering the consequences that the measurement system poses on the results, the influence of the X offsets and the performance peaks for different XY coordinates, one could experiment with different sensor configurations and with different setups for sensitivity, gain, etc., to fine tune to the desired readout region, also taking advantage of the minima and maxima found along the XY plane. ACKNOWLEDGMENT The Competence Center CTR is funded within the R&D Program COMET - Competence Centers for Excellent Technologies by the Federal Ministries of Transport, Innovation and Technology (BMVIT), of Economics and Labor (BMWA) and it is managed on their behalf by the Austrian Research Promotion Agency (FFG). The Austrian provinces (Carinthia and Styria) provide additional funding. REFERENCES [1] E. Hall, On a New Action of the Magnet on electric Currents, American Journal of Mathematics 2: , [2] T. Bratland, M. J. Caruso, R. W. Schneider, A New Perspective on Magnetic Field Sensing, Sensors, [3] D. Niarchos, Magnetic MEMS: key issues and some applications, Sensors and Actuators A: Physical 106, 1-3: , [4] A. Fert, The origin, development and future of spintronics, Soviet Physics Uspekhi 178, 12: , [5] J. Lenz and A.S. Edelstein, Magnetic Sensors and their Applications, IEEE Sens. J. 6: , [6] M. Diaz-Michelena, Small Magnetic Sensors for Space Applications, Sensors 9: , [7] M. Ortner, Improving magnetic linear position measurement by field shaping, Publication at the 9th International Conference on Sensing Technology (ICST 2015), , Auckland, New Zealand. [8] Melexis, Magnets for MLX90333 linear position sensor, Appl. Note LP-AP , 2007 [9] D. J. Griffiths, Introduction to Electrodynamics, 1999, Prentice- Hall, Inc. Upper Saddle River, New Jersey [10] A. R. Hambley, Electrical Engineering: Principles and Applications, 4th ed., Pearson Education, Inc., [11] Melexis, Precision Programmable Linear Hall Effect Sensor, MLX90215 datasheet, Sep [12] M. Ribeiro, M. Ortner, and M. Seger, Long range magnetic field measurement with magnetic sensors, in Proceedings of the International Conference on Industrial Automation, Information and Communications Technology (IAICT 2014), A. H. Gunawan, Ed., Telkom Indonesia, pp DOI: /IAICT