High-precision angular or linear non-contact magnetic position sensor

Abstract

The invention relates to an, angular or linear, non-contact magnetic position sensor (C) comprising a permanent magnet (A) generating a magnetic field that varies non-linearly and at least a first pair of probes associated with the permanent magnet (A), each probe being able to measure at least two field components at the same point. A processing circuit is connected to the first pair and is configured to exploit at least some of the measurements furnished by the first pair of probes and to deliver a position signal representative of the absolute position of the permanent magnet (A).

Description

High precision non-contact angular or linear magnetic position sensor FIELD OF THE INVENTION

[0001] The present invention relates to the field of contactless, magnetic and / or electromagnetic position sensors for the precise measurement of an absolute angular or linear position. Such sensors make it possible to detect an angular position or a linear displacement with a high precision of the order of 0.1% of the total travel, i.e. < 0.5° for detection carried out over a complete revolution. Such systems, which can be robust to disturbances and of high precision, are particularly used in the automotive industry and in the robotics industry. TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0002] In the state of the art, document EP1083406 is known describing a rotary sensor composed of a ring magnet and two magneto-sensitive detection elements measuring the radial component of the field generated by the ring magnet, diametrically magnetized. Placed in quadrature, these two magneto-sensitive elements generate two sinusoidal signals which make it possible to detect the angular position of the rotating magnet via the implementation of the arc tangent function (Atan) of these two signals.

[0003] The disadvantage of this solution is its sensitivity to the placement of the magneto-sensitive detection elements to calculate an accurate angular value. This solution is also sensitive to any external field. Indeed, any stray field having a component measurable by at least one of the two magneto-sensitive elements will induce a significant non-linearity error.

[0004] We also know the document US7741839 presenting the principle of a rotary sensor composed of a ring magnet and two magneto-sensitive detection elements measuring at the same point two quadrature components of the field generated by the ring magnet. These two elements generate two sinusoidal signals which again make it possible to detect the angular position of the rotating magnet via the implementation of the Atan function of these two signals.

[0005] The accuracy of the detected signal is not satisfactory, limited by very different signal amplitudes, which leads to an erroneous calculation of the angular position. This solution is also sensitive to any external field.

[0006] US7030608 describes some configurations for achieving equality between the field components, but the impact is generally very limited and requires a large footprint. In addition, the problem of sensitivity to any external field is not solved, so this solution can only be considered very insufficient for many applications.

[0007] In order to improve accuracy, document FR2893410 proposes applying to the ratio of signals delivered by magneto-sensitive detection elements located at the same point, a compensation coefficient (gain) equal to the ratio of the maximum amplitudes of the measured quadrature field components. Thus the non-linearity of the signal is improved, but the configuration remains limited to the case of a diametrically magnetized ring magnet.

[0008] Document EP1989505 describes a linear or rotary sensor having a magnet whose magnetization is linearly variable. By applying a proportionality factor (gain), it is possible to determine the value of the linear or angular displacement of the magnet relative to the probe. But here again, the precision remains insufficient, particularly in the case where the magnetization harmonics of the magnet are significant or if the magnet has inhomogeneities.

[0009] Document FR2965347 reports a definition of the gain value on a case-by-case basis, through the performance of numerous tests or numerous calculations, to define a gain value different from the simple ratio of the maximum amplitudes of the components. This solution makes it possible to improve measurement accuracy, but remains very complicated to implement (test bench, simulations) and remains limited to certain configurations.

[0010] The disadvantage of these latter solutions remains their sensitivity to any external field.

[0011] Document FR2923903 describes an angular or linear sensor comprising a magnet with a magnetization direction varying linearly according to the displacement, using 2 sets of probes placed in quadrature, each probe comprising a pair of magneto-sensitive detection elements measuring quadrature field components. By combining these different field components to obtain two signals of the same intensity in quadrature, it is possible to obtain an angular measurement insensitive to the external field under certain conditions. However, the angular accuracy of the sensor remains modest. The principle remains valid for a linear sensor.

[0012] Document WO2009 / 101270 extends the principle described above and introduces a third and a fourth set of magneto-sensitive detection elements, "judiciously" offset from the first and second sets, and positioned between them identically to the first and second sets. By combining the signals from these 4 sets of magneto-sensitive detection elements, it is possible to obtain an angular measurement insensitive to the external field under certain conditions and to improve the accuracy (at least over a mechanical half-turn, i.e. 180°) of the sensor by correcting the magnetization error linked to the geometry of the magnet (in particular by compensating for the 3rd harmonic of magnetization). Considering the number of elements involved (4 separate probes), this solution remains very sensitive to probe positioning errors and is still too imprecise, with a non-linearity of + / -0.1% of the 180° angular range, or at best a precision of + / - 0.18° over this angular range.It is still too imprecise to be applied to certain automotive and robotic applications.

[0013] Document EP2711663 describes this time a sensor with two tracks and at least 2 probes, allowing a very high angular precision by combining a global detection carried out on a 1st track and a finer measurement resulting from the measurement of a second multipolar track. This solution nevertheless requires a good precision on the global detection, possibly via the use of additional probes (4 in total) on this sensor 1. The disadvantage of this solution remains its sensitivity to the external field, and its sensitivity to positioning errors of the probes and the two magnetic tracks.

[0014] Also known in the state of the art is document FR3118804 providing a simple device for angular or linear displacement measurement using (at least) a magneto-sensitive detection element capable of measuring at least two field components and a magnetized magnet with a magnetization profile comprising a combination of at least two periodic contributions making it possible to simultaneously calculate a global position in the revolution (or multi-turn) and a more precise local position of the displacement value.

[0015] A disadvantage of some of the state-of-the-art solutions is sensitivity to an external magnetic field.

[0016] It is always possible to shield the sensor, but this adds parts and significantly increases the manufacturing cost as well as the size of the sensor. SUBJECT OF THE INVENTION

[0017] The present invention proposes to solve, at least in part, the problems mentioned above by taking advantage of the principles disclosed in document FR3118804. The invention makes it possible in particular to facilitate the calculation of the global position and the local position. The invention makes it possible in particular to simply carry out a precise measurement of an angular (or linear) position and can be configured to be insensitive to external fields. BRIEF DESCRIPTION OF THE INVENTION

[0018] In order to achieve this aim, the subject of the invention proposes a contactless magnetic position sensor, angular or linear, comprising:a permanent magnet generating a magnetic field varying non-linearly according to the direction of movement, said variation comprising in the different field components a combination of at least a first quasi-periodic signal, called "carried signal" and at least a second quasi-periodic signal, called "carrier signal", the first and second signals being different from each other,at least a first pair of probes comprising a first main probe and a first secondary probe, the first pair being associated with the permanent magnet and each probe being capable of measuring at least two field components at the same point;anda processing circuit connected to the probes and configured to exploit at least some of the measurements provided by the probes and to deliver a position signal representative of the absolute position of the permanent magnet.;

[0019] According to the invention, the two probes of the first pair of probes are positioned relative to each other to allow, by differential combination of the measurements of the field components, to isolate a first quasi-periodic signal representative of a global position and a second quasi-periodic signal representative of a local position.

[0020] According to other advantageous and non-limiting characteristics of the invention, taken alone or in any technically feasible combination: the magnetic position sensor comprises a second pair of probes comprising a second main probe and a second secondary probe, the second pair of probes being associated with the permanent magnet and each probe of the second pair of probes being capable of measuring at least two field components at the same point, the two pairs of probes being positioned relatively to each other so as to obtain a quadrature of the signals obtained by differential combination; the processing circuit established: the measurement of the overall position of the sensor by combining the field components according to the formula: Atan2 ((Br1+Bt3)+(Br2+Bt4); Gain*((Br3-Bt1)+(Br4-Bt2))) or the measurement of the local position of the sensor by combining the field components according to the formula: Atan2 ((Br3-Bt1)-(Br4-Bt2); Gain*(Bz1-Bz2)),

[0021] in which Bri, Bti and Bzi represent respectively the radial, tangential and axial field measured by a probe of index i and in which Gain is a quantity chosen so as to minimize the angular error calculated on the excursion of the sensor; the processing circuit establishes the measurement of the local position of the sensor by combining the field components according to the formula: Atan2 ((Bz3-Bt1)-(Bz4-Bt2) ; Gain*((Bz1+Bt3)-(Bz2+Bt4))), in which Bri, Bti and Bzi represent respectively the radial, tangential and axial field measured by a probe of index i and in which Gain is a quantity chosen so as to minimize the angular error calculated on the excursion of the sensor; the processing circuit establishes the measurement of the global position and / or the measurement of local position on different faces of the permanent magnet simultaneously;the permanent magnet is a multi-pole magnet or an assembly of magnets or an equivalent machined magnet.the permanent magnet is constituted by a set of current loops;the permanent magnet is constituted by all or part of the rotor of an electric motor, a generator, an actuator, a reducer, a coupler, a gearbox, or an oscillator;each probe comprises a plurality of magneto-sensitive elements chosen from the list formed by a Hall probe, a magnetoresistive element, an eddy current element, a detection coil;the permanent magnet has a disc, ring or cylinder shape, the probes of the at least one pair of probes being arranged on the periphery of the magnet;the probes of the at least one pair of probes are separated by a first angular value corresponding to a half-period of the signal carried;the probes of the first pair of probes and the probes of the second pair of probes are respectively separated from each other by a second angular value substantially equal to a quarter of the period of the carrier signal; the first angular value and / or the second angular value is adjustable; the processing circuit is configured to use an angular compensation factor to correct the calculated angular values; the correction of the calculated angular values ​​is based on trigonometric functions; the processing circuit is configured to combine analog detection and digital detection of the field components.;

[0022] According to another aspect, the invention proposes to use a contactless magnetic position sensor as defined above for the measurement of additional values ​​on a complex system, comprising a force, a torque, an acceleration, a braking, a phase shift, an overall speed, a direction of movement, a distance, a number of revolutions, an inertia, an unbalance, a vibration, a noise, a harmonic content, a temperature, a pressure, an electric current, an electric voltage, an electric current, a frequency, an information coding. BRIEF DESCRIPTION OF THE FIGURES

[0023] Other characteristics and advantages of the invention will emerge from the detailed description of the invention which follows with reference to the appended figures in which:

[0024]

[0025] The figure represents a rotary sensor structure in a configuration according to the invention;

[0026]

[0027] La represents the evolution curve of the field components measured by a quasi-point probe of a sensor according to the invention;

[0028] La represents the evolution curve of the radial field components measured by a first pair of probes phase-shifted by a first chosen angular value and the combination, by addition, of these measurements;

[0029] La represents the evolution curve of the radial field components measured by the two pairs of probes arranged relatively to each other in a configuration in accordance with the invention, with a view to calculating the overall angular position;

[0030]

[0031] La represents the evolution curve of the overall angular value obtained by combining the radial component measured by four probes arranged relatively to each other in a configuration in accordance with the invention;

[0032]

[0033]

[0034] Figures 6 and 7 respectively represent the combinations of the tangential and axial components of the field, obtained by combining the measurements provided by four probes arranged relatively to each other in a configuration in accordance with the invention;

[0035]

[0036] La represents the evolution curve of the radial field components measured by the two pairs of probes arranged relatively to each other in a configuration in accordance with the invention, with a view to calculating the local angular position;

[0037]

[0038] La represents the evolution curve of the local angular value obtained by combining the radial component measured by 4 probes arranged relatively to each other in a configuration in accordance with the invention;

[0039]

[0040] La represents the curves of evolution of the field components measured by the two pairs of probes arranged relatively between them in a configuration in accordance with the invention, the measurements being combined with each other to be less sensitive to an external field in the calculation of the overall angular position;

[0041]

[0042] Illustrates the result of the calculation of the angular position provided by a sensor according to the invention implementing the combinations of the;

[0043]

[0044] La represents the curves of evolution of the field components measured by the two pairs of probes arranged relatively between them in a configuration in accordance with the invention, the measurements being combined with each other to be less sensitive to an external field in the calculation of the local angular position;

[0045]

[0046] Illustrates the result of the calculation of the angular position provided by a sensor according to the invention implementing the combinations of the;

[0047]

[0048] The represents the evolution curves of the field components measured by the two pairs of probes arranged relatively to each other in a configuration in accordance with the invention, the measurements being combined with each other according to an alternative combination to that represented on the;

[0049]

[0050] Illustrates the result of the calculation of the angular position provided by a sensor according to the invention implementing the combinations of the;

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057] Figures 16a,16b,17a,17b,18a,18b illustrate the impact of an external magnetic field on angular errors;

[0058]

[0059] La represents a sensor C according to the invention in a linear configuration;

[0060]

[0061]

[0062] Figures 20 and 21 illustrate the result of the calculation of the angular position provided by a sensor according to the invention operating in a degraded mode with only a pair of probes;

[0063]

[0064] The figure represents a rotary sensor structure in another configuration in accordance with the invention. DETAILED DESCRIPTION OF THE INVENTION

[0065] The represents a magnetic position sensor structure C according to an embodiment. The magnetic flux generated by a permanent magnet A in the form of a disc is collected at the periphery thereof by at least four sets of detection elements 1-4 (or magneto-sensitive elements) located radially or axially, without contact with the magnet. The permanent magnet may take a shape other than that of the disc of the magnet A shown in the, it may in particular be a ring or cylinder shape.

[0066] Each set of detection elements measures, at the same point, at least two components of the magnetic induction, if necessary by means of flux collectors defining an air gap in which the detection elements 1-4 are placed. The magneto-sensitive elements may comprise, for example, Hall probes, magneto-resistive elements, eddy current elements, detection coils. As is well known per se, each set of magneto-sensitive elements may be integrated into a housing to form a magnetic probe, capable of detecting two or three field components. For simplicity of expression, the term "probe" will be used in the remainder of this description to refer to a set of magneto-sensitive elements capable of detecting, almost punctually, two or three field components.This designation does not, however, in any way limit the implementation of the principles of the invention, this implementation being able to provide for placing, in the same housing, several sets of magneto-sensitive elements to detect, at a plurality of points, the components of the field.

[0067] The magnetic position sensor C therefore has four probes 1-4, comprising on the one hand a first main probe 1 and a second main probe 2, and on the other hand a first secondary probe 3 and a second secondary probe 4. The first main probe 1 and the first secondary probe 2 form a first pair of probes 1,2 angularly phase-shifted from each other by a first separation angle theta. The second main probe 3 and the second secondary probe 4 form a second pair of probes 3,4, also angularly phase-shifted from each other by the first separation angle theta. The first pair of probes 1,2 and the second pair of probes 3,4 are angularly phase-shifted from each other by a second separation angle beta.

[0068] The full interest of such a configuration will be detailed in a subsequent section of this description.

[0069] Permanent magnet A may be a multi-pole magnet or an assembly of magnets or an equivalent machined magnet. This magnet may consist of all or part of the rotor of an electric motor, a generator, an actuator, a reducer, a coupler, a gearbox, an oscillator. Alternatively, the permanent magnet may consist of a set of current loops.

[0070] The sensor also includes a processing circuit connected to probes 1-4 and configured to deliver a position signal depending on the absolute position (excursion) of permanent magnet A.

[0071] A magnetic position sensor C according to the invention has a magnetization profile incorporating the teaching of document FR3118804 cited in the introduction to the present application. This profile is complex and results in multi-periodic variations of the magnetization profile (or its orientation relative to an axis or reference point) as a function of the relative trajectory of the measuring air gap and the magnet. The permanent magnet A generates a magnetic field varying non-linearly along the direction of movement, said variation having, depending on the different field components, a form corresponding to a combination of at least two different quasi-periodic contributions.

[0072] This magnetization profile can thus present, according to a first pattern, called "carried", of period p and comprising P measurement increments. It can also comprise a second pattern, called "carrier", of period p*n (n real >0; constant or variable) comprising N increments. This double pattern makes it possible to combine a coarse detection (global output of the sensor) and a finer detection of the absolute position (local output of the sensor). A measurement increment is for example constituted by the measurement of a magnetic pole. Two poles of opposite polarity can thus constitute a pattern of given period. It is also possible to provide, in the magnetization profile, at least one magnetic anomaly leading to atypical fluxes serving as a "top turn" index making it possible to count the number of turns or events carried out beyond a first excursion. This output signal comprises a predetermined number of T measurement increments over the measurement interval.

[0073] By way of illustration, the three field components (radial component Br, tangential component Bt, and axial component Bz) from one of the probes 1-4 are shown for a magnetization profile comprising N=2 increments on the carrier signal, and P=48 increments on the carried signal, for an angular travel of 360° (one complete revolution). The number of magnetic anomalies T serving as a "top revolution" index that can help to count the number of revolutions is here considered to be zero so as not to unnecessarily complicate the description of the invention.

[0074] By judiciously combining the measurements made by two probes of the same pair of probes, we note that there is always (at least) a first angular value separating these two probes (denoted theta) making it possible to generate, by combining a measurement of a field component respectively produced by these two probes, a sinusoid of period representative of the carrier signal. This first angular value theta corresponds substantially to a half-period of the carried signal. This illustrates the result of the radial field measurements Br1, Br2 respectively provided by the first main probe 1 and the first secondary probe 2 for a first angular value of separation theta of the order of 7.5° (in this configuration P=48) of these two probes 1, 2, as well as the addition Br1+Br2 of these two signals.

[0075] By judiciously combining the measurements made by the second main probe 3 and the second secondary probe 4 phase-shifted from each other by the same first angular value theta, it is noted that there always exists (at least) a second angular value (denoted beta), respectively separating the first main probe 1 and the first secondary probe 2 from the second main probe 3 and the second secondary probe 4, making it possible to generate, by combining a measurement of a field component produced by these probes 3, 4, a second sinusoid with a period representative of the carrier signal, in quadrature of the sinusoid obtained from the first pair of probes 1, 2. This second angular value beta corresponds substantially to a quarter of the period of the carrier signal.

[0076] Illustrates, for a second angular value of beta separation of the order of 90° (for this configuration N=2), the result of the measurements of the radial fields Br3, Br4 respectively provided by the probes of the second pair of probes 3,4, as well as the addition Br3+Br4 of these two measurements.

[0077] Illustrates the result of the calculation of an angular position provided by a sensor C having the configuration in accordance with the invention which led to the results presented on the. This calculation uses the arctangent of the two sinusoids in quadrature resulting from the measured radial components Brad and combined by addition. The angular error of the calculated global position is + / -5° over a complete revolution.

[0078] The same strategy can be advantageously applied for the tangential Bt and axial Bz components of the field, as shown in Figures 6 and 7 respectively. The tangential Bt and axial Bz components are combined two by two by addition. It is thus possible to obtain three sets of curves, of similar intensities two by two, allowing an accurate calculation of the overall angular value.

[0079] For the calculation of the local position, the same approach can be used, this time combining the field components in order to highlight the signal carried, here by subtracting the components provided by the probes of the same pair two by two, as shown in the figure.

[0080] This illustrates the result of the calculation of the angular position provided by a sensor C having the configuration in accordance with the invention which led to the results presented on the. This calculation uses the arc tangent of the two sinusoids in quadrature resulting from the radial components Brad. The angular error of the calculated local position is + / - 3.5° over a fraction of a turn (a pair of increments of the signal carried, or approximately 15° mechanical, in the configuration taken as an example). The angular precision on the measurement of the mechanical angle is therefore 3.5 / 24 pairs of increments (P = 48), or + / -0.15°.

[0081] The same strategy can be advantageously applied for the tangential Bt and axial Bz components. It is possible to obtain three sets of curves (not shown), of similar intensities two by two, allowing a precise calculation of the local angular value.

[0082] This very general approach allows us to clearly understand the advantages of differential processing of at least one of the measured field components. By "differential processing or combination" we mean the combination, by difference or by addition, of the field components from at least part of the probes 1-4 of the sensor C. This makes it possible to highlight the carrier signal for the calculation of a global angle (position in the revolution) and the carried signal for the calculation of a local angle (position in the increment considered) by means of elementary calculations.

[0083] The simultaneous calculation of the two angular values ​​guarantees instantaneous and absolute knowledge of a precise angular (or linear) position, without the need for a minimum excursion to finalize a first calculation.

[0084] The use of a component ratio that is exploited in arctangent calculations ensures intrinsic insensitivity of the angular calculation to temperature variations. The components are affected in the same way, the ratio remains unchanged.

[0085] However, the field components used previously remain sensitive to an external field. To compensate for this phenomenon, it is possible to combine the signals provided by the probes so as to integrate external field compensation to obtain detection insensitive to a stray external field.

[0086] With reference to the, we note that: The first secondary probe 2 arranged along a first x axis measures Br2+Hx the radial field Br2 of the magnet to which is added the projection Hx along the first x axis of an external field Hext. Similarly, the second secondary probe 4 arranged along a second y axis, in quadrature with the x axis, measures Bt4-Hx the tangential field Bt4 of the magnet 4 to which is subtracted the projection Hx along the first x axis of the external field Hext. The sum of these two measurements (Br2+Hx)+(Bt4-Hx) is therefore immune to the component Hx along the first x axis of an external field. The same observation can be made by choosing the other pair of probes in quadrature 1,3. The first secondary probe 2 arranged along the first x axis measures Bt2+Hy the tangential field Bt2 of the magnet to which is added the projection Hy along the second y axis of an external field Hext.Similarly, the second secondary probe 4 arranged along the second y axis measures Br4+Hy the radial field Br4 of the magnet 4 to which is also added the projection Hy along the second y axis of the external field Hext. The difference of these two measurements (Bt2+Hy)-(Br4+Hy) is therefore immune to this component Hy along the second y axis of the external field Hext. The same observation can be made by choosing the other pair of probes in quadrature 1,3. Finally, the first secondary probe 2 arranged along the first x axis measures Bz2+Hz the axial field of the magnet to which is added the projection Hz along a third z axis (forming a trihedron with the first and second x,y axes) of an external field Hext. Similarly, the second secondary probe 4 arranged along the second y axis measures Bz4+Hz the axial field Bz4 of the magnet 4 to which is also added the projection Hz along the z axis of the external field Hext.The difference of these two measurements (Bz2+Hz)-(Bz4+Hz) is therefore immune to this Hz component along the third z axis of the external field Hext.

[0087] This description is based on the assumption that the external field impacts the four probes in a homogeneous manner. In reality, the small, but necessarily existing, distance between the probes results in small variations without significant impact on the final angular precision.

[0088] By combining the field components according to the formula (whose expression is simplified): Br1+Bt3 and Br2+Bt4, and their addition; Br3-Bt1 and Br4-Bt2, and their addition;

[0089] it is possible to obtain robust detection to the external magnetic field for the calculation of the global angular position, as shown in the.

[0090] Illustrates the result of the calculation of the angular position provided by a sensor C implementing the combinations just presented. This calculation uses the arc tangent (or the Atan2 function) of the ratio of the two curves of the same intensity: Atan2 ((Br1+Bt3)+(Br2+Bt4) ; Gain*((Br3-Bt1)+(Br4-Bt2))) the value of the Gain is adjusted so as to minimize the angular error calculated on the sensor excursion (here 0.998)

[0091] The angle measured over a complete mechanical revolution allows obtaining a value of the angular position at + / - 5° ().

[0092] Advantageously, by combining the field components according to the formula (shown on the): Br3-Bt1 and Br4-Bt2, and their subtraction; Bz1-Bz2;

[0093] it is possible to obtain robust detection to the external magnetic field for the calculation of the local angular position.

[0094] The value of the local angular position is calculated using the arc tangent (or the Atan2 function) of the ratio of the two curves of the same intensity: Atan2 ((Br3-Bt1)-(Br4-Bt2); Gain*(Bz1-Bz2)) the Gain value is adjusted to minimize the angular error calculated on the sensor excursion (here 0.933)

[0095] The angle calculated over one period of the carried signal (one pair of increments) provides a value of the local angular position to + / - 5° (electrical), as shown in the figure. In relation to the mechanical angle, the measured angular accuracy is 5 / 24 = + / - 0.21° mechanical over one complete revolution, the angular error measured for each increment pair over the revolution remaining very similar.

[0096] According to a particular embodiment, illustrated in the, the combination of the field components according to the formula: Bz1+Bt3 and Bz2+Bt4, and their subtraction; Bz3-Bt1 and Bz4-Bt2, and their subtraction;

[0097] it is possible to obtain a second detection for the calculation of the local angular position, with generally improved precision.

[0098] The value of the local angular position is calculated using the arc tangent (or Atan2) of the ratio of the two curves of the same intensity: atan2 ((Bz3-Bt1)-(Bz4-Bt2) ; Gain*((Bz1+Bt3)-(Bz2+Bt4))) the value of the Gain is adjusted so as to minimize the angular error calculated on the sensor excursion (here 0.995).

[0099] The angle calculated over one period of the carried signal (one pair of increments) provides a value of the local angular position to + / - 3° (electrical) as shown in the figure. In relation to the mechanical angle, the measured angular accuracy is 3 / 24 = + / - 0.125° mechanical over a complete revolution, the angular error measured for each increment pair over the revolution remaining very similar.

[0100] Figures 16a and 16b illustrate the impact of an external magnetic field Hext (measured alone at a probe: Bx=10 mT, By=-3mT and Bz=7 mT; homogeneous across all probes) on the measured field components () without external field Hext, () with the external field Hext.

[0101] Figures 17a,17b and 18a,18b allow us to compare the angular errors calculated in the two cases.

[0102] - in figure 17, in the presence of the external field Hext for (a) for the global angular position and (b) for the local position measured for an increment.

[0103] - in figure 18, without external field Hext (a) for the global angular position and (b) for the local position measured for an increment.

[0104] As expected, the signals are unchanged. Other tested field values ​​confirm these measurements.

[0105] The person skilled in the art will easily be able to differentiate between the present invention and the solution proposed by the prior art WO2009 / 101270 for calculating an angular travel carried out in radial detection on a diametrically magnetized ring for a field varying linearly with the angular position. In the context of a sensor C according to the present invention, the field measured by a magneto-sensitive element does not vary linearly with the angular position. This is particularly apparent on the. The two approaches cannot therefore be confused.

[0106] A sensor C according to the present description may comprise a signal processing circuit for combining the components and determining an instantaneous global and local angular value. This circuit may also carry out the various compensations enabling these angular values ​​to be corrected.

[0107] In order to reduce the calculation time, this device can combine analog detection and digital detection of field components.

[0108] According to a particular embodiment, the sensor C further comprises means for storing the angular values ​​and linearity coefficients used to compensate for the linearity of the sensor. These storage means are accessible to the signal processing circuit, which can use them to apply the processing aimed at compensating for the linearity of the sensor.

[0109] According to a particular embodiment, the sensor further comprises means for determining, as a function of the overall angular position, the increment or pair of increments opposite. This increment or pair of increments is said to be "active".

[0110] Each active increment is (for example) associated with an order number or even with a global angular position (identified for example with the "zero-crossing" of one of the field components or a composition of components). This method makes it possible to easily and precisely ensure that the read increment corresponds to the global value calculated, without risk of error with a neighboring increment.

[0111] Each local angular value calculated at the active increment can be used as a local angular value or percentage of full scale, or directly as a global angular value (by dividing the read value by the number of increment pairs (here 24), and then adding it to the global angular position associated with the active increment pair: For example, 183.4° calculated in the increment corresponding to a global position (zero crossing) of 157.3° instantly gives an angular position of 157.30+183.4 / 24 = 164.94°.

[0112] Since each increment may contain magnetization anomalies, the exact angular value of each increment can be used to respect the true angular values ​​read on the magnetic strip. Thus, for a pair of increments of 15.7° (and not 15° as theoretically expected): 157.30+183.4 / (360 / 15.7) = 165.29°.

[0113] According to a particular embodiment, an angular compensation coefficient is calculated for each increment (i) according to: Increment coefficient (i) = angular value of the theoretical increment / value of the measured increment. This coefficient is used to correct the local and / or global angular values.

[0114] According to a particular embodiment, the mode of detection of the field components can be of the axial type or of the radial type, or a combination.

[0115] According to a particular embodiment, the mode of detection of the field components can be carried out on different faces of the magnetic target simultaneously.

[0116] According to a particular embodiment, the sensor further comprises self-calibration means, making it possible to adjust certain calculation values ​​during the life of the sensor to maintain an optimum level of performance.

[0117] According to a particular embodiment, the sensor further comprises diagnostic means, making it possible to alert on a possible failure (probe, magneto-sensitive element, magnetic track, communication, calculation, temperature, external field, the travel of the sensor, the speed, etc.), and to engage a possibly degraded mode allowing the operation of the sensor with reduced performance. This degraded mode allows in particular detection using a minimum number of magneto-sensitive elements.

[0118] This gives a good indication of a global position calculation using only the signals from 2 probes out of phase by theta (for example the first pair of probes 1,2 or the second pair of probes 3,4 – left graph). The global position (right graph) is obtained with an Atan(Bz / Btan) calculation or with an Asin(Brad), in which Brad, Btan and Brz correspond respectively to the sum of the radial, tangential and axial field components provided by each of the two probes. The calculated linearity is + / -6° over a complete mechanical revolution.

[0119] This gives a good indication of a local position calculation using only the signals from two probes phase-shifted by the first angular value theta. The local position is obtained with an Atan(Bz / Btan) calculation or with an Asin(Brad), in which Brad, Btan and Brz correspond respectively to the differences in the radial, tangential and axial field components provided by each of the two probes (.b). The calculated linearity is + / -3° over a pair of increments, i.e. + / -0.125° or a complete mechanical revolution.

[0120] Other combinations of field components allowing the isolation of a global angular position and / or a local angular position may be retained to implement a valid detection mode according to the operating conditions envisaged.

[0121] In particular, it will be possible to use, as shown in the figure: a first pair of probes 1,2 phase-shifted by the first angular value theta corresponding to a half-period of the signal carried. This first pair of probes 1,2 can be used to determine a local position. a second pair of probes 3,4 phase-shifted by an angular value corresponding to a period of the signal. This second pair of probes 3,4 can be used to determine a global position.

[0122] the first pair and the second probe pair can in this case be separated by any second angular value (beta), for example chosen to be zero.

[0123] A method for determining the precise angular position based on the calculated angular measurements will firstly include the calculation of the global position and the local position. Secondly, a compensated value of these angular positions can be defined. Finally, a detection diagnosis can be carried out in order to validate the measurement and, if necessary, engage a degraded detection mode. The speed and direction of rotation values ​​will be defined from the calculated angular values.

[0124] According to a particular embodiment, a mathematical function (a polynomial for example) can be used to compensate for the angular value (or linear excursion) calculations. Each increment will be associated with the set of coefficients required to define this function (6 coefficients for a polynomial of order 5) and a combination of functions.

[0125] According to a particular embodiment, a trigonometric function (or its equivalent) may be used to compensate for angular value (or linear excursion) calculations. Each increment will be associated with the set of coefficients required to define this function (for example: intensity, period, angular shift, offset), and a combination of functions.

[0126] According to a particular embodiment, the sensor may use only a very small number of magneto-sensitive elements capable of measuring one or more field components, or any combination, to meet space requirements or even price. This solution may also be considered for detections over very limited travels (<360° for a rotary sensor).

[0127] According to another variant, the measuring device may include a plurality of probes, each arranged at a specific point, making it possible to double or triple the operating mode. By calculating an average of all the values ​​obtained by all of these probes, it will be possible to obtain an averaged angular value making it possible to reduce the uncertainty linked to a single set of probes (for example, a set of 4 probes capable of measuring one or more field components). The use of 5, 6, 7 or 8 probes may thus be envisaged, for example to produce a fully redundant sensor for applications requiring a maximum level of integrity and operational safety (ASIL D type for automobiles). This may also make it possible to enrich the combination of the different field components.

[0128] Of course, the invention is not limited to the method of implementation described and variant embodiments can be made without departing from the scope of the invention as defined by the claims.

[0129] Thus, the sensor C can take a linear configuration as shown. This linear configuration can be obtained by mentally unfolding a solution obtained for axial or radial angular detection. Since the field profile of the different field components is similar, the processing remains identical, taking care to properly position (linearly space) the sets of magneto-sensitive elements to obtain the first and second angular values ​​beta, theta valid for differential analysis.

[0130] Whether the C-sensor takes a linear or disc shape, the sets of magneto-sensitive elements can be repositioned relative to the permanent magnet, so that the first angular value theta and / or the second angular value beta can be adjusted.

[0131] The sensor can find an application for the measurement of additional quantities to that of an angular or linear displacement in a complex system. This quantity can correspond to a force, a torque, an acceleration, a braking, a phase shift, an overall speed, a direction of movement, a distance, a number of revolutions, an inertia, an unbalance, a vibration, a noise, a harmonic content, a temperature, a pressure, an electric current, an electric voltage, an electric current, a frequency, an information coding.

Claims

Contactless, angular or linear magnetic position sensor (C), comprising:a permanent magnet (A) generating a magnetic field varying non-linearly along the direction of movement, said variation comprising in the different field components a combination of at least a first quasi-periodic signal, called "carried signal" and at least a second quasi-periodic signal, called "carrier signal", the first and second signals being different from each other,at least a first pair of probes (1, 2; 3, 4) comprising a first main probe (1; 3) and a first secondary probe (2; 4), the first pair being associated with the permanent magnet (A) and each probe being capable of measuring at least two field components at the same point;anda processing circuit connected to the probes and configured to exploit at least some of the measurements provided by the probes and to deliver a position signal representative of the absolute position of the permanent magnet (A), the sensor being characterized in that the two probes of the first pair of probes (1, 2) are separated by a first angular value (theta) corresponding to a half-period of the signal carried to allow, by differential combinations of the measurements of the field components, to isolate a first quasi-periodic signal representative of a global position and a second quasi-periodic signal representative of a local position.; Magnetic position sensor (C) according to the preceding claim comprising a second pair of probes (3, 4) comprising a second main probe (3) and a second secondary probe (4) separated by an angular value corresponding to a period of the signal carried. Magnetic position sensor (C) according to claim 1 comprising a second pair of probes (3, 4) comprising a second main probe (3) and a second secondary probe (4), the second pair of probes (3, 4) being associated with the permanent magnet (A) and each probe of the second pair of probes (3, 4) being capable of measuring at least two field components at the same point, the two pairs of probes (1, 2; 3, 4) being positioned relatively to each other so as to obtain a quadrature of the signals obtained by differential combination. Contactless magnetic position sensor (C) according to the preceding claim in which the probes of the first pair of probes (1, 2) and the probes of the second pair of probes (3, 4) are respectively separated from each other by a second angular value (beta) substantially equal to a quarter of the period of the carrier signal. Contactless magnetic position sensor (C) according to the preceding claim in which the first angular value (theta) and / or the second angular value (beta) is adjustable. Contactless magnetic position sensor (C) according to one of claims 3 to 5, in which the processing circuit establishes: the measurement of the overall position of the sensor by combining the field components according to the formula: Atan2 ((Br1+Bt3)+(Br2+Bt4); Gain*((Br3-Bt1)+(Br4-Bt2))) or the measurement of the local position of the sensor by combining the field components according to the formula: Atan2 ((Br3-Bt1)-(Br4-Bt2); Gain*(Bz1-Bz2)), in which Bri, Bti and Bzi represent respectively the radial, tangential and axial field measured by a probe of index i and in which Gain is a quantity chosen so as to minimize the angular error calculated on the excursion of the sensor. Contactless magnetic position sensor (C) according to one of claims 3 to 6 in which the processing circuit establishes the measurement of the local position of the sensor by combining the field components according to the formula: Atan2 ((Bz3-Bt1)-(Bz4-Bt2); Gain*((Bz1+Bt3)-(Bz2+Bt4))), in which Bri, Bti and Bzi represent respectively the radial, tangential and axial field measured by a probe of index i and in which Gain is a quantity chosen so as to minimize the angular error calculated on the excursion of the sensor. Contactless magnetic position sensor (C) according to one of the preceding claims, in which the permanent magnet (A) is a multi-pole magnet or an assembly of magnets or an equivalent machined magnet. Contactless magnetic position sensor (C) according to one of claims 1 to 7, in which the permanent magnet (A) is constituted by a set of current loops. Contactless magnetic position sensor (C) according to one of the preceding claims, in which the permanent magnet (A) consists of all or part of the rotor of an electric motor, a generator, an actuator, a reducer, a coupler, a gearbox, or an oscillator. Contactless magnetic position sensor (C) according to one of the preceding claims, in which each probe (1, 2; 3, 4) comprises a plurality of magneto-sensitive elements chosen from the list formed by a Hall probe, a magneto-resistive element, an eddy current element, a detection coil. Contactless magnetic position sensor (C) according to one of the preceding claims, in which the permanent magnet (A) has the shape of a disc, ring or cylinder, the probes of the at least one pair of probes (1, 2; 3, 4) being arranged on the periphery of the magnet. Contactless magnetic position sensor (C) according to one of the preceding claims, in which the processing circuit is configured to exploit an angular compensation factor to correct the calculated angular values. Contactless magnetic position sensor (C) according to the preceding claim in which the correction of the calculated angular values ​​is based on trigonometric functions. Contactless magnetic position sensor (C) according to one of the two preceding claims in which the processing circuit is configured to combine analog detection and digital detection of the field components. Use of a contactless magnetic position sensor (C) according to any one of the preceding claims for measuring additional values ​​on a complex system, comprising force, torque, acceleration, braking, phase shift, overall speed, direction of movement, distance, number of revolutions, inertia, unbalance, vibration, noise, harmonic content, temperature, pressure, electric current, electric voltage, electric current, frequency, information coding.

Images