Methods for compensating angular errors

Abstract

Method for compensating angular errors in the determination of angular values ​​from an arrangement with a permanent magnet and at least two magnetic field sensors with a distance formed between the permanent magnet and the magnetic field sensors, wherein a sinusoidal signal and a cosine signal are generated in the magnetic field sensors, and an evaluation signal corresponding to the angular values ​​is formed from the two signals by means of an arctangent function, wherein arrangement errors and errors of the measuring and evaluation unit change the shape of the evaluation signal and lead to an angular deviation in the evaluation signal, and in order to minimize the angular deviation, offset values ​​of the signals are changed, and offset values ​​are determined by determining a first harmonic in the evaluation signal and / or a second harmonic in the signals and / or density differences of measurement points of a Lissajous plot.

Description

[0001] The invention relates to a method for compensating angular errors.

[0002] DE 10 2014 109 693 A1 discloses a method in which the angular position of a magnetic device with a rotational axis is determined by a non-contact measurement. The magnetic device comprises a magnet and at least three magnetic field sensors, and a sinusoidal and a cosinusoidal output signal are calculated from the signals of the magnetic field sensors using a transformation.

[0003] From the two output signals, or from the instantaneous values ​​of the two output signals, the angle of rotation, i.e. the angular position of the magnet or sensor arranged on the axis of rotation, is calculated directly in an evaluation device using an arctangent function.

[0004] By using at least three magnetic field sensors, magnetic interference fields can be compensated, which are particularly common in automotive electronics applications. The device and method are not limited to three magnetic field sensors. Using more than three sensors allows for more precise calculation of the angular position.

[0005] However, such methods are not suitable for compensating for errors due to positioning and position inaccuracies. For example, both the magnet and the sensor device may be axially offset from each other due to tolerances, or exhibit an inclination or tilt.

[0006] Further errors in the signals and, consequently, in the angle determination are caused by non-ideal magnetic properties, particularly temperature changes and aging. Basically, the errors can be divided into three classes: - offset errors that cause DC components of the signals, and - amplification errors causing deviations in the peak values ​​of the amplitudes of the signals, and - Orthogonality errors that cause deviations in the phase shift of the signals.

[0007] Further methods for error compensation are known from "Application Note APN000202_001EN." The underlying devices have means for implementing compensation methods.

[0008] However, due to the variety of error sources and the superposition of error sources, such as an axial misalignment between the magnet and sensor device in conjunction with a tilting of the magnet, effects occur that cannot be corrected with the compensation methods mentioned.

[0009] Against this background, the object of the invention is to provide a method which further develops the state of the art.

[0010] The object is achieved by a method for compensating angular errors with the features of patent claim 1. Advantageous embodiments of the invention are the subject of subclaims.

[0011] According to the subject matter, a method for compensating angular errors is provided.

[0012] The method is used to determine angle values ​​from an arrangement with a permanent magnet and at least two magnetic field sensors.

[0013] The two magnetic field sensors are connected to a measuring and evaluation device.

[0014] A gap is formed between the permanent magnet and the magnetic field sensors.

[0015] By means of a forced mechanical movement, a sinusoidal signal and a cosinusoidal signal are generated in the magnetic field sensors or the sinusoidal signal and the cosinusoidal signal are generated by means of a transformation.

[0016] From the two signals, an evaluation signal corresponding to the angle values ​​is formed using an arctangent function, whereby arrangement errors and errors from the measuring and evaluation unit change the shape of the evaluation signal.

[0017] In the evaluation signal, the errors lead to an angular deviation, whereby in order to minimize the angular deviation, values ​​for correcting the offset error from the signals, hereinafter referred to as offset values, are changed.

[0018] In this process, the offset values ​​are determined in one step by determining a first harmonic in the evaluation signal and / or a second harmonic in the signals and / or density differences of measuring points of a Lissajous plot.

[0019] It should be noted that the measuring and evaluation unit generally does not determine the offset values, but rather applies values ​​to correct the offset, which have already been determined, for example, during a calibration. The latter usually takes place outside the measuring and evaluation unit and is generally stored.

[0020] In other words, offset errors, ie DC components in the signals, are known before the procedure is carried out and are taken into account by the measuring and evaluation unit, ie the signals are corrected by the DC components.

[0021] It should be noted that in contrast to the prior art, in which, for example, the minimum and maximum values ​​of the signals are considered, whereby the resulting angular deviation is in some cases even increased instead of reduced, in the present invention the offset values ​​are determined using a different method in order to reliably correct first-order angular deviations.

[0022] It is understood that the term "magnetic field sensors" includes, in particular, vertical Hall sensors and / or lateral Hall sensors and / or TMR and / or AMR and / or other types of magnetic field sensors. Furthermore, different types of magnetic field sensors can also be used in the arrangement.

[0023] It should be noted that even in a design with just two sensors, the sinusoidal signal and the cosinusoidal signal are either directly obtained as a measurement signal from the two magnetic field sensors or can only be derived by transforming the measurement signals. Especially with more than two magnetic field sensors, the sinusoidal signal and the cosinusoidal signal can generally only be generated by transforming them.

[0024] Even with exactly two magnetic field sensors, a transformation may be necessary depending on the arrangement of the two magnetic field sensors and / or when using different types of magnetic field sensors, for example, one of the two magnetic field sensors is designed as a TMR and the other as a Hall sensor.

[0025] It should be noted that by means of the method, the second harmonic in the signals is minimized by changing offset values.

[0026] The determination of the second harmonic can be determined directly in the signals or in the resulting effects, such as the resulting angular deviation of the evaluation signals or the density function of the Lissajous plot of the two signals.

[0027] It should be noted that in an alternative approach, the angular deviation in the evaluation signals should be minimized, which occurs as the first harmonic. In other words, the first harmonic is minimized.

[0028] In another alternative, the density differences of the measurement points are compensated during the evaluation using the Lissajous plot.

[0029] It should be noted that the approaches described above for reducing errors are equivalent to each other in terms of their effect.

[0030] It is understood that error values ​​can be determined and adjusted directly in the signals and only then are the offset values ​​determined using the method according to the invention.

[0031] An advantage of the method is that errors resulting from axial misalignment between the magnet and sensor device or from tilting of the magnet and other error causes that cause higher-order effects in the signals can be compensated.

[0032] In particular, a second harmonic can be determined in the sinusoidal and cosinusoidal signals, which can be seen as the first harmonic of the angle of rotation value and the resulting deviation of the angle of rotation. It is understood that it is advantageous to consider at least one 360° sweep.

[0033] Furthermore, it should be noted that the method significantly extends the compensation of angular errors, which occur in the form of the second harmonic of the signals.

[0034] This increases the achievable precision of angle measurement, while also reducing manufacturing tolerance requirements and reducing the yield and effort involved in assembling the fixture. In other words, manufacturing costs can be reduced.

[0035] It is understood that the method is not limited to angle measurement. The method can be used for all forms of contactless displacement measurement in which at least one permanent magnet, in combination with at least two magnetic field sensors, generates a sine and cosine signal. This also encompasses any type of linear movement whose magnetic structure is designed accordingly.

[0036] In one embodiment, after calculating the angle values ​​using the arctangent function, the offset values ​​are determined from the amplitude and phase of the sinusoidal component, ie the first harmonic of the resulting angular deviation of a 360° rotation.

[0037] In one embodiment, to determine the offset values ​​from the density differences of the measurement points of the Lissajous plot, the measurement points are recorded at a constant period speed / frequency or rotation speed with a fixed sampling rate.

[0038] In another embodiment, the offset values ​​are determined from the component of the first harmonic calculated by DFT or FFT in the resulting course of the error of the rotation angle of a 360° revolution.

[0039] In a further development, the offset values ​​are determined from the change in correction values ​​already generated by other methods in conjunction with the second harmonic of the sinusoidal and cosinusoidal signals calculated by DFT or FFT.

[0040] In other words, by means of the method according to the invention, after the correction from the determination of the first harmonic in the signals, the offset values ​​from the determination of the second harmonic are calculated.

[0041] In a further development, the DFT calculation is performed using a Goertzel algorithm, or, alternatively, the first harmonic or second harmonic is directly determined using a sinusoidal fitting function. In other words, in this alternative, no FFT or DFT calculation is performed. One advantage is that the computing power required to implement the method is reduced.

[0042] In one embodiment, it is preferred that the offset values ​​be determined multiple times, iteratively. This allows the accuracy of the angle calculation to be further increased.

[0043] In a further development, the magnetic field sensors are spaced along an X-direction.

[0044] In another embodiment, the magnetic field sensors are arranged directly next to one another in the X-direction and / or are arranged partially or completely next to one another in a Z-direction.

[0045] In a further embodiment, three magnetic field sensors are arranged in the form of a pixel cell, with the three magnetic field sensors measuring three different directions of a magnetic field. In a further development, several magnetic field sensors are arranged in an XY plane, preferably linearly or in a circle.

[0046] According to another further development, after correction values ​​have already been determined, a second harmonic is determined from a comparison of the signal curve with an ideal sine curve and from a comparison of the signal curve with an ideal cosine curve directly at the measured values.

[0047] Offset values ​​are determined from the correction values ​​already determined and the deviation from the ideal curves.

[0048] In one embodiment, the method is used to determine the rotational angle position in an arrangement with permanent magnets and magnetic field sensors.

[0049] The arrangement has an axis with an end face and at least one permanent magnet arranged on the end face or laterally on the axis.

[0050] Magnetic field sensors are arranged along an imaginary extension of the axis. Alternatively, the magnetic field sensors are arranged laterally, i.e., spaced apart from the imaginary extension of the axis.

[0051] It is understood that there is a distance between the axis and the magnetic field sensors.

[0052] A measuring and evaluation device is connected to the magnetic field sensors, whereby when the axis rotates, a sinusoidal signal or a cosinusoidal signal is present at an output of the respective magnetic field sensor or is only obtained by means of a transformation.

[0053] In a further development, the magnetic field sensors are spaced apart along an X-direction or arranged directly adjacent to one another in the X-direction and / or are arranged partially or completely adjacent to one another in a Z-direction. If the magnetic field sensors are arranged adjacent to one another or adjacent to one another, different directions of the magnetic field are measured by the two magnetic field sensors.

[0054] The measuring and evaluation device is designed to calculate an evaluation signal corresponding to the angle of rotation values ​​from the signals using an arctangent function.

[0055] Furthermore, the measuring and evaluation device is designed to calculate offset values ​​for a change in the evaluation signal in order to minimize the angular deviation.

[0056] In an alternative embodiment, the method is applied to a different arrangement for determining angle values ​​of a translational movement path.

[0057] Here, the other arrangement comprises a plurality of permanent magnet pairs arranged along the translational path and at least two magnetic field sensors arranged along the translational movement path.

[0058] It is understood that the magnetic field sensors are spaced from the permanent magnet pairs and are arranged to be movable along the translational movement path.

[0059] Furthermore, a measuring and evaluation device connected to the magnetic field sensors is provided.

[0060] During a translational movement of the permanent magnet pairs or the magnetic field sensors along the path, a sinusoidal signal or a cosinusoidal signal is present at an output of the respective magnetic field sensor or the two signals are generated by means of a transformation.

[0061] The measuring and evaluation device is designed to calculate an evaluation signal corresponding to the angle values ​​from the signals using an arctangent function.

[0062] Furthermore, the measuring and evaluation device is designed to calculate offset values ​​for a change in the evaluation signal in order to minimize the angular deviation.

[0063] In a preferred embodiment, the measuring and evaluation device is configured to determine offset values ​​by means of a DFT or FFT evaluation.

[0064] In a further development, the measuring and evaluation device comprises a Goertzel filter or the measuring and evaluation device is configured to perform a sinusoidal adaptation function.

[0065] In an alternative development, the measuring and evaluation device is set up to determine density differences between the measuring points of a Lissajous plot.

[0066] In a preferred embodiment, the measuring and evaluation device with the magnetic field sensors is integrated monolithically or as a multi-chip solution in an IC housing.

[0067] The invention will be explained in more detail below with reference to the drawings. Similar parts are labeled with identical designations. The illustrated embodiments are highly schematic, ie the distances and the lateral and vertical extensions are not to scale and, unless otherwise stated, do not have any deducible geometric relationships to one another. Fig. 1 faulty sinusoidal and cosinusoidal signals, Fig. 2 a spectrum of the signals or the evaluation signal, Fig. 3 a representation of the resulting angular error of the evaluation signal, Fig. 4 a representation of a Lissajous plot of erroneous sinusoidal and cosinusoidal signals in the form of density changes of the measuring points, Fig. 5 a Goertzel filter, Fig. 6 a rotating device, Fig. 7 a linear device, Fig. 8 a representation of ideal sinusoidal and cosinusoidal signals.

[0068] It should be noted that in the following figures, the sinusoidal and cosinusoidal signals from the two magnetic field sensors S1 and S2 are either directly available as measurement signals or only after a transformation.

[0069] The illustration of the Fig. 1 shows a view of the uncompensated sinusoidal and cosinusoidal signals as a function of the rotation angle value.

[0070] The signals exhibit three types of errors. A DC component 3, often referred to as offset, causes one or both measurement signals to modulate in a positive or negative direction.

[0071] A gain error 4 leads to a difference in the peak values ​​of the two measurement signals without DC component.

[0072] An orthogonality error 5 results in a phase shift of other than 90° in the signals. The orthogonality error describes the deviation of the two measured magnetic field vector components from an ideal 90° orientation.

[0073] The illustration in Fig. Figure 2 shows the spectrum of the signals. In the case of the spectrum of the angular deviation of the evaluation signal, the deviation would occur in the form of a first harmonic 6. By performing a compensation according to the invention, the resulting error is reduced.

[0074] The figure also serves to display the spectra of the measurement signals. In measurement signals in which the DC component 3, the gain error 4, and the orthogonality error 6 have already been adjusted, a second harmonic, often referred to as the first harmonic, can still occur.

[0075] The first harmonic creates a difference to the first harmonic, which forms the actual signal and is often referred to as the fundamental.

[0076] The deviation causes an error delta Phi in the resulting rotation angle Phi when calculating the arctangent.

[0077] The application of the method according to the invention changes the offset values ​​and thereby minimizes angle errors Delta Phi caused by the second harmonics present in the signals.

[0078] The zeroth harmonic corresponds to the DC component. The method according to the invention is characterized by changing the offset values ​​of the two measurement signals.

[0079] It is inherent in the method that in the case of a second harmonic present in the measurement signals, a different offset leads to a smaller resulting angular error than one according to the known correction, in which an offset in the signals is directly adjusted.

[0080] This can be described as overcompensation. In other words, using the method according to the invention, the overcompensation of the second harmonic error in the signals leads to an increase in the accuracy of the angle calculation.

[0081] In the figure in Fig. Figure 3 shows a typical curve of the resulting angle error Delta Phi of the evaluation signal. In this case, the error occurs as a first harmonic oscillation. The characteristic shown shows a deviation in the measurement signals in the form of a pronounced second harmonic. It should be noted that an offset also leads to the displayed characteristic.

[0082] In the illustration of the Fig. Figure 4 shows a Lissajous plot of the sinusoidal and cosinusoidal signals.

[0083] The plot is not further deformed, which means that the three errors in question are already perfectly aligned. However, the plot does not show any equidistant measurement points.

[0084] The density distribution indicates the presence of a second harmonic in the measurement signals. Changing the offset would shift the plot but would smooth out the density distribution and thus minimize the resulting angle error, delta phi.

[0085] In the illustration of the Fig. Figure 5 shows an example of a second-order Görtzel filter. The use of the Görtzel filter allows the targeted determination of individual spectral components of a signal. The Görtzel filter is particularly suitable for determining the second harmonic of signals according to the invention.

[0086] In the illustration of the Fig. 6 depicts a rotating device DV, wherein the rotating device DV has a rotation axis DA formed in a Z direction and a permanent magnet PM arranged on an end face of the rotation axis DA in an XY plane. The permanent magnet PM has a north pole N and a south pole S. A first magnetic field sensor S1 and a second magnetic field sensor S2 are arranged on a carrier board TR along an imaginary extension of the rotation axis DA.

[0087] The two magnetic field sensors S1, S2 are spaced apart in the X-direction by a distance x1. Both magnetic field sensors S1, S2 are also spaced apart from the head end of the rotation axis DA. If the carrier board with the two magnetic field sensors S1, S2 is tilted relative to the front end, or if the magnetic field sensors S1, S2 are misaligned, the signals will be subject to different errors.

[0088] In an embodiment not shown, the distance x1 between the two magnetic field sensors S1 and S2 is zero or very small, or the two magnetic field sensors S1 and S2 are arranged one above the other in the Z direction, wherein in the arrangement one above the other the two magnetic field sensors preferably measure different components of the magnetic field.

[0089] In the illustration of the Fig. Figure 7 shows an arrangement of pairs of permanent magnets along a straight line.

[0090] Using the linearly movable permanent magnet PM, which has the pole pairs 10-16 arranged in a row, the spaced-apart magnetic field sensors S1 and S2 can generate a sine and a cosine signal, respectively. The inventive compensation of angular errors due to second-order deviations in the measurement signals can also be applied to linear motion, including translational displacement / position measurement.

[0091] The illustration of the Fig. Figure 8 shows an ideal curve of the two signals with a sinusoidal signal 20 and a cosinusoidal signal 10.

[0092] Both the sinusoidal signal and the cosinusoidal signal are ideal, i.e. error-free. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2014 109 693 A1

[0002] Cited non-patent literature

[0000] Application Note APN000202_001EN

[0007]

Claims

[1] Method for compensating angular errors when determining angular values ​​from an arrangement with a permanent magnet (PM) and at least two magnetic field sensors (S1, S2), - a measuring and evaluation device connected to the magnetic field sensors (S1, S2), - a distance formed between the permanent magnet and the magnetic field sensors (S1, S2), - a sinusoidal signal and a cosinusoidal signal are generated by means of a forced mechanical movement in the magnetic field sensors (S1, S2) or the sinusoidal signal and the cosinusoidal signal are generated by means of a transformation, and an evaluation signal corresponding to the angle values ​​is formed from the two signals by means of an arctangent function, whereby - Arrangement errors and errors from the measuring and evaluation unit change the shape of the evaluation signal and lead to an angular deviation in the evaluation signal, and - to minimize the angular deviation, offset values ​​of the measuring signals are changed, characterized by that the offset values ​​are determined by determining a first harmonic in the evaluation signal and / or a second harmonic in the signals and / or density differences of measuring points of a Lissajous plot. [2] Method according to claim 1, characterized by that after calculating the angle values ​​using the arctangent, the offset values ​​are determined from the amplitude and phase of the first harmonic of the resulting angular deviation of a 360° rotation. [3] Method according to claim 1, characterized by that in order to determine the offset values ​​from the density differences of the measuring points of the Lissajous plot, the measuring points are recorded at a constant period velocity with a fixed sampling rate. [4] Method according to claim 1, characterized bythat the offset values ​​are determined from the component of the first harmonic calculated by DFT or FFT from the course of the error of the rotation angle of a 360° revolution. [5] Method according to claim 1, characterized by that offset values ​​are determined by changing correction values ​​already generated by other methods and the second harmonics of the sinusoidal and cosinusoidal signals calculated by DFT or FFT. [6] Method according to claim 4 or 5, characterized by that the DFT calculation is carried out using a Goertzel algorithm or the first harmonic or the second harmonic is determined directly using a sinusoidal fitting function. [7] Method according to one of the preceding claims, characterized by that the determination of offset values ​​is carried out several times, in an iterative manner. [8] Method according to one of the preceding claims, characterized bythat after correction values ​​have already been determined, a second harmonic is determined directly at the signals from a comparison of the signal curve with an ideal sine curve and from a comparison of the signal curve with an ideal cosine curve, and the offset values ​​to be set are determined from the correction values ​​already determined and the deviation from the ideal curves. [9] Method according to one of the preceding claims, characterized by that the magnetic field sensors (S1, S2) are spaced apart along an X-direction or the magnetic field sensors (S1, S2) are arranged one above the other in a Z-direction. [10] Use of the method according to one of the preceding claims for determining angle of rotation values, comprising - an axle with an end face and at least one permanent magnet arranged on the end face or laterally on the axle, - at least two magnetic field sensors (S1, S2) arranged along an imaginary extension of the axis or magnetic field sensors arranged laterally to the axis, wherein a distance is formed between the axis and the magnetic field sensors (S1, S2), - a measuring and evaluation device connected to the magnetic field sensors (S1, S2), wherein - when the axis rotates, a sinusoidal signal or a cosinusoidal signal is present at an output of the respective magnetic field sensor, the measuring and evaluation device is set up to calculate an evaluation signal corresponding to the rotation angle values ​​from the signals by means of an arctangent function in order to minimize the angular deviation. [11] Use of the method according to one of the preceding claims 1 to 8 for determining angle values ​​of a translational movement path comprising - a plurality of permanent magnet pairs arranged along the translational path, - at least two magnetic field sensors arranged along the translational movement path, which are spaced apart from the permanent magnet pairs and are arranged to be movable along the translational movement path, - a measuring and evaluation device connected to the magnetic field sensors (S1, S2), wherein - during a translational movement of the permanent magnet pairs or the magnetic field sensors (S1, S2) along the path, a sinusoidal signal or a cosinusoidal signal is present at an output of the respective magnetic field sensor, the measuring and evaluation device is designed to calculate an evaluation signal corresponding to the angle values ​​from the signals by means of an arctangent function in order to minimize the angular deviation. [12] Use of the method according to claim 10 or claim 11, characterized bythat the measuring and evaluation device is set up to determine offset values ​​using a DFT or FFT evaluation. [13] Use of the method according to any one of claims 10 to 12, characterized by that the measuring and evaluation device includes a Goertzel filter or the measuring and evaluation device is set up to perform a sinusoidal adaptation function. [14] Use of the method according to claim 10 or claim 11, characterized by that the measuring and evaluation device is set up to determine density differences between the measuring points of a Lissajous plot. [15] Use of the method according to one of the preceding claims, characterized by that the measuring and evaluation device with the magnetic field sensors is integrated monolithically or as a multi-chip solution in an IC housing

Images