This application claims priority to German Patent Application No. 102019126430.8 filed on Oct. 1, 2019, the content of which is incorporated by reference herein in its entirety.
In many applications it is not an option to directly access the end of a shaft for angle measurement purposes. As a solution, a magnetic 3D sensor can be used by measuring the X-Y (X-Z or Y-Z) components. The sensor is located out of the shaft, e.g. not on the rotation axis of the shaft. The shaft itself has a magnetic encoder with at least two poles (one referred as N, the other as S).
Implementations described herein may improve existing solutions for sensing of a rotational angle of, e.g., a shaft.
The examples suggested herein may in particular be based on at least one of the following solutions. Combinations of the following features may be utilized to reach a result. The features of the method could be combined with any feature(s) of the device, apparatus or system or vice versa.
A rotation angle sensing device is suggested comprising:
The at least two magnetic field sensor elements are located out of shaft by not being rigidly attached to the shaft. The position of the at least two magnetic field sensor elements is fixed independently from the rotation of the shaft and the magnetic field source (which is coupled to the shaft).
The magnetization points in axial direction. Hence, the magnetized volume caused by the magnetic field source is not rotationally symmetric which results in a variation of the magnetic field in view of the rotation angle. This variation of the magnetic field is detected by the at least two magnetic field sensor elements.
According to an implementation, the magnetic field source is fixed on the shaft and the magnetic field source comprises at least one of the following:
The magnetic field source may be directly or indirectly, e.g., via a holder, fixed to the shaft.
According to an implementation, the magnetic field source comprises at least one region of homogeneous axial magnetization, which is not rotationally symmetric to the rotation axis.
According to an implementation, the magnetic field source has a homogenous magnetization in a direction along its thinnest dimension.
The thinnest dimension may in particular be the thickness of the magnetic field source. This allows magnetizing a large ring magnet such that it provides a substantially homogenous magnetic field.
For example, a permanent magnet with a diameter of more than 15 mm is difficult to magnetize such that it supplies a homogenous magnetic field. Therefore, it may be beneficial to use a ring magnet as magnetic field source, which is magnetized along its thickness. The thickness may be in a range between 1 mm and 20 mm.
This approach is substantially different from prior art solutions, which magnetize permanent magnets not along its thinnest dimension, but along its elongated direction.
According to an implementation, the at least two magnetic field sensor elements are located in the same plane, which is perpendicular to the rotation axis.
According to an implementation, at least three magnetic field sensor elements are substantially equally distributed across a circle around the rotation axis.
For example, N magnetic field sensor elements may be provided with N being at least three and the N magnetic field sensor elements are positioned on the circle around the rotation axis with an azimuthal spacing amounting to 360°/N.
According to an implementation, the at least two magnetic field sensor elements are mounted on the same component board.
According to an implementation, the component board has a central bore to accommodate the shaft and an aperture to mount the component board in a diametrical direction.
Advantageously, due to the aperture, there is no need to pull the component board over the end of the shaft and therefore to demount the shaft or conduct a costly (dis)assembly.
According to an implementation,
According to an implementation, the at least two magnetic field sensor elements are positioned on the mid-plane at a radial clearance to the magnetic field source.
According to an implementation, the center of gravity of the respective magnetic field sensor element is positioned on the mid-plane.
According to an implementation, at least two magnetic field sensor elements are positioned not on the mid-plane, but above or below this mid-plane at a radial distance from the rotation axis which is smaller than half of the maximum diameter of the magnetic field source, thereby having an axial clearance to the top and/or bottom surfaces of the rotatable magnet.
According to an implementation, the axial or radial clearance is between 1 mm and 5 mm.
According to an implementation, the magnetic field source has at least one of the following characteristics:
According to an implementation, the magnetic field source comprises
According to an implementation, the magnetic field source comprises
According to an implementation,
The rotational position of the shaft determined by the processing device may be an estimate of the true rotational position of the shaft.
According to an implementation, the processing device is arranged to determine a corrected rotation angle according to
φ″=φ′+ƒ(φ′),
wherein φ″ is the corrected rotation angle of the shaft, φ′ is the previously determined rotational angle and f(φ′) is a correction term, which considers N periods in the full 360° revolution of the shaft, with N being the number of the at least two magnetic field sensor elements.
The processing device may comprise at least one of the following: a processor, a microcontroller, a hard-wired circuit, an ASIC, an FPGA, a logic device.
According to an implementation, the magnetic field sensor element may be one of the following:
Also, a method is provided for sensing a rotational angle of a shaft that is arranged rotatably around a rotation axis, wherein a magnetic field source is capable of being mechanically coupled to a rotatable shaft, wherein the magnetic field source provides a magnetization which is substantially parallel to the rotation axis and not rotationally symmetric to the rotation axis, wherein the magnet field source comprises or is a permanent magnet, the method comprising:
According to an implementation, the method further comprises:
Implementations are shown and illustrated with reference to the drawings. The drawings serve to illustrate the basic principle, so that only aspects necessary for understanding the basic principle are illustrated. The drawings are not to scale. In the drawings the same reference characters denote like features.
The magnet 103 may preferably be homogeneously magnetized in axial direction. The center axis of the shaft 101 is identical with the rotation axis 107. The sensor elements 104 to 106 are placed symmetrical around the rotation axis, e.g. their radial distance Rs to the rotation axis is identical, they are positioned on a regular grid of 120° and they are positioned in an axial mid-plane of the magnet at z=0.
The sensor elements 104 to 106 each responds to an axial magnetic field component Bz. Each of the sensor elements may preferably comprise Hall plates. In such case, each of the sensor elements represent dies or chips of the semiconductor Hall plate.
The ring magnet 103 is preferably rotationally symmetric around its center axis and has an outer diameter amounting to 2*Rm. However, the center axis of the magnet 103 is mounted eccentrically to the rotation axis 107 (by an amount e).
The center axis of the magnet 103 and the rotation axis 107 are both parallel to each other. The magnet 103 is fixed to the shaft 101 via the holder 102 so that they rotate coherently.
When the shaft 101 rotates around its rotation axis 107, a clearance CL of each of the sensor elements 104 to 106 to the magnet 103 varies due to this eccentricity 303.
The Minimum Clearance Amounts To
CLmin=Rs−Rm−e
and the maximum clearance amount to
CLmax=Rs−Rm+e,
wherein Rs is the radial distance of the sensor element to the rotation axis 107, Rm is the outer radius of the ring magnet 103 and e is the eccentricity 303.
As the magnetic field Bz decays with an increasing distance from the magnet, the magnetic field fluctuates on each sensor element. The sensors 104 to 106 detect the magnetic field, which can then be combined according to
The two numbers CO and SI can be interpreted as ξ- and η-components of a point, which rotates coherently with the shaft 101. The angle of a pointer to any fixed reference axis in a (ξ, η)-space is an estimated rotation angle φ′. If the ξ-axis is chosen as reference axis, the following applies:
φ′=arctan 2(CO;SI),
wherein arctan 2 is the unique arctan over 360°. Usually, the rotation angle φ′ differs from a true rotation angle φ by an error, which stems from the nonlinear decay of the magnetic field Bz versus the radial distance. Hence, the sensor system may utilize a correction formula and/or a look-up table to improve the determination of the rotation angle. For example, a correction function f(φ′) may be applied on the estimated (potentially incorrect) rotation angle φ′ resulting in an improved (corrected) rotation angle φ″:
φ″=φ′+ƒ(φ′).
Example:
The magnet 103 has an inner diameter of 35 mm, an outer diameter of 50 mm and 7 mm thickness in axial direction. The magnet 103 is mounted on the shaft 101 with an eccentricity 303 amounting to 1 mm.
The sensor elements 104 to 106 are placed such that their sensitive spots are on a circle with 56 mm diameter around the rotation axis 107.
The systematic error hence varies between ±4.5°. This systematic error can be measured on prototype systems in a laboratory characterization, and it can be plotted as a function over φ′ (instead of φ). This function over φ′ can be parametrized with a small number of parameters. For example, it can be expanded into a Fourier series of the argument φ′ like
ƒ(φ′)=k·sin(3φ′),
wherein the number k may depend on
The number k may amount to, e.g., −4.5°.
Eventually, the corrected rotation angle
φ″=φ′+ƒ(φ′),
e.g.,
φ″=φ′+k·sin(3φ′)=−4.5°·sin(3φ′)
can be determined.
As an option, an offset-calibration may be applied for the signals CO and SI. In case the signals CO and SI have an offset, e.g. they are not centered around zero, the system may be configured to determine maxima and minima of the curves for several revolutions, and calculate the offset as
offset=½(max+min).
Then, this offset may be subtracted. Such offset-calibration may be used to further improve the accuracy of the rotation angle determined.
It is another option to conduct an amplitude-calibration: The system may thus compute the amplitudes
for both signals CO and SI. If the amplitudes are substantially different from each other (e.g., the difference is above a predefined threshold), the system may normalize the amplitudes to the same number by dividing each of the signals CO, SI by the value (max−min)/2 and by using the result the same way as the signals CO, SI without amplitude-calibration.
It is another option that an additional step is conducted prior to any other processing explained so far: The system may compute three signals U, V, W based on the signals Bz(0°), Bz(120°), Bz(−120°) as follows:
U=Bz(0°)−Bz(120°);
V=Bz(120°)−Bz(−120°);
W=Bz(−120°)−Bz(0°).
The advantage of using the signals U, V, W is that they are robust against interfering homogeneous external magnetic fields, because interferences are cancelled out by the subtractions. The system may use the signals U, V, W the same way as the signals Bz(0°), Bz(120°), Bz(−120°) have been used. Also, it is more robust to use the signals U, V, W for any kind of calibration.
Further Advantages and Implementations:
As shown in
However, the inner bore of the magnet may also be concentric to the rotation axis. In this case the holder does not need to provide the eccentricity, because the magnet does this already.
The axial thickness of the magnet may be significantly (e.g., two times) larger than an axial play plus any axial mounting tolerances between the sensor element and the magnet. In practice, the magnet may preferably be thicker than 3 mm if the sensor elements are placed in the mid-plane of the magnet.
It is favorable to place the sensor elements in the mid-plane of the magnet, because there the magnetic field shows the highest symmetry. Hence, the magnetic field changes little if the sensor element slightly moves in z-direction (up or down).
Alternative Implementation: Sensor Elements Below (or Above) the Magnet
This example allows the magnet 103 to be thin in axial direction. For example, a thickness of the magnet amounting to 1 mm may suffice.
It is an option that the magnet 103 is mounted to a supporting layer, e.g., a steel-back as it is known from multipole-encoder wheels. In case of a steel-back it is favorable that the steel-back is rotationally symmetric with respect to the rotation axis and to the sensor elements. Then a homogeneous external magnetic disturbance generates identical disturbance on all sensor elements, which are cancelled out in the CO and SI signals or also in the U,V,W signals.
If the sensor elements 104 to 106 are placed below the magnet 103, their radial distance to the rotation axis 107 may be smaller than the radial distance of the outer perimeter of the magnet 103 to the rotation axis 107. Hence, the sensor elements may have an axial clearance instead of a radial clearance.
If the magnet has an eccentricity, this may have the disadvantage of a large centrifugal force if the shaft rotates at high rpm.
Example: Eccentric Magnetization
This effect can be mitigated or avoided by utilizing a magnet that shows rotational symmetry with regard to the rotation axis, but shows an eccentric magnetization, e.g. has only eccentric parts magnetized. The remaining parts of the magnet may not be magnetized (or do show a significant degree of magnetization) such that the dominant contribution to the magnetic field supplied by the magnet stems from its magnetized eccentric portion.
Further Examples and Advantages
The magnet shown in
It is also an option to magnetize both portions 701 and 702 of the magnet in opposite axial directions. The portion 702 may thus be magnetized upwards and the portion 702 may be magnetized downwards.
The cross section of the ring magnet is rectangular and may have a constant axial thickness. It is also an option to vary the thickness to shape the field pattern Bz(φ) to become more similar to a sinusoidal shape, e.g. with no mean value and thus no offset, having no higher harmonics, only a fundamental frequency component that varies once over a range of 360°. Preferably, the shape of the magnet remains rotationally symmetric to avoid centrifugal forces acting on the magnet.
It is also an option that the geometry of the magnet deviates from any rotational symmetry shape in case such deviation does not create any or at least no significant centrifugal force. This may result in a criterion for the shape of the magnet that only the gravity center of the magnet is placed on the rotation axis. An example is an ellipsoidal magnet.
It is another option that an interface between the portion 701 and the portion 702, which may be parts with homogeneous magnetization in opposite axial direction, may be a cylinder with elliptical cross-section, whereby the cylinder axis is parallel to the rotation axis and shifted eccentrically to it by an amount ranging from 1 mm to 4 mm.
The magnet may comprise inexpensive material, e.g., ferrite, because the volume of the magnet is considerably large which on the other hand allows using material of rather weak magnetization.
In particular in combination with the supporting layer (e.g., steel-back) it is an option to use plastic-bound and/or rubber-bound magnetic layers as they are known, e.g., from multipole-encoder discs in ABS systems.
Although various example implementations of the implementation have been disclosed, it will be apparent to those skilled in the art that various changes and modifications can be made which will achieve some of the advantages of the implementation without departing from the spirit and scope of the implementation. It will be obvious to those reasonably skilled in the art that other components performing the same functions may be suitably substituted. It should be mentioned that features explained with reference to a specific figure may be combined with features of other figures, even in those cases in which this has not explicitly been mentioned. Further, the methods of the implementation may be achieved in either all software implementations, using the appropriate processor instructions, or in hybrid implementations that utilize a combination of hardware logic and software logic to achieve the same results. Such modifications to the inventive concept are intended to be covered by the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
102019126430.8 | Oct 2019 | DE | national |
Number | Name | Date | Kind |
---|---|---|---|
20050007104 | Lequesne | Jan 2005 | A1 |
20160138940 | Kurmaniak et al. | May 2016 | A1 |
20160216132 | Ausserlechner | Jul 2016 | A1 |
20190170497 | Ausserlechner | Jun 2019 | A1 |
20200271479 | Wang | Aug 2020 | A1 |
Number | Date | Country |
---|---|---|
102015115247 | Mar 2016 | DE |
1577182 | Sep 2005 | EP |
1339956 | Oct 1963 | FR |
Entry |
---|
Tumanski, “Induction Coil Sensors—a Review,” Mar. 2007 [Retrieved from http://www.tumanski.x.pl/coil.pdf]. |
“Out of Shaft with magnetic 3D sensor,” Infineon Technologies AG, Jul. 31, 2018 [Retrieved from https://www.infineon.com/dgdl/Infineon-Out_of_Shaft-AN-v01_00-EN.pdf?fileld=5546d46265257de801653898ba536074]. |
Number | Date | Country | |
---|---|---|---|
20210095999 A1 | Apr 2021 | US |