Torque sensor and vehicle steering device

Abstract

A torque sensor (1) is provided with a storage unit (47c) that stores a correction map indicating a relationship between a detection value output by a magnetic detection element (47a) and a correction value. A larger number of correction points are set in a first region (R1) and a fourth region (R4) than in a second region (R2) and a third region (R3) when a change range between a first detection value (Tdmax), which is detected when a relative rotation amount in a first direction between an input shaft (82a) and an output shaft (82b) reaches a maximum allowable amount, and a second detection value (Tdmin), which is detected when a relative rotation amount in a second direction therebetween reaches a maximum allowable amount, is equally divided into four regions, and the first region (R1) to the fourth region (R4) are defined in ascending order of the detection values.

Description

Torque sensor and vehicle steering device

[0001] The present invention relates to a torque sensor and a vehicle steering device.

[0002] Conventionally, a magnetic torque sensor including a torsion bar, a permanent magnet, and a magnetic sensor has been known. For example, Japanese Patent Application Laid-Open Publication No. 2003-129998 (Patent Document 1) describes a magnetic torque sensor in which a permanent magnet and a stator are fixed to both ends of a torsion bar, and input torque is detected by a magnetic detection element detecting the magnetic flux density flowing from the magnet to the stator. This magnetic torque sensor stores calibration data in a memory unit to correct a linearity error between the input torque and the detection value output by the magnetic detection element. The calibration data is composed of correction values ​​at multiple points (e.g., calibration points) within a range of possible detection values ​​output by the magnetic detection element. In the following description, points within the range of detection values ​​of the magnetic detection element for which correction values ​​are stored are referred to as "correction points."

[0003] Japanese Patent Application Laid-Open No. 2022-000621

[0004] The storage capacity of the storage unit is limited, and so is the number of correction points at which correction values ​​can be stored. Therefore, the correction points are set discretely. Therefore, correcting the detected value by interpolation between adjacent correction points can result in a decrease in linearity between the adjacent correction points. The present invention was devised in light of the above circumstances, and aims to accurately correct linearity errors in the detected value of a magnetic torque sensor.

[0005] In order to achieve the above object, a torque sensor according to one aspect of the present invention comprises an input shaft, an output shaft, a torsion bar having one end fixed to the input shaft and the other end fixed to the output shaft, a permanent magnet fixed to one of the input shaft and the output shaft, a stator having a first stator member and a second stator member fixed to the other of the input shaft and the output shaft facing the permanent magnet, a magnetic detection element that outputs a detection value corresponding to the magnetic flux density of the magnetic flux flowing from the first stator member to the second stator member, a memory unit that stores a correction map that shows the relationship between the detection value and a correction value, and a correction circuit that corrects the detection value with the correction value and outputs an output signal.

[0006] The correction map stores the correspondence between detection values ​​and correction values ​​at three or more correction points, and divides the range of change between a first detection value, which is the detection value when the relative rotation amount in a first direction between the input shaft and the output shaft reaches the maximum allowable amount, and a second detection value, which is the detection value when the relative rotation amount in a second direction reaches the maximum allowable amount, into four equal regions.When the four regions are defined as a first region, a second region, a third region, and a fourth region in order of decreasing detection value, more correction points are set in the first and fourth regions than in the second and third regions.

[0007] A torque sensor according to another aspect of the present invention comprises an input shaft and an output shaft, a torsion bar having one end fixed to the input shaft and the other end fixed to the output shaft, a sensor shaft portion made of magnetic material and fixed to one of the input shaft and output shaft and having a plurality of axially extending ridges formed along the circumferential direction, a cylindrical member made of conductive and non-magnetic material and fixed to the other of the input shaft and output shaft so as to be arranged coaxially with the sensor shaft portion and having a window formed in a position opposite the ridges, a detection coil arranged to surround the window and outputting a detection value corresponding to inductance, a memory portion storing a correction map similar to that described above, and a correction circuit that corrects the detection value with the correction value and outputs an output signal.

[0008] A vehicle steering device according to yet another aspect of the present invention includes the torque sensor described above, an actuator that generates a steering reaction force to be applied to the steering system of the vehicle, and an actuator control unit that drives and controls the actuator in accordance with the detection result of the torque sensor.

[0009] According to the present invention, linearity errors in the detected values ​​of a magnetic torque sensor can be corrected with high precision.

[0010] 1 is a schematic diagram of an electric power steering device of an embodiment. FIG. 2 is an exploded perspective view of an electric power steering device of an embodiment. FIG. 3 is a cross-sectional view of the electric power steering device of an embodiment. FIG. 4 is an enlarged view of a portion of FIG. 4. FIG. 5 is an exploded perspective view showing a magnet, a stator, etc. of an embodiment. FIG. 6 is an explanatory diagram of an example of a circumferential positioning means between an output shaft and a second sleeve. FIG. 7 is a conceptual diagram of a linearity error of a torque sensor. FIG. 8 is a block diagram of an example of the functional configuration of a Hall IC. FIG. 9 is a table conceptually showing an example of calibration data. (a) and (b) are conceptual diagrams of reasons why a linearity error remains. (a) and (b) are conceptual diagrams of a linearity error when correction points are set at equal intervals, and (c) and (d) are conceptual diagrams of a linearity error when correction points are set at unequal intervals. FIG. 10 is a conceptual diagram of an example of setting of correction points in an embodiment. FIG. 11 is a schematic diagram of a configuration of a torque sensor of a modified example.

[0011] Embodiments of the present invention will be described in detail with reference to the drawings. Note that the embodiments of the present invention shown below are merely examples of devices and methods for embodying the technical concept of the present invention, and the technical concept of the present invention does not limit the configuration, arrangement, etc. of component parts to those described below. Various modifications can be made to the technical concept of the present invention within the technical scope defined by the claims.

[0012] (Configuration) Fig. 1 is a schematic diagram of an electric power steering device according to an embodiment, Fig. 2 is an exploded perspective view of the electric power steering device according to the embodiment, and Fig. 3 is a cross-sectional view of the electric power steering device according to the embodiment.

[0013] As shown in Fig. 1, the electric power steering device 80 includes, in the order in which force applied by the operator is transmitted, a steering wheel 81, a steering shaft 82, a steering force assist mechanism 83, a universal joint 84, an intermediate shaft 85, and a universal joint 86, which are joined to a pinion shaft 87. In the following description, the front of a vehicle in which the electric power steering device 80 is installed will simply be referred to as the front, and the rear of the vehicle will simply be referred to as the rear. Furthermore, as shown in Fig. 2, the electric power steering device 80 includes a gearbox 920, an intermediate plate 10, and a column housing 820. The gearbox 920 is attached to the vehicle, and the column housing 820 is fixed to the gearbox 920 via the intermediate plate 10.

[0014] As shown in FIGS. 1 and 3 , the steering shaft 82 includes an input shaft 82a, an output shaft 82b, and a torsion bar 82c. The input shaft 82a is supported by a column housing 820 shown in FIG. 3 via a bearing. The input shaft 82a can rotate relative to the column housing 820. One end of the input shaft 82a is connected to the steering wheel 81. The other end of the input shaft 82a is connected to the torsion bar 82c. The torsion bar 82c is fitted into a hole provided in the center of the input shaft 82a and is fixed to the input shaft 82a via a pin. In the following description, a direction parallel to the rotation axis Z of the input shaft 82a will be referred to as the axial direction. A direction perpendicular to the rotation axis Z and parallel to a line passing through the rotation axis Z will be referred to as the radial direction. A direction along the circumference of a circle centered on the rotation axis Z will be referred to as the circumferential direction.

[0015] 3, the output shaft 82b is supported by the intermediate plate 10 via a bearing 71, and is supported by the gear box 920 via a bearing 72. For example, the bearing 71 is press-fitted into the intermediate plate 10, and the bearing 72 is press-fitted into the gear box 920. The output shaft 82b can rotate relative to the intermediate plate 10 and the gear box 920. One end of the output shaft 82b is connected to a torsion bar 82c. The other end of the output shaft 82b is connected to a universal joint 84. The torsion bar 82c is press-fitted into a hole provided in the center of the output shaft 82b, and is thereby fixed to the output shaft 82b.

[0016] The front end of the input shaft 82a is located inside the output shaft 82b. A convex portion provided on one of the outer circumferential surface of the input shaft 82a and the inner circumferential surface of the output shaft 82b fits into a concave portion provided on the other. A circumferential gap is provided between the convex portion and the concave portion. This allows torque to be transmitted between the input shaft 82a and the output shaft 82b even if the torsion bar 82c no longer functions as a connecting member.

[0017] The input shaft 82a and the output shaft 82b are provided with anti-rotation devices (not shown). The anti-rotation devices may include, for example, a plurality of (e.g., eight) protrusions provided near the end of the input shaft 82a and protruding radially outward from the end of the input shaft 82a, and a plurality of (e.g., eight) protrusions provided near the end of the output shaft 82b and protruding radially outward from the end of the output shaft 82b. Relative rotation between the input shaft 82a and the output shaft 82b is restricted by the abutment of these protrusions.

[0018] As shown in FIG. 1 , the intermediate shaft 85 connects the universal joint 84 and the universal joint 86. One end of the intermediate shaft 85 is connected to the universal joint 84, and the other end is connected to the universal joint 86. One end of the pinion shaft 87 is connected to the universal joint 86, and the other end of the pinion shaft 87 is connected to the steering gear 88. The universal joints 84 and 86 are, for example, Cardan joints. The rotation of the steering shaft 82 is transmitted to the pinion shaft 87 via the intermediate shaft 85. That is, the intermediate shaft 85 rotates together with the steering shaft 82.

[0019] As shown in Figure 1, the steering gear 88 includes a pinion 88a and a rack 88b. The pinion 88a is connected to the pinion shaft 87. The rack 88b meshes with the pinion 88a. The steering gear 88 converts the rotational motion transmitted to the pinion 88a into linear motion by the rack 88b. The rack 88b is connected to a tie rod 89. The angle of the wheels changes as the rack 88b moves.

[0020] As shown in FIG. 1, the steering force assist mechanism 83 includes a reduction gear 92 and an electric motor 93. The reduction gear 92 is, for example, a worm reduction gear, and includes a gear box 920, a worm wheel 921, and a worm 922, as shown in FIGS. 2 and 3. Torque generated by the electric motor 93 is transmitted to the worm wheel 921 via the worm 922, causing the worm wheel 921 to rotate. The worm 922 and the worm wheel 921 increase the torque generated by the electric motor 93. The worm wheel 921 is fixed to the output shaft 82b. For example, the worm wheel 921 is press-fitted into the output shaft 82b. Therefore, the reduction gear 92 applies an assisting steering torque to the output shaft 82b.

[0021] As shown in Fig. 1, the electric power steering device 80 includes a controller 90, which is an electronic control unit (ECU), a torque sensor 1, and a vehicle speed sensor 95. The electric motor 93, the torque sensor 1, and the vehicle speed sensor 95 are electrically connected to the controller 90. The torque sensor 1 outputs the steering torque transmitted to the input shaft 82a to the controller 90 via communication. The vehicle speed sensor 95 detects the traveling speed (vehicle speed) of the vehicle body on which the electric power steering device 80 is mounted. The vehicle speed sensor 95 is provided on the vehicle body, and outputs the vehicle speed to the controller 90 via communication.

[0022] The controller 90 controls the operation of the electric motor 93. The controller 90 acquires signals from the torque sensor 1 and the vehicle speed sensor 95. When the ignition switch 98 is on, the controller 90 is supplied with power from a power supply device 99 (e.g., an on-board battery). The controller 90 calculates an auxiliary steering command value based on the steering torque and the vehicle speed. The controller 90 adjusts the value of power supplied to the electric motor 93 based on the auxiliary steering command value. The controller 90 acquires information on the induced voltage of the electric motor 93 or information output from a resolver or the like provided in the electric motor 93. When the controller 90 controls the electric motor 93, the force required to operate the steering wheel 81 is reduced.

[0023] Fig. 4 is a cross-sectional view of the electric power steering device of this embodiment taken along a plane different from that of Fig. 3. Fig. 5 is an enlarged view of a portion of Fig. 4. Fig. 6 is an exploded perspective view showing the magnet, stator, etc. of this embodiment. As shown in Fig. 3, the torque sensor 1 is disposed between the column housing 820 and the gear box 920. More specifically, the torque sensor 1 is located in the space sandwiched between the column housing 820 and the intermediate plate 10.

[0024] As shown in FIGS. 3 to 5 , the torque sensor 1 includes a first sleeve 21, a magnet 25, a second sleeve 31, a carrier 32, a stator 35, a sensor housing 40, a magnetic flux collecting member 46, a printed circuit board 43, a Hall IC 47, a first cover 48, and a second cover 49. The first sleeve 21 is a non-magnetic metal. A specific example of the non-magnetic metal is austenitic stainless steel (SUS304). The first sleeve 21 is a cylindrical member and is attached to the input shaft 82a as shown in FIG. 3 . The first sleeve 21 is formed, for example, by deep drawing.

[0025] The magnet 25 includes a hard magnetic material. Specific examples of hard magnetic materials include neodymium and ferrite. The magnet 25 is formed, for example, by solidifying a material made by mixing magnetic powder and resin. The magnet 25 is called a bonded magnet. The magnet 25 is formed into a cylindrical shape using, for example, neodymium and polyphenylene sulfide, or ferrite and polyphenylene sulfide. In the magnet 25, south poles and north poles are arranged alternately in the circumferential direction. As shown in FIG. 5 , the magnet 25 is attached to the first sleeve 21 with an adhesive 27.

[0026] The second sleeve 31 is a non-magnetic metal. A specific example of the non-magnetic metal is austenitic stainless steel (SUS304). The second sleeve 31 is a cylindrical member and is attached to the output shaft 82b as shown in FIG. 3. Specifically, the second sleeve 31 is press-fitted onto the outer peripheral surface of the output shaft 82b. The carrier 32 is a non-magnetic member. For example, the carrier 32 is made of resin. Specific examples of resin include polybutylene terephthalate (PBT) and polyacetal resin (POM). The carrier 32 is a cylindrical member and is attached to the output shaft 82b via the second sleeve 31. The carrier 32 is integrally formed with the second sleeve 31 by injection molding.

[0027] As shown in FIG. 6 , the stator 35 includes a first stator member 351 and a second stator member 352. The first stator member 351 and the second stator member 352 are made of a soft magnetic material. A specific example of a soft magnetic material is a nickel-iron alloy. The first stator member 351 and the second stator member 352 are fixed to the carrier 32. The first stator member 351 and the second stator member 352 rotate together with the output shaft 82b, the second sleeve 31, and the carrier 32. The first stator member 351 includes a first flange portion 351a and a plurality of first teeth (tooth portions) 351b. The first flange portion 351a is a plate that is perpendicular to the axial direction. The first teeth 351b protrude forward from the first flange portion 351a. The plurality of first teeth 351b are arranged at equal intervals in the circumferential direction. For example, the first flange portion 351a and the first teeth portion 351b may be formed by bending a single sheet metal member.

[0028] The second stator member 352 includes a second flange portion 352a and a plurality of second teeth 352b. The second flange portion 352a is a plate parallel to the first flange portion 351a and is located in front of the first flange portion 351a. The second teeth 352b protrude rearward from the second flange portion 352a. The plurality of second teeth 352b are arranged at equal intervals in the circumferential direction. For example, the second flange portion 352a and the second teeth 352b may be formed by bending a single sheet metal member. Each second tooth 352b is located between two first teeth 351b. In other words, the first teeth 351b and the second teeth 352b are arranged alternately in the circumferential direction. The first teeth 351b and the second teeth 352b face the magnet 25.

[0029] Furthermore, the circumferential positional relationship between the output shaft 82b and the stator 35 is such that, in a neutral state where no input torque is applied to the torque sensor 1 (i.e., a state where the torsion angle of the torsion bar 82c is approximately zero), the circumferential center positions of the first teeth portion 351b and the second teeth portion 352b coincide with the boundary between the south pole and north pole of the magnet 25, and the magnetic flux of the magnet 25 does not flow from the first stator member 351 to the second stator member 352, or from the second stator member 352 to the first stator member 351.

[0030] See Fig. 7. For example, the circumferential positional relationship between the output shaft 82b and the stator 35 may be determined by engaging a convex portion and a concave groove formed on the inner peripheral surface of the second sleeve 31 and the outer peripheral surface of the output shaft 82b. In the example of Fig. 7, the inner peripheral surface of the second sleeve 31 is formed with a convex portion 311 by embossing, and the outer peripheral surface of the output shaft 82b is formed with a concave groove 82b1 extending along the axial direction, and the circumferential positional relationship between the output shaft 82b and the stator 35 is determined by engaging the convex portion 311 and the concave groove 82b1.

[0031] The sensor housing 40 is made of a non-magnetic material. For example, the sensor housing 40 is made of resin. Specific examples of resin include polybutylene terephthalate (PBT) and polyamide 66. The sensor housing 40 is fixed to the intermediate plate 10 by fastening members such as bolts.

[0032] As shown in FIG. 5 , the magnetic flux collector 46 includes a first magnetic flux collector 461 and a second magnetic flux collector 462. The first magnetic flux collector 461 and the second magnetic flux collector 462 are made of a soft magnetic material, such as a nickel-iron alloy. The first magnetic flux collector 461 and the second magnetic flux collector 462 are fixed to the sensor housing 40. As shown in FIG. 5 , the first magnetic flux collector 461 faces the first flange portion 351a. A gap C2 is provided between the first magnetic flux collector 461 and the first flange portion 351a. The first magnetic flux collector 461 is magnetized in response to the magnetization of the first stator member 351. The second magnetic flux collector 462 faces the second flange portion 352a. A gap C3 is provided between the second magnetic flux collector 462 and the second flange portion 352a. The second magnetic flux collector 462 is magnetized in response to the magnetization of the second stator member 352. For example, the axial length of the gap C3 is approximately equal to the axial length of the gap C2.

[0033] The printed circuit board 43 is fixed to the sensor housing 40. The Hall IC 47 is attached to the printed circuit board 43. The Hall IC 47 is disposed between the first magnetic flux collecting member 461 and the second magnetic flux collecting member 462. There are gaps between the Hall IC 47 and the first magnetic flux collecting member 461 and between the Hall IC 47 and the second magnetic flux collecting member 462. The Hall IC 47 changes the signal it outputs in response to changes in the magnetic flux density between the first magnetic flux collecting member 461 and the second magnetic flux collecting member 462. The Hall IC 47 outputs the signal to the controller 90.

[0034] When the steering wheel 81 is operated, torque is transmitted to the input shaft 82a. Because the output shaft 82b is connected to the input shaft 82a via the torsion bar 82c, the input shaft 82a rotates relative to the output shaft 82b. This causes the magnet 25 to rotate relative to the first teeth 351b and the second teeth 352b. This changes the magnetization strength of the first stator member 351 and the second stator member 352. This changes the magnetic flux density between the first magnetic flux collector 461 and the second magnetic flux collector 462. The Hall IC 47 detects this change in magnetic flux density. The controller 90 controls the electric motor 93 using the steering torque calculated based on the output signal of the Hall IC 47.

[0035] The first cover 48 is made of a non-magnetic material. For example, the first cover 48 is made of resin. Specific examples of resin include polybutylene terephthalate (PBT) and polyamide 66. As shown in FIG. 4 , the first cover 48 is attached to the rear end of the sensor housing 40. The first cover 48 covers the printed circuit board 43.

[0036] The second cover 49 is made of a non-magnetic material. For example, the second cover 49 is made of resin. Specific examples of resin include polybutylene terephthalate (PBT) and polyamide 66. As shown in FIG. 4 , the second cover 49 is attached to the front end of the sensor housing 40. As shown in FIG. 5 , the second cover 49 includes an annular main body 491 and a plurality of claws 492. The claws 492 are arranged at equal intervals in the circumferential direction. The claws 492 protrude forward from the main body 491. The claws 492 are inserted into the intermediate plate 10 by light press-fitting and are in contact with the inner circumferential surface of the intermediate plate 10.

[0037] The first sleeve 21 does not necessarily have to be attached to the input shaft 82a. For example, the first sleeve 21 and the magnet 25 may be attached to the output shaft 82b, and the second sleeve 31 and the stator 35 may be attached to the input shaft 82a. Even when the first sleeve 21 is attached to the output shaft 82b, it is press-fitted onto the outer peripheral surface of the output shaft 82b.

[0038] Next, we will explain the linearity error of the torque sensor 1. It is desirable that the output torque Tout, which is the output signal of the torque sensor 1, be proportional to the input torque Tin input to the torque sensor 1. However, the strength of the radial component of the magnetic flux density of the magnet 25 varies in a substantially sinusoidal manner, reaching maximum values ​​at the north and south poles, and therefore increases nonlinearly as it approaches the magnetic poles. Therefore, the magnetic flux density detected by the magnetic detection element of the Hall IC 47 varies nonlinearly with changes in the input torque Tin input to the torque sensor 1 (i.e., changes in the relative position between the magnet 25 and the stator 35 in the circumferential direction). As a result, the detection value of the magnetic detection element of the Hall IC 47 is not proportional to the input torque Tin (i.e., has a linearity error). In FIG. 8 , the solid line L0 represents the input / output characteristic curve of the detection value Td of the magnetic detection element, and the dashed-dotted line Lid represents the ideal straight line, which is the ideal input / output characteristic of the torque sensor 1 before origin correction. The error between the input / output characteristic curve L0 and the ideal straight line Lid is the linearity error.

[0039] As described above, the input shaft 82a and the output shaft 82b are provided with detents, which limit the amount of relative rotation between the input shaft 82a and the output shaft 82b (i.e., the torsion angle of the torsion bar 82c) to less than or equal to the maximum allowable amount. The first detection value Tdmax is the detection value of the magnetic detection element when the amount of relative rotation in the first direction between the input shaft 82a and the output shaft 82b reaches the maximum allowable amount, and the second detection value Tdmin is the detection value of the magnetic detection element when the amount of relative rotation in the first direction between the input shaft 82a and the output shaft 82b reaches the maximum allowable amount. The input torques Timax and Timin are the input torque Ti when the amount of relative rotation in the first direction and the second direction between the input shaft 82a and the output shaft 82b reach the maximum allowable amount, respectively.

[0040] In order to improve the linearity of the detection value Td of the magnetic detection element, the Hall IC 47 stores calibration data for correcting linearity errors. Fig. 9 is a block diagram showing an example of the functional configuration of the Hall IC 47. The Hall IC 47 includes a Hall sensor 47a, which is a magnetic detection element, a detection circuit 47b, a storage unit 47c, and a subtractor 47d. The detection circuit 47b converts the analog signal output from the Hall sensor 47a in accordance with the magnetic flux density into a digital signal.

[0041] The storage unit 47c stores, as calibration data, correspondence relationships between the detected value Td and the correction value dc at a plurality of correction points in the range of the detected value Td (i.e., the range from the first detected value Tdmax to the second detected value Tdmin). The circle plots in Fig. 8 represent the correction points set for the detected value Td of the value td. Similarly, the correction points are represented by circle plots in Figs. 11(a) and 11(b), 12(a) and 12(b), and 13.

[0042] FIG. 10 is a table conceptually illustrating an example of calibration data. The calibration data is data that represents a one-to-one correspondence between a detected value Td (detected value td in the example of FIG. 9 ) that serves as a correction point and a correction value dc for this detected value td. For example, the calibration data may be a correction map in which the detected value Td and the correction value dc are stored in correspondence with each other. The correction value dc may be obtained by calculating in advance the difference between the detected value td at each correction point and the ideal line Lid. See FIG. 9 . The subtractor 47d corrects the detected value Td by subtracting the correction value dc from the detected value Td output by the detection circuit 47b, and outputs the difference (Td - dc) as the output torque Tout of the Hall IC 47.

[0043] However, even if the detected value Td is corrected using such calibration data, the linearity error in the corrected output torque Tout may not be zero. In particular, a large linearity error may remain in the portion of the input-output characteristic curve L0 in FIG. 8 with a large curvature. FIGS. 11(a) and 11(b) are conceptual diagrams illustrating why a linearity error remains in the corrected output torque Tout. Note that FIGS. 11(a) and 11(b) are partial illustrations of the input-output characteristic curve between the detected value Td and the input torque Tin before correction and the input torque Tin after correction. The dashed line indicates the input-output characteristic curve between the detected value Td and the input torque Tin before correction with the correction value dc, the solid line indicates the input-output characteristic curve between the corrected output torque Tout and the input torque Tin, and the dashed-dotted line Lid indicates an ideal straight line. In other words, the detection value Td before correction is offset by the correction amounts dc1 and dc2 at the correction points td1 and td2. Note that between the correction points td1 and td2, the detection value Td before correction is offset by the correction amounts obtained by linearly interpolating dc1 and dc2.

[0044] As described above, the storage capacity of the storage unit 47c is limited, and the number of correction points at which correction values ​​can be stored is also limited. Therefore, the correction points are set discretely, and in the section between adjacent correction points td1 and td2, the correction value is determined by interpolating the correction values ​​dc1 and dc2 corresponding to the correction points td1 and td2, respectively. For example, when the detection value Td output by the detection circuit 47b is greater than the correction point td1 but smaller than the correction point td2 (td1<Td<td2), the detection value Td is corrected by the correction value (dc1+(Td-td1)×(dc2-dc1) / (td2-td1)) to obtain the output torque Tout.

[0045] Therefore, in the section between the adjacent correction points td1 and td2, the input / output characteristic curve of the pre-correction detection value Td (dashed line) is offset by the correction values ​​dc1 and dc2 while retaining the curvature, forming the input / output characteristic curve of the post-correction output torque Tout (solid line). As a result, as shown in Figures 11(a) and 11(b), linearity errors ea and eb remain in the section between the adjacent correction points td1 and td2, and when the curvature of the section between the correction points td1 and td2 is large, the linearity error eb becomes larger than the linearity error ea when the curvature is small. In other words, in the section where the curvature of the input / output characteristic curve of the pre-correction detection value Td is large, a larger linearity error remains in the post-correction output torque Tout.

[0046] Therefore, in the present invention, the number of correction points set in sections with small curvature of the input / output characteristic curve of the uncorrected detected value Td (i.e., sections with good linearity) is reduced, and the number of correction points set in sections with large curvature (i.e., sections with poor linearity) is increased, thereby suppressing an increase in linearity error in sections with large curvature. A method for setting correction points in the present invention will be explained schematically with reference to Figures 12(a) to 12(d). In Figures 12(a) to 12(d), the solid line L0 represents the input / output characteristic curve of the uncorrected detected value Td, the solid line L1 represents the input / output characteristic curve of the corrected output torque Tout, and the dashed dotted line Lid represents an ideal straight line.

[0047] 12(a) and 12(b) are schematic diagrams illustrating the results of correction when correction points are set so that the input torques Ti1, Ti2, Ti3, and Ti4 at the correction points are equally spaced. In the input torque range from Ti1 to Ti3, the curvature of the input / output characteristic curve L0 of the detected value Td before correction is small and the linearity is good. On the other hand, in the input torque range from Ti3 to Ti4, the curvature of the input / output characteristic curve L0 is large and the linearity is poor. Therefore, the linearity of the input / output characteristic curve L1 of the output torque Tout after correction is good in the range from Ti1 to Ti3, but is poor in the range from Ti3 to Ti4, resulting in a large linearity error e.

[0048] Here, we focus on a correction point within a section where the uncorrected detected value Td originally has good linearity (for example, correction point Pc1 where the input torque is Ti2). Because the linearity is good in the range before and after correction point Pc1, the linearity of the corrected input / output characteristic curve L1 will be good in the section Ti1 to Ti3 regardless of whether correction point Pc1 is set. In other words, omitting correction point Pc1 will not worsen the linearity error.

[0049] 12(c) and 12(d), the correction point Pc1 in the section Ti1-Ti3 is omitted, and a new correction point Pc2 is added at the point where the input torque is Tiα in the section Ti3-Ti4, which has a large curvature. Adding the new correction point Pc2 reduces the linearity error e in the section Ti3-Ti4. This makes it possible to improve the linearity of the corrected output torque Tout without changing the total number of correction points for which the correction values ​​tc are stored in the memory unit 47c.

[0050] FIG. 13 is a conceptual diagram of an example of how correction points are set in this embodiment. Due to the characteristic that the magnetic flux density generated by the magnet 25 changes in a substantially sinusoidal manner, the curvature of the input / output characteristic curve of the detected value Td output from the Hall sensor 47a is greater in the range where the absolute value of the detected value Td is large than in the range where the absolute value of the detected value Td is small. In other words, the curvature of the input / output characteristic curve of the detected value Td is greater in the range close to the first detected value Tdmax and the second detected value Tdmin than in the range close to the point where the detected value Td is zero. Therefore, correction points are set so that the density of correction points in the range where the absolute value of the detected value Td is large is higher than the density of correction points in the range where the absolute value of the detected value Td is small. As shown in FIG. 13 , the ideal line Lid may be set as a line connecting the first detected value Tdmax and the second detected value Tdmin. Alternatively, Lid may be defined as a line passing through a predetermined point and having a predetermined slope. For example, if the detected value Td is Tdn when the input torque Tin is zero (Tin=0), a straight line that passes through the points Tin=0 and Td=Tdn and has a predetermined slope may be defined as Lid.

[0051] Specifically, the range of change between the first detected value Tdmax and the second detected value Tdmin (i.e., the range of the detected value Td) is divided equally into four regions, and these four regions are defined as a first region R1, a second region R2, a third region R3, and a fourth region R4 in order of decreasing detected value Td. More correction points are set in the first region R1 and the fourth region R4 than in the second region R2 and the third region R3.

[0052] Ideally, the input / output characteristic curve of the detected value Td should be such that the detected value Td is zero when the input torque Tin is zero. However, in reality, the detected value Td may not be zero when the input torque Tin is zero. In the example of FIG. 13 , when the input torque Tin is zero, the detected value Td has a positive value Tdn≠0. In the following description, the fact that the detected value Td is not zero when the input torque Tin is zero is referred to as "origin deviation." Origin deviation occurs mainly for the following reasons.

[0053] The circumferential positional relationship between the output shaft 82b and the stator 35 described above is determined so that when the input torque Tin is zero, the magnetic flux of the magnet 25 does not flow from the first stator member 351 to the second stator member 352 or from the second stator member 352 to the first stator member 351. In other words, it is determined so that the Hall IC 47 does not detect the magnetic flux between the first magnetic flux collecting member 461 and the second magnetic flux collecting member 462 when the input torque Tin is zero.

[0054] However, it is difficult to achieve such an ideal positional relationship, and in practice, positional deviation may occur due to mechanical errors, etc. For example, if the circumferential positional relationship between the output shaft 82b and the stator 35 is determined by the engagement between the embossed convex portion 311 formed on the second sleeve 31 and the concave groove 82b1 formed on the output shaft 82b as described above, positional deviation may occur due to processing errors in the embossing. As a result, when magnetic flux is generated between the first magnetic flux collecting member 461 and the second magnetic flux collecting member 462 when the input torque Tin is zero, the Hall IC 47 detects this magnetic flux, and therefore the detected value Td does not become zero when the input torque Tin is zero, resulting in a deviation from the origin.

[0055] When the origin shift occurs in this way, a region with large curvature (a region with poor linearity) occurs in either the first detection value Tdmax or the second detection value Tdmin. In the example of Fig. 13, a region with large curvature occurs in the first detection value Tdmax. Furthermore, it is difficult to control the direction of the origin shift (i.e., whether the detection value Td is positive or negative when the input torque Tin is zero).

[0056] Therefore, the number of correction points set in the first region R1 and the fourth region R4 may be set to be approximately the same (or exactly the same). Even if a region with poor linearity occurs in either the first detection value Tdmax or the second detection value Tdmin, the linearity error in this region can be corrected using many correction points. This reduces the linearity error remaining after correction.

[0057] Furthermore, the correction point Pn may be set to the detection value Tdn that is output when the input torque Tin is approximately zero (when the torsion angle of the torsion bar 82c is approximately zero). When the input torque Tin is approximately zero, a minute input torque Tin is input when the steering wheel 81 starts to move. Therefore, by setting the correction point Pn to the detection value Tdn when the input torque Tin is approximately zero, the minute input torque Tin when the steering wheel 81 starts to move can be detected with high accuracy.

[0058] Furthermore, correction points may be set near the first detection value Tdmax and near the second detection value Tdmin. The linearity error is likely to be greatest either near the first detection value Tdmax or near the second detection value Tdmin. By setting correction points near the first detection value Tdmax and near the second detection value Tdmin, it is possible to prevent a large linearity error from remaining after correction. Note that the vicinity of the first detection value Tdmax may be, for example, a range from 95% to 100% of the first detection value Tdmax, and the vicinity of the second detection value Tdmin may be, for example, a range from 95% to 100% of the second detection value Tdmin.

[0059] For example, one correction point Pn may be provided at a point where the input torque Tin is approximately zero, four correction points may be provided in a region where the input torque Tin is greater than 0, and four correction points may be provided in a region where the input torque Tin is less than 0. Of the four correction points provided in the region where the input torque Tin is greater than 0, three may be provided in a fourth region, and the remaining one may be provided in a third region. Of the four correction points provided in the region where the input torque Tin is less than 0, three may be provided in a first region, and the remaining one may be provided in a second region.

[0060] (Effects of the embodiment) (1) The torque sensor 1 includes an input shaft 82 a and an output shaft 82 b, a torsion bar 82 c having one end fixed to the input shaft 82 a and the other end fixed to the output shaft 82 b, a magnet 25 fixed to one of the input shaft 82 a and the output shaft 82 b, a stator 35 including a first stator member 351 and a second stator member 352 fixed to the other of the input shaft 82 a and the output shaft 82 b facing the magnet 25, a Hall sensor 47 a that outputs a detection value corresponding to the magnetic flux density of the magnetic flux flowing from the first stator member 351 to the second stator member 352, a memory unit 47 c that stores a correction map showing the relationship between the detection value and a correction value, and a correction circuit that corrects the detection value with the correction value and outputs an output signal.

[0061] The correction map stores correspondence relationships between detection values ​​and correction values ​​at three or more correction points, and divides the range of change between a first detection value Tdmax, which is the detection value when the amount of relative rotation in the first direction between the input shaft 82 a and the output shaft 82 b reaches the maximum allowable amount, and a second detection value Tdmin, which is the detection value when the amount of relative rotation in the second direction reaches the maximum allowable amount, into four equal regions. If the four regions are defined as a first region R1, a second region R2, a third region R3, and a fourth region R4 in order of decreasing detection value, more correction points are set in the first region R1 and the fourth region R4 than in the second region R2 and the third region R3. This makes it possible to reduce linearity errors remaining after correction in regions with large curvatures of the input / output characteristic curve of the detection value with respect to the input torque, thereby enabling accurate correction of linearity errors in the detection value of the torque sensor 1.

[0062] (2) The first and fourth regions may have approximately the same number of correction points. Even if a region with poor linearity occurs in either the first detection value Tdmax or the second detection value Tdmin, the linearity error in this region can be corrected using many correction points. This reduces the linearity error remaining after correction.

[0063] (3) A correction point may be set to the detection value output when the torsion angle of the torsion bar 82c is approximately zero. When the torsion angle of the torsion bar 82c is approximately zero, a minute input torque when the steering wheel 81 starts to move is input to the torque sensor 1. Therefore, by setting a correction point to the detection value when the torsion angle of the torsion bar 82c is approximately zero, it is possible to accurately detect the minute input torque when the steering wheel 81 starts to move.

[0064] (4) Correction points may be set near the first detection value Tdmax and near the second detection value Tdmin. The linearity error is likely to be greatest near either the first detection value Tdmax or the second detection value Tdmin. By setting correction points near the first detection value Tdmax and near the second detection value Tdmin, it is possible to prevent a large linearity error from remaining after correction.

[0065] (5) The stator 35 may include a flange portion and teeth formed by bending a single sheet metal member. When forming the teeth by bending in this manner, the machining accuracy of the teeth tends to be lower than when cutting, and origin deviation, which is a condition in which the detected value Td does not become zero when the input torque Tin is zero, tends to occur. According to the embodiment, even if origin deviation occurs and a region of poor linearity occurs in either the first detected value Tdmax or the second detected value Tdmin, the linearity error in this region can be corrected at many correction points. This reduces the linearity error remaining after correction.

[0066] (6) The stator 35 is fixed to the output shaft 82b via the second sleeve 31. The inner peripheral surface of the second sleeve 31 is formed with a convex portion by embossing, and the outer peripheral surface of the output shaft 82b is formed with a concave groove along the axial direction, with the convex portion engaging the concave groove. If the circumferential positional relationship between the stator 35 and the output shaft 82b is determined by such engagement of the convex portion and the concave groove, forming the convex portion by embossing tends to result in lower machining accuracy than by cutting, making origin misalignment more likely to occur. According to this embodiment, even if origin misalignment occurs and a region of poor linearity occurs in either the first detection value Tdmax or the second detection value Tdmin, the linearity error in this region can be corrected using many correction points. This reduces the linearity error remaining after correction.

[0067] (Modifications) In the above embodiment, the present invention is applied to a torque sensor including a permanent magnet, a stator having a first stator member and a second stator member, and a magnetic detection element that outputs a detection value corresponding to the magnetic flux density of the magnetic flux flowing from the first stator member to the second stator member, but the present invention is not limited to this. The present invention can be applied to correcting linearity errors in the detection values ​​of various types of torque sensors.

[0068] Fig. 14 is a schematic diagram of the configuration of a torque sensor according to a modified example. The torque sensor 100 according to the modified example includes a sensor shaft portion 111 formed on the input shaft 82a, a pair of detection coils 113a and 113b arranged inside a column housing 820, and a cylindrical member 112 arranged between the two. The sensor shaft portion 111 is made of a magnetic material, and on the surface of the sensor shaft portion 111, as shown in Fig. 14, a plurality of axially extending ridges 111a (nine in the example of Fig. 14) are formed at equal intervals along the circumferential direction. In addition, grooves 111b are formed between the ridges 111a.

[0069] A cylindrical member 112 made of a conductive and non-magnetic material, such as aluminum, is arranged coaxially around the sensor shaft portion 111 and close to the sensor shaft portion 111, and the cylindrical member 112 is fixed to the output shaft 82b. In this modified example, the sensor shaft portion 111 is fixed to the input shaft 82a and the cylindrical member 112 is fixed to the output shaft 82b, but the cylindrical member 112 may be fixed to the input shaft 82a and the sensor shaft portion 111 may be fixed to the output shaft 82b.

[0070] The cylindrical member 112 is provided with a first window row consisting of a plurality of (nine in the example of FIG. 14 ) rectangular windows 112a equally spaced in the circumferential direction at a position facing the ridge 111a on the surface of the sensor shaft portion 111, and a second window row consisting of a plurality of (nine in the example of FIG. 14 ) rectangular windows 112b, which have the same shape as the windows 112a but are out of phase with each other in the circumferential direction, at a position axially offset from the first window row. The outer periphery of the cylindrical member 112 is surrounded by yokes 115a and 115b, which hold coil bobbins 118 wound with detector coils 113a and 113b of the same standard. That is, the detector coils 113a and 113b are arranged coaxially with the cylindrical member 112, with the detector coil 113a surrounding the first window row portion consisting of the windows 112a, and the detector coil 113b surrounding the second window row portion consisting of the windows 112b.

[0071] The output lines of the detection coils 113a and 113b are connected to a detection circuit 47b instead of the Hall sensor 47a shown in FIG. 9. The detection circuit 47b converts the output voltages of the detection coils 113a and 113b into digital signals and obtains the difference between these output voltages as a detection value Td. When the torsion bar 82c twists, the sensor shaft portion 111 and the cylindrical member 112 undergo relative angular displacement. One of the windows 112a and 112b moves closer to the ridge 111a, while the other moves away. As a result, the inductance of one of the detection coils 113a and 113b increases and the other decreases. By detecting this inductance difference, the amount of torsion between the input shaft 82a and the output shaft 82b can be detected, as with the torque sensor 1 of the embodiment. Although linearity error can be a problem with this type of torque sensor, the linearity can be improved by applying the present invention.

[0072] REFERENCE SIGNS LIST 1...torque sensor, 10...intermediate plate, 21...first sleeve, 25...magnet, 27...adhesive, 31...second sleeve, 311...protrusion, 32...carrier, 35...stator, 351...first stator member, 351a...first flange portion, 351b...first teeth portion, 352...second stator member, 352a...second flange portion, 352b...second teeth portion, 40...sensor housing, 43...printed circuit board, 46...magnetic collecting member, 461...first magnetic collecting member, 462...second magnetic collecting member, 47a...magnetic detection element, 47a...Hall sensor, 47b...detection circuit, 47c...storage unit, 47d...subtractor, 48...first cover, 49...second cover, 491...main body portion, 492...claw portion, 66...polyamide, 7 DESCRIPTION OF SYMBOLS 1...bearing, 72...bearing, 80...electric power steering device, 81...steering wheel, 82...steering shaft, 82a...input shaft, 82b...output shaft, 82b1...groove, 82c...torsion bar, 820...column housing, 83...steering force assist mechanism, 84...universal joint, 85...intermediate shaft, 86...universal joint, 87...pinion shaft, 88...steering gear, 88a...pinion, 88b...rack, 89...tie rod, 90...controller, 92...reduction device, 920...gearbox, 921...worm wheel, 922...worm, 93...electric motor, 95...vehicle speed sensor, 98...ignition switch, 99...power supply device

Claims

1. A torsion bar having one end fixed to the input shaft and the other end fixed to the output shaft, a permanent magnet fixed to one of the input shaft and the output shaft, a stator having a first stator member and a second stator member fixed to the other of the input shaft and the output shaft facing the permanent magnet, a magnetic detection element that outputs a detection value corresponding to the magnetic flux density of magnetic flux flowing from the first stator member to the second stator member, a memory unit that stores a correction map that indicates the relationship between the detection value and a correction value, and a correction circuit that corrects the detection value with the correction value and outputs an output signal, wherein the correction map stores the correspondence between the detection value and the correction value at three or more correction points, a torque sensor characterized in that, when a change range between a first detection value, which is the detection value when the amount of relative rotation in a first direction between the input shaft and the output shaft reaches a maximum allowable amount, and a second detection value, which is the detection value when the amount of relative rotation in a second direction reaches a maximum allowable amount, is divided equally into four regions, and the four regions are defined as a first region, a second region, a third region, and a fourth region in order of decreasing magnitude of the detection value, more correction points are set in the first region and the fourth region than in the second region and the third region.

2. The torque sensor according to claim 1, wherein the first region and the fourth region have substantially the same number of correction points.

3. The torque sensor according to claim 1 or 2, wherein the correction point is set to the detection value that is output when the torsion angle of the torsion bar is approximately zero.

4. A torque sensor according to any one of claims 1 to 3, characterized in that the correction points are set near the first detection value and near the second detection value.

5. A torque sensor according to any one of claims 1 to 4, characterized in that the stator has a flange portion and teeth portion formed by bending a single sheet metal member.

6. A torque sensor according to any one of claims 1 to 5, characterized in that the stator is fixed to the other of the input shaft and the output shaft via a sleeve, the inner peripheral surface of the sleeve is formed with a convex portion formed by embossing, and the outer peripheral surface of the other shaft is formed with a concave groove along the axial direction, and the convex portion and the concave groove are engaged with each other.

7. A device comprising: an input shaft and an output shaft; a torsion bar having one end fixed to the input shaft and the other end fixed to the output shaft; a sensor shaft portion made of a magnetic material fixed to one of the input shaft and the output shaft and having a plurality of axially extending protrusions formed along the circumferential direction; a cylindrical member made of a conductive and non-magnetic material fixed to the other of the input shaft and the output shaft so as to be coaxial with the sensor shaft portion and having a window formed in a position facing the protrusions; a detection coil disposed to surround the window and outputting a detection value corresponding to inductance; a memory portion storing a correction map indicating the relationship between the detection value and a correction value; and a correction circuit which corrects the detection value with the correction value and outputs an output signal, wherein the correction map stores the correspondence between the detection value and the correction value at three or more correction points, a torque sensor characterized in that, when a change range between a first detection value, which is the detection value when the amount of relative rotation in a first direction between the input shaft and the output shaft reaches a maximum allowable amount, and a second detection value, which is the detection value when the amount of relative rotation in a second direction reaches a maximum allowable amount, is divided equally into four regions, and the four regions are defined as a first region, a second region, a third region, and a fourth region in order of decreasing magnitude of the detection value, more correction points are set in the first region and the fourth region than in the second region and the third region.

8. A vehicle steering device comprising: a torque sensor according to any one of claims 1 to 7; an actuator that generates a steering reaction force to be applied to a steering system of a vehicle; and an actuator control unit that drives and controls the actuator in accordance with the detection result of the torque sensor.

Images