JP5315212B2 - Rotation angle sensor abnormality detection device, motor control system, electric power steering - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve availability of a system by not only detecting abnormality of a resolver or a magnetoresistive element but also judging whether or not the influence of the abnormality on the system is serious. <P>SOLUTION: The rotation angle sensor abnormality detector includes a magnetic field sensor 201 and a microcomputer 10. The magnetic field sensor 201 detects a change in magnetic field caused by rotation of a rotary shaft, and outputs signals according to the sine and cosine of a rotation angle of the rotary shaft, respectively. The microcomputer 10 performs determination of a Lissajous trajectory amplitude 121 and determination of a Lissajous trajectory quadrant 122 in an error detection function 12, and thereby detects abnormality of the magnetic field sensor 201 based on the Lissajous waveform of each signal output from the magnetic field sensor 201. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a device for detecting an abnormality of a rotation angle sensor, and a motor control system and an electric power steering system having the device.

  In the servo control system, a rotation angle sensor is necessary to detect the rotation angle and perform feedback control. Further, in brushless motor control, since it is necessary to energize the motor coil in accordance with the rotation angle of the motor, a rotation angle sensor is required in addition to the servo control system. Conventionally, resolvers have been widely used as rotation angle sensors because of their robustness and environmental resistance resulting from their simple structure.

In addition, a servo control system applied to electric power steering, x-by-wire, especially steer-by-wire, fly-by-wire, etc. requires safety and reliability, and therefore requires a failure detection function. For example, Patent Document 1 describes that a resolver failure is detected by utilizing the property of a trigonometric function that is an output of a resolver of sin ² θ + cos ² θ = 1. Patent Document 2 describes that the calculation is simplified by using a determination area as a regular polygon. Furthermore, Patent Document 3 describes that the rotation angle is detected using a magnetoresistive element such as a giant magnetoresistive element (GMR element) instead of the resolver.

JP-A-9-280890 JP 2005-308634 A JP-A-2005-49097

According to the prior art disclosed in Patent Documents 1 and 2 above, if the signal from the resolver or to the resolver is disconnected, the signal from the resolver becomes abnormal and the relationship between sinθ and cosθ does not hold, so sin ² θ + cos ^Since the value of 2θ deviates from 1, it can be detected as a failure and notified to the microcomputer.

  In the conventional technology described above, further consideration is desirable not only in detecting an abnormality, but also in determining whether or not the influence of the abnormality on actual system operation is important. The determination of whether or not a failure has a significant effect on the operation of the system is important in increasing the availability of the system without securing the safety and immediately shutting down the system. In particular, in an electric power steering applied to an automobile having a large vehicle weight, it is desirable from the viewpoint of ensuring safety that the assist by the motor is continued unless a dangerous event occurs due to a control abnormality. However, according to the prior art, since such a determination is not made, even if it is preferable to continue the control, it may be determined as a failure and the control may be stopped. In particular, in the related art disclosed in Patent Document 3, for example, even if a slight burnout that does not affect the control occurs in the magnetoresistive element, the control may be stopped.

  Therefore, an object of the present invention is to provide a technique for increasing the availability of the system by determining whether the influence on the system is serious, as well as detecting the abnormality of the resolver and the magnetoresistive element.

A rotation angle sensor abnormality detection device according to the present invention detects a magnetic field change caused by rotation of a rotation shaft, and according to a sine signal corresponding to the sine of the rotation angle of the rotation shaft and a cosine of the rotation angle of the rotation shaft. Rotation angle sensor comprising: a magnetic field sensor that outputs a cosine signal; and an abnormality detection means that detects an abnormality of the magnetic field sensor based on the sine signal output from the magnetic field sensor and a Lissajous waveform of the cosine signal. An abnormality detection device, wherein the abnormality detection means detects an abnormality of the magnetic field sensor based on an amplitude of the Lissajous waveform and a quadrant through which the Lissajous waveform passes. Is in a predetermined first detection area, and there is a quadrant in which the Lissajous waveform does not pass within a predetermined time, an abnormality of the magnetic field sensor is detected. And butterflies.
A motor control system according to the present invention includes the rotation angle sensor abnormality detection device, a motor connected to the rotation shaft, a magnet installed on the rotation shaft, a motor drive circuit that drives the motor, and the rotation. A motor drive control circuit that controls driving of the motor by the motor drive circuit based on a detection result of the angle, and the magnetic field sensor of the rotation angle sensor abnormality detection device is configured to detect the magnet according to the rotation of the rotation shaft. A change in the magnetic field caused by a change in the rotational position is detected, and when an abnormality of the magnetic field sensor is detected by the abnormality detection means, the driving of the motor is prohibited according to the detection result.
An electric power steering according to the present invention includes the rotation angle sensor abnormality detection device, a motor coupled to the rotation shaft, a magnet installed on the rotation shaft, a motor drive circuit that drives the motor, and the motor. A steering mechanism for a wheel mechanically connected via a column shaft, a steering wheel mechanically connected to the column shaft, a torque sensor for detecting an operation input via the steering wheel, A motor drive control circuit for controlling the drive of the motor by the motor drive circuit based on the detection result of the operation by the torque sensor and the detection result of the rotation angle, and the rotation angle sensor abnormality detection device has The magnetic field sensor is a magnetic field change caused by a change in the rotational position of the magnet according to the rotation of the rotating shaft. Detecting an abnormality of the magnetic sensor is detected by the abnormality detecting means, and inhibits the driving of the motor according to the detection result.

  As described above, according to the present invention, it is possible to increase system availability while ensuring safety.

It is a figure which shows the basic structural example of the resolver abnormality detection apparatus by one Embodiment of this invention. It is a figure for demonstrating the processing content of Lissajous locus | trajectory amplitude determination by one Embodiment of this invention. It is a figure for demonstrating the detection principle of Lissajous locus quadrant determination. It is a figure for demonstrating the processing content of the Lissajous locus quadrant determination by one Embodiment of this invention. It is a figure for demonstrating the processing content of Lissajous locus | trajectory amplitude determination by the 2nd Embodiment of this invention. It is a figure which shows the basic structural example of the resolver abnormality detection apparatus by the 3rd Embodiment of this invention. It is a figure for demonstrating the processing content of Lissajous locus | trajectory amplitude determination by the 3rd Embodiment of this invention. It is a figure which shows the example which applied the resolver abnormality detection apparatus by this invention to the motor control system. It is a figure which shows the example which applied the resolver abnormality detection apparatus by this invention to electric power steering. It is a figure for demonstrating the processing content of the Lissajous locus quadrant determination by the 4th Embodiment of this invention. It is the list which put together the determination result by each case for every quadrant which the Lissajous trajectory passed. It is the list which put together the determination result of the resolver abnormality at the time of applying the method of the Lissajous locus quadrant determination by the 4th Embodiment of this invention to the apparatus shown in FIG. It is the list which summarized the determination result of the resolver abnormality at the time of applying the method of the Lissajous locus quadrant determination by the 4th Embodiment of this invention to the apparatus shown in FIG. It is a figure which shows the basic structural example of the rotation angle sensor abnormality detection apparatus by the 5th Embodiment of this invention. It is sectional drawing which shows the structure of the rotation angle detection apparatus to which the rotation angle sensor abnormality detection apparatus by the 5th Embodiment of this invention is applied. It is a figure which shows the basic composition of the GMR element used for a magnetic field sensor. It is the figure which showed typically the cross section of the free magnetic layer in a GMR element, a spacer layer, and a pinned magnetic layer. It is a figure which shows the structure of the bridge circuit which a magnetic field sensor has. It is a block diagram which shows the structure of a rotation angle detection apparatus. It is the figure which showed typically a mode when a failure generate | occur | produced in the GMR element. It is a figure which shows the Lissajous locus | trajectory by the output signal of the magnetic field sensor in the 5th Embodiment of this invention. It is a figure which shows an example of a signal upper limit / lower limit determination circuit. It is a figure which shows the example which applied the rotation angle sensor abnormality detection apparatus by this invention to the motor control system. It is a figure which shows the example which applied the rotation angle sensor abnormality detection apparatus by this invention to electric power steering.

-First embodiment-
One embodiment of the present invention will be described below together with the principle of the present invention.
(1) Monitor the amplitude of the signal from the resolver, and monitor whether the value of sin ² θ + cos ² θ deviates from 1.
(2) If the value of sin ² θ + cos ² θ deviates from 1 in (1), plot sin θ and cos θ on orthogonal coordinates and coordinates (sin θ, cos θ) within a predetermined time. When plots of coordinates (sin θ, cos θ) exist in all four quadrants, it is considered normal, and when it is not, motor driving is prohibited.

  The influence on the system caused by the abnormality of the resolver is torque fluctuation due to the magnetic pole position detection error of the motor. The relationship between the magnetic pole position detection error and the torque fluctuation is expressed by the following equation.

τm = K ・ iq ・ cosθe
However, τm: Motor output torque K: Torque constant iq: q-axis current θe: Magnetic pole position detection error From this equation, torque fluctuation occurs due to the magnetic pole position detection error within the range of −90 ° <θe <+ 90 °. However, it can be seen that controllability and operability are only deteriorated. Further, when θe <−90 ° or θe> + 90 °, the value of cosθe becomes negative and the signs of iq and τm are reversed, so that it turns out that the motor rotates in the opposite direction. . When such a state occurs in the feedback control system, the polarity of the feedback is reversed, so that the negative feedback becomes positive feedback and the control system diverges. When such a state occurs in, for example, an electric power steering apparatus, a phenomenon such as self-steering that turns the steering without permission occurs. Such a phenomenon must be avoided.

  When the resolver is normal, the plot of the coordinates (sinθ, cosθ) on the sinθ, cosθ orthogonal coordinates, that is, the Lissajous locus draws a unit circle with the origin O as the locus A in FIG. In this case, it is known that the unit circle deviates from the unit circle, such as the locus B and the locus C. Further, as a result of the failure analysis of the resolver, the origin O is included in the Lissajous locus when −90 ° <θe <+ 90 °, and the origin O is included in the Lissajous locus when θe <−90 ° or θe> + 90 °. I understood that it does not contain.

  Here, including the origin O within the Lissajous trajectory means that the Lissajous trajectory passes through all four quadrants, and not including the origin O within the Lissajous trajectory indicates that the Lissajous trajectory is one of the four quadrants. It will not pass through the quadrant.

  Therefore, when the coordinates (sinθ, cosθ) are plotted on the sinθ, cosθ orthogonal coordinates and the plots of the coordinates (sinθ, cosθ) exist in all four quadrants within a predetermined time, it is normal. Can be considered abnormal. However, in the case of electric power steering, θ may be almost constant when going straight, and the plot of coordinates (sinθ, cosθ) may not pass through all four quadrants within a predetermined time. It should be determined in combination with another diagnosis, for example, the diagnosis of (1) above.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a diagram showing a basic configuration example of an apparatus for detecting a resolver abnormality according to an embodiment of the present invention. When the excitation signal f (t) from the excitation signal generator 2 is applied to the primary winding, the resolver 1 receives the signal f (t) according to the rotation angle θ of the rotary shaft 7 attached to the resolver 1 and the motor 5. sinθ, f (t) cosθ is induced in the secondary winding. The excitation signal generator 2, the motor 5, and the rotary shaft 7 are not shown in FIG. 1, but are shown in FIGS. The signals f (t) sinθ and f (t) cosθ induced in the secondary winding of the resolver 1 are converted into digital signals by the A / D converter 11 and input to the error detection function 12. The error detection function 12 executes a Lissajous locus amplitude determination 121 and a Lissajous locus quadrant determination 122 together with other diagnoses 129. When the Lissajous locus amplitude determination 121 determines that the detection range is 1 and the Lissajous locus quadrant determination 122 determines that the abnormality is detected, it is determined that motor driving is prohibited. Further, when it is determined that there is an abnormality in the other diagnosis 129, it is determined that the motor drive is prohibited.

  In addition, it is desirable that the error detection function 12 is realized by software in the microcomputer 10 and that the A / D converter 11 includes the microcomputer 10 because the amount of hardware can be reduced.

  According to the apparatus described above, it is possible to distinguish between a resolver failure that does not affect the safety of system operation and a resolver failure that affects the safety of system operation. In the case of a resolver failure that does not affect performance, the operation can be continued, and the availability of the system can be increased while ensuring safety. In the case of a resolver failure that does not affect the safety of system operation, it is possible to turn on a warning lamp or limit the motor drive output.

  FIG. 2 is a diagram for explaining the processing contents in the Lissajous locus amplitude determination 121.

  When the coordinates (sin θ, cos θ) are plotted on the sin θ, cos θ orthogonal coordinates from the signals f (t) sin θ, f (t) cos θ, and the plot is in the detection area 1, it is regarded as abnormal.

  In the present embodiment, the detection area 1 becomes a unit circle, for example, when normalized by the amplitude of a unit circle (f (t). In this embodiment, the detection region 1 is a circle having a radius r0 including the amplitude of f (t). )) With a certain width inside and outside. This constant width may be determined in consideration of individual differences of resolvers and electronic circuits, temperature characteristics, and deterioration over time. In FIG. 2, the center value of the radius in the normal state is r0, the upper limit is r1, and the lower limit is r2.

  FIG. 3 is a diagram for explaining the detection principle of the Lissajous locus quadrant determination 122. When the resolver is normal, the plot of the coordinates (sinθ, cosθ) on the sinθ, cosθ orthogonal coordinates, that is, the Lissajous locus draws a unit circle like the locus A in FIG. 3, and when the resolver fails, the locus B, It is known that it deviates from the unit circle like the locus C. Further, as a result of the failure analysis of the resolver, the origin O is included in the Lissajous locus when −90 ° <θe <+ 90 °, and the origin O is included in the Lissajous locus when θe <−90 ° or θe> + 90 °. I know it does n’t.

  FIG. 4 is a diagram for explaining the processing contents in the Lissajous locus quadrant determination 122. When passing through all four quadrants within a predetermined time such as the trajectories A and B, it is considered normal, and when passing only two quadrants as with the trajectory C, it is considered abnormal.

  In addition, whether it exists in each quadrant can be judged from the positive / negative of sin (theta) and cos (theta) as follows.

Quadrant 1: sinθ> 0, cosθ> 0
Second quadrant: sinθ <0, cosθ> 0
Third quadrant: sinθ <0, cosθ <0
Quadrant 4: sinθ> 0, cosθ <0

  The determination of whether all four quadrants have been passed within a predetermined time is stored, for example, by storing the time at which each quadrant has passed, and whether or not the difference from the current time is within the predetermined time. Judgment can be made.

  In the first embodiment described above, the resolver 1 is normal when the Lissajous trajectory determination 122 determines that the Lissajous trajectory has passed all four quadrants within a predetermined time as shown by the trajectories A and B shown in FIG. In the case where only two quadrants are passed as in the locus C, it is regarded as abnormal. However, since the rotating shaft 7 rotates only in a limited electric angle when the vehicle is traveling straight, there is a possibility of causing an erroneous determination that it is normal but abnormal. In order to avoid such an erroneous determination, in this embodiment, the Lissajous locus amplitude determination 121 and the Lissajous locus quadrant determination 122 are combined.

  Also, after determining that the resolver 1 is abnormal and stopping the steering assist by the motor 5, if the Lissajous trajectory passes through all four quadrants, this is regarded as normal and the steering assist is resumed. Also good. In this way, even if the steering assist is stopped when traveling straight due to an erroneous determination, the rotating shaft 7 is rotated by 360 ° by an electrical angle after that, and the Lissajous locus has all four quadrants. If the vehicle passes, steering assist is resumed.

-Second Embodiment-
In the present embodiment, an example will be described in which the detection area of the Lissajous locus amplitude determination 121 is a square instead of the detection area 1 shown in FIG. FIG. 5 is a diagram for explaining the processing contents in the Lissajous locus amplitude determination 121 in the present embodiment. According to the present embodiment, an abnormality can be detected only by determining whether or not the absolute values of sinθ and cosθ are within a predetermined range. Therefore, detection is performed as described with reference to FIG. 3 in the first embodiment. The amount of calculation can be reduced as compared with the case where the area is a circle. In this case, it can be determined by the following logical expression.

| Sinθ | <b and | cosθ | <b,
And | sinθ |> a And | cosθ |> a
Then normal,
Otherwise, it is abnormal.

  Similarly, the detection area can be a regular polygon.

-Third embodiment-
In the present embodiment, an example will be described in which the detection range of the Lissajous locus amplitude determination 121 is set in two stages, ie, a detection area 1 and a detection area 2. FIG. 6 is a diagram showing a basic configuration example of an apparatus for detecting a resolver abnormality according to the present embodiment. FIG. 7 is a diagram for explaining the processing contents in the Lissajous locus amplitude determination 121 in the present embodiment. In FIG. 6, if the detection range 2 is determined in the Lissajous locus amplitude determination 121, the motor drive is prohibited regardless of the result of the Lissajous locus quadrant determination 122. However, if it is determined to be abnormal, motor drive is prohibited.

  As shown in FIG. 7, the detection area 2 is set to an area having a smaller amplitude than that of the detection area 1 (in a circle having a radius r3). In this embodiment, motor drive inhibition is prohibited when the amplitude of the signal from the resolver decreases and S / N decreases to θe <−90 ° or θe> + 90 °.

  The method of setting the detection area in two steps as in the present embodiment is not limited to the case where the detection area of the Lissajous locus amplitude determination 121 is circular as shown in FIG. As described in the second embodiment, the present invention can be similarly applied to a case where the detection area of the Lissajous locus amplitude determination 121 is a square or a regular polygon.

-Fourth embodiment-
In the present embodiment, a method for further reducing the probability of erroneous determination in the Lissajous locus quadrant determination 122 will be described. FIG. 10 is a diagram for explaining the processing contents in the Lissajous locus quadrant determination 122 according to the present embodiment. Here, when the Lissajous trajectory does not pass through one quadrant of the four quadrants, the trajectory of the quadrant is interpolated by connecting the trajectories of two quadrants adjacent to the quadrant with a straight line. Then, whether the resolver is normal or abnormal is determined based on whether the locus obtained by adding the interpolated portion surrounds the origin.

Specifically, when a straight line connecting the trajectories of two quadrants adjacent to the quadrant is represented as y = ax + b, the trajectory obtained by adding an interpolation amount from the positive / negative of b representing the y intercept of the straight line surrounds the origin. Can be determined. That is, if the trajectories of two adjacent quadrants are P1 (x1, y1) and P2 (x2, y2), respectively, the slope a and the y-intercept b of the straight line connecting these are expressed by the following equations (1) and (2). Desired.
a = (y2-y1) / (x2-x1) (1)
b = (x2y1-x1y2) / (x2-x1) (2)

  FIG. 11 summarizes the determination results for each quadrant through which the Lissajous trajectory passes in a list. Case 1 shown at the top of the list shows the determination result when the Lissajous trajectory passes through the four quadrants. In this case, it can be determined that the resolver is normal without interpolating the Lissajous trajectory.

  Case 2 shown in the second row and Case 3 shown in the third row show the determination results when passing through three quadrants other than the first quadrant. In this case, if the y-intercept b of the straight line connecting the trajectory of the second quadrant and the fourth quadrant adjacent to the first quadrant is both positive (Case 2), the resolver is regarded as normal, otherwise (Case 3) considers the resolver to be abnormal. When there are a plurality of trajectories in the second quadrant and the fourth quadrant, the determination is performed based on a straight y-intercept b using a trajectory closer to the first quadrant. That is, in the second quadrant, the locus having the maximum y coordinate is used. Here, the reason why the locus having the maximum x coordinate (the minimum absolute value) is not used is to prevent an error due to a small absolute value of the coordinate, that is, a small signal intensity, from increasing. Similarly, in the fourth quadrant, the locus having the maximum x coordinate is used.

  Case 4 shown in the fourth row and Case 5 shown in the fifth row show the determination results when passing through three quadrants other than the second quadrant. In this case, if the y-intercept b of the straight line connecting the trajectory of the first quadrant and the third quadrant adjacent to the second quadrant is positive (Case 4), the resolver is regarded as normal, otherwise (Case 5) considers the resolver to be abnormal. When there are a plurality of trajectories in the first quadrant and the third quadrant, the determination is performed based on a straight y-intercept b using a trajectory closer to the second quadrant. That is, the locus with the maximum y coordinate is used in the first quadrant, and the locus with the minimum x coordinate (maximum absolute value) is used in the third quadrant.

  Case 6 shown in the sixth row and Case 7 shown in the seventh row show the determination results when passing through three quadrants other than the third quadrant. In this case, if the y-intercept b of the straight line connecting the fourth quadrant and the second quadrant adjacent to the third quadrant is negative (Case 6), the resolver is regarded as normal, otherwise (Case 7) considers the resolver to be abnormal. When there are a plurality of trajectories in the fourth quadrant and the second quadrant, the determination is performed based on a straight y-intercept b using a trajectory closer to the third quadrant. That is, in the fourth quadrant, the locus with the minimum y coordinate (maximum absolute value) is used, and in the second quadrant, the locus with the minimum x coordinate (maximum absolute value) is used.

  Case 8 shown in the eighth row and Case 9 shown in the ninth row show the determination results when passing through three quadrants other than the fourth quadrant. In this case, if the y-intercept b of the straight line connecting the trajectory of the third quadrant and the first quadrant adjacent to the fourth quadrant is negative (Case 8), the resolver is regarded as normal, otherwise (Case 9) considers the resolver to be abnormal. When there are a plurality of trajectories in the third quadrant and the first quadrant, the determination is performed based on a straight y-intercept b using a trajectory closer to the fourth quadrant. That is, in the third quadrant, the trajectory having the minimum y coordinate (maximum absolute value) is used, and in the first quadrant, the trajectory having the maximum x coordinate is used.

  Case 10 shown in the tenth stage shows a determination result when the Lissajous trajectory does not pass through two or more quadrants out of the four quadrants. In this case, it can be determined that the resolver is abnormal without interpolating the Lissajous locus.

  12 shows a resolver abnormality caused by a combination of the Lissajous locus amplitude determination 121 and the Lissajous locus quadrant determination 122 when the determination method of the Lissajous locus quadrant determination 122 according to the present embodiment as described above is applied to the apparatus shown in FIG. The judgment results are summarized in a list. Note that the Lissajous locus amplitude determination 121 determines whether the resolver 1 is normal or abnormal by using the determination method described in the first or second embodiment. Case 0 shown at the top of the list shows a determination result when the Lissajous locus amplitude determination 121 determines that the Lissajous locus is not within the detection area 1 shown in FIG. 2 or FIG. In this case, regardless of the determination result of the Lissajous locus quadrant determination 122, it can be determined that the resolver is normal.

  In Case 1 shown in the second row from Case 1 shown in the second row, the Lissajous locus passes when the Lissajous locus amplitude determination 121 determines that the Lissajous locus is within the detection area 1 shown in FIG. 2 or FIG. The determination results for each quadrant are shown. Here, it is determined whether the resolver is normal or abnormal according to the determination result of the Lissajous locus quadrant determination 122 described with reference to FIG. That is, whether the resolver is normal or abnormal is determined based on the positive / negative of the y-intercept b of the straight line connecting two quadrants adjacent to the quadrant where the Lissajous trajectory has not passed.

  Further, FIG. 13 shows the determination result of the resolver abnormality by the combination of the Lissajous locus amplitude determination 121 and the Lissajous locus quadrant determination 122 when the determination method of the Lissajous locus quadrant determination 122 according to this embodiment is applied to the apparatus shown in FIG. These are summarized in a list. Note that the Lissajous trajectory amplitude determination 121 determines whether the resolver 1 is normal or abnormal by using the determination method as described in the third embodiment. Case 0 shown at the top of the list shows a determination result when the Lissajous locus amplitude determination 121 determines that the Lissajous locus is not within the detection area 1 shown in FIG. In this case, it is possible to determine that the resolver is normal regardless of the determination result of the Lissajous locus quadrant determination 122, as described with reference to FIG.

  Case 11 shown at the bottom indicates the determination result when the Lissajous locus amplitude determination 121 determines that the Lissajous locus is within the detection area 2 shown in FIG. In this case, regardless of the determination result of the Lissajous locus quadrant determination 122, it can be determined that the resolver is abnormal.

  In Case 1 shown in the second stage from Case 1 shown in the second stage, when the Lissajous trajectory amplitude determination 121 determines that the Lissajous trajectory is within the detection area 1 shown in FIG. The determination results for each quadrant through which the Lissajous trajectory has passed are shown. Here, similarly to FIG. 12, it is determined whether the resolver is normal or abnormal according to the determination result of the Lissajous locus quadrant determination 122 described with reference to FIG. That is, whether the resolver is normal or abnormal is determined based on the positive / negative of the y-intercept b of the straight line connecting two quadrants adjacent to the quadrant where the Lissajous trajectory has not passed.

  FIG. 8 is a diagram showing an example in which any of the resolver abnormality detection devices according to the first to fourth embodiments described above is applied to a motor control system.

  The microcomputer 10 controls the motor driver 4, and the motor driver 4 drives the motor 5. A rotating shaft 7 of the motor 5 is connected to the resolver 1 and the controlled object 6. The resolver 1 detects the rotation angle (magnetic pole position) of the motor 5 via the rotating shaft 7, and the microcomputer 10 drives and controls the motor 5 by the motor driver 4 based on the result. Further, the motor 5 controls the control object 6 via the rotating shaft 7.

  The error detection function 12 executes any of the resolver abnormality detection methods described in the first to fourth embodiments. When the motor detection prohibition is stopped by the error detection function 12, the control system stops the driving of the motor 5. In this case, there are several methods for stopping the driving of the motor 5, but a method of cutting off the control signal to the motor driver 4, a method of stopping the power supply to the motor driver 4, and an output from the motor driver 4 to the motor 5 There is a way to shut off.

  FIG. 9 is a diagram showing an example in which any of the resolver abnormality detection devices according to the first to fourth embodiments described above is applied to electric power steering.

  The motor 5 is mechanically connected to a wheel steering mechanism acting as the control target 6 via a reduction gear, a column shaft, a pinion, and a rack. In this steering mechanism, the column shaft is mechanically connected to the steering wheel 8 and the torque sensor 9. An operation input from the driver via the steering wheel 8 is detected by a torque sensor 9, a detection signal of the torque sensor 9 is input to the microcomputer 10, and the microcomputer 10 detects the driver's operation torque detected by the torque sensor 9 ( The driver's steering is assisted by controlling the output torque of the motor 5 in accordance with the steering force.

-Fifth embodiment-
In each of the embodiments described above, the resolver diagnosis method has been described. However, the present invention can also be applied to a rotation angle sensor diagnosis method. This method will be described below as a fifth embodiment of the present invention. FIG. 14 is a diagram illustrating a basic configuration example of an apparatus for detecting an abnormality of the rotation angle sensor according to the present embodiment. As a rotation angle sensor, this apparatus detects a magnetic field change that occurs according to the rotation of the rotary shaft 7 as a change in the rotation angle instead of the resolver 1 in the apparatus shown in FIG. 1 as the first embodiment. The magnetic field sensor 201 which acts is provided. The abnormality of the magnetic field sensor 201 is detected and the driving of the motor is prohibited.

  The magnetic field sensor 201 uses a magnetoresistance effect (magnetoresistance, MR) to change the magnetic field when the magnetic field generated by the magnet installed on the rotating shaft 7 (see FIGS. 23 and 24) changes according to the rotation of the rotating shaft 7. Is detected as a change in resistance value. The magnetic field sensor 201 includes a cosine sensor 201A and a sine sensor 201B. The cosine sensor 201A outputs voltage signals Vc1 and Vc2 corresponding to the cosine component cosθ of the rotation angle θ of the rotating shaft 7 to the microcomputer 10. On the other hand, the sine sensor 201 </ b> B outputs voltage signals Vs <b> 1 and Vs <b> 2 corresponding to the sine component sin θ of the rotation angle θ of the rotating shaft 7 to the microcomputer 10.

  The microcomputer 10 detects the difference between the voltage signals Vc1 and Vc2 from the cosine sensor 201A and the difference between the voltage signals Vs1 and Vs2 from the sine sensor 201B. These differences are converted into digital signals by the A / D converter 11 and then used to calculate the values of the sine component sinθ and the cosine component cosθ of the rotation angle θ. A method of calculating sin θ and cos θ will be described later. The error detection function 12 executes the Lissajous trajectory amplitude determination 121 and the Lissajous trajectory quadrant determination 122 based on the calculated values of sin θ and cos θ by any of the methods described in the above-described embodiments. A diagnosis 129 is also performed. As a result, when the Lissajous locus amplitude determination 121 determines that the detection area is 1 and the Lissajous locus quadrant determination 122 determines that there is an abnormality, it is determined that motor driving is prohibited. Further, when it is determined that there is an abnormality in the other diagnosis 129, it is determined that the motor drive is prohibited.

  According to the apparatus of the present embodiment, as in the case of the resolver failure in the first to fourth embodiments, the rotation angle sensor that does not affect the safety of the system operation and the system operation A failure of the rotation angle sensor that affects safety can be distinguished. That is, in the case of a failure of the rotation angle sensor that does not affect the safety of the system operation, the operation can be continued, and the availability of the system can be increased while ensuring the safety. In the case of a failure of the rotation angle sensor that does not affect the safety of the operation of the system, it is possible to turn on a warning lamp or limit the motor drive output.

  FIG. 15 is a cross-sectional view showing a configuration of a rotation angle detection device to which the rotation angle sensor abnormality detection device is applied. This rotation angle detection device includes a motor unit 100 and a rotation angle detection unit 200.

  The motor unit 100 generates rotational torque by rotating a plurality of rotating magnetic poles according to a magnetic action between the plurality of fixed magnetic poles and the plurality of rotating magnetic poles, and constitutes the plurality of fixed magnetic poles. A stator 110, a rotor 130 constituting a plurality of rotating magnetic poles, and a casing are formed. The stator 110 includes a stator core 111 and a stator coil 112 attached to the stator core 111. The rotor 130 is opposed to the inner peripheral side of the stator 110 via a gap and is rotatably supported. In the present embodiment, a three-phase AC type surface magnet type synchronous motor is used as the motor 100.

  The housing of the motor unit 100 includes a cylindrical frame 101 and a first bracket 102 and a second bracket 103 provided at both ends of the frame 101 in the axial direction. A bearing 106 is provided in the hollow portion of the first bracket 102, and a bearing 107 is provided in the hollow portion of the second bracket 103. These bearings support the rotary shaft 7 so as to be rotatable.

  A seal member (not shown) is provided between the frame 101 and the first bracket 102. The seal member is an O-ring provided in an annular shape, and is sandwiched between the frame 101 and the first bracket 102 while being compressed in the axial direction and the radial direction, respectively. Thereby, between the flame | frame 101 and the 1st bracket 102 can be sealed, and the front side can be waterproofed. Further, the frame 101 and the second bracket 103 are also waterproofed by a similar seal member (not shown).

  The stator 110 includes a stator core 111 and a stator coil 112 attached to the stator core 111, and is installed on the inner peripheral surface of the frame 101. The stator core 111 is a magnetic body (magnetic path forming body) formed by laminating a plurality of silicon steel plates in the axial direction. The stator core 111 protrudes inward in the radial direction from the annular back core and the inner peripheral portion of the back core. It consists of a plurality of teeth arranged at equal intervals in the direction.

  A winding conductor constituting the stator coil 112 is intensively wound around each of the plurality of teeth. The plurality of winding conductors are electrically connected for each phase by a connecting member juxtaposed in the axial end of one coil end portion (on the second bracket 103 side) of the stator coil 112, and further, a three-phase winding. As electrically connected. In addition, although there are a Δ (delta) connection method and a Y (star) connection method as the connection method of the three-phase windings, the Δ (delta) connection method is adopted in this embodiment.

  The rotor 130 includes a rotor core fixed on the outer peripheral surface of the rotating shaft 7, a plurality of magnets fixed to the outer peripheral surface of the rotor core, and a magnet cover provided on the outer peripheral side of the magnet. The magnet cover is for preventing the magnet from scattering from the rotor core, and is a member having a cylindrical or tubular shape formed of a non-magnetic material such as stainless steel.

  Next, the configuration of the rotation angle detection unit 200 will be described. The rotation angle detection unit 200 includes a magnetic field sensor 201, a sensor magnet 202, and a housing 203. The rotation angle detection unit 200 is installed in a space surrounded by the housing 203 and the second bracket 103. The sensor magnet 202 is installed on a shaft that rotates in conjunction with the rotation shaft 7, and when the rotation shaft 7 changes its rotation position, the direction of the magnetic field generated changes accordingly. By detecting this magnetic field direction with the magnetic field sensor 201, the rotation angle (rotation position) of the rotating shaft 121 can be measured.

  The sensor magnet 202 is a two-pole magnet magnetized to two poles or a multi-pole magnet magnetized to four or more poles.

  The magnetic field sensor 201 has an output signal that changes according to the direction of the magnetic field, and is configured using a magnetoresistive element. A magnetoresistive element is an element that produces a phenomenon (magnetoresistance effect) in which a resistance value changes according to a magnetic field. As such magnetoresistive effect, for example, anisotropic magnetoresistive effect (anisotropic magnetoresisitance, AMR), tunneling magnetoresistance effect (tunneling magnetoresistance, TMR), giant magnetoresistance effect (giant magnetoresistance, GMR), etc. are known. Yes. In this embodiment, an example in which a giant magnetoresistive element (GMR element) that produces a giant magnetoresistive effect is used for the magnetic field sensor 201 will be described.

  The magnetic field sensor 201 detects the direction of the magnetic field at the installation location with reference to a preset reference angle, and outputs a signal corresponding to the detection result. That is, a signal corresponding to the rotation angle θ of the rotating shaft 7 is output. The magnetic field sensor 201 used in the present embodiment has two sets of bridge circuits each composed of a plurality of GMR elements as will be described later. Each bridge circuit outputs a signal proportional to cos θ and sin θ.

  The magnetic field sensor 201 is fixed to the housing 203. The housing 203 is preferably made of a material having a magnetic susceptibility of a predetermined value or less, for example, a magnetic susceptibility of 1.1 or less, such as aluminum or resin, so as not to affect the direction of magnetic flux. In this embodiment, an example made of aluminum will be described.

  The magnetic field sensor 201 needs to be arranged so that its position is fixed with respect to the motor unit 100. This is because the rotation angle of the rotation shaft 7 can be detected by detecting the change in the magnetic field direction by the magnetic field sensor 201 when the rotation angle of the rotation shaft 7 changes and the direction of the sensor magnet 202 changes. It is. In the present embodiment, the magnetic field sensor 201 is disposed on the rotation axis center line 226 of the rotation shaft 7. This is preferable because distortion of the magnetic field formed by the sensor magnet 202 can be suppressed and measurement accuracy of the rotation angle can be improved. Note that the magnetic field sensor 201 may be fixed to components other than the housing 203.

  A sensor wiring 208 is connected to the magnetic field sensor 201. An output signal of the magnetic field sensor 201 is transmitted by the sensor wiring 208.

  Next, the configuration of the magnetic field sensor 201 will be described. A basic configuration of the GMR element used in the magnetic field sensor 201 is shown in FIG. The GMR element is formed by overlapping a free magnetic layer 21, a spacer layer 22, and a pinned magnetic layer 23 as shown in the figure. The free magnetic layer 21 and the pinned magnetic layer 23 are magnetic layers magnetized in predetermined magnetization directions θf and θp, respectively, whereas the spacer layer 22 is a non-magnetized nonmagnetic layer. The spacer layer 22 is sandwiched between the free magnetic layer 21 and the pinned magnetic layer 23. When a magnetic field is applied to the GMR element from the outside, the magnetization direction of the pinned magnetic layer 23 does not change and remains fixed, whereas the magnetization direction θf of the free magnetic layer 21 is the direction of the external magnetic field as indicated by reference numeral 20. It changes according to. The pinned magnetic layer 23 is also called a pinned magnetic layer.

  When a voltage is applied across the GMR element, a current corresponding to a predetermined element resistance flows. The magnitude of the element resistance changes depending on the difference Δθ = θf−θp between the magnetization direction θp of the pinned magnetic layer 23 and the magnetization direction θf of the free magnetic layer 21. Therefore, if the magnetization direction θp of the pinned magnetic layer 23 is known, the resistance value of the GMR element is measured to detect the magnetization direction θf of the free magnetic layer 21, that is, the direction of the external magnetic field by utilizing this property. Can do.

  A mechanism by which the resistance value of the GMR element changes according to Δθ = θf−θp will be described below. The magnetization directions in the thin free magnetic layer 21 and the pinned magnetic layer 23 are related to the spin directions of electrons in the magnetic material included in each magnetic layer. Therefore, when Δθ = 0, the electrons in the free magnetic layer 21 and the electrons in the pinned magnetic layer 23 have a high proportion of electrons having the same spin direction. Conversely, when Δθ = 180 °, the electrons in both magnetic layers have a high proportion of electrons whose spin directions are opposite to each other.

  FIG. 17 schematically shows a cross section of the free magnetic layer 21, the spacer layer 22, and the pinned magnetic layer 23 in the GMR element. Each arrow in the free magnetic layer 21 and the pinned magnetic layer 23 schematically shows the spin direction of electrons occupying a large number in each magnetic layer.

  FIG. 17A shows a case where Δθ = 0, and the spin directions of the free magnetic layer 21 and the pinned magnetic layer 23 are aligned. In this case, electrons in the rightward spin emitted from the pinned magnetic layer 23 occupy a large number of electrons in the same spin direction in the free magnetic layer 21, so that there is little scattering in the free magnetic layer 21. As a result, electrons from the pinned magnetic layer 23 can move in the GMR element through a locus such as an electron locus 810.

  On the other hand, FIG. 17B shows a case where Δθ = 180 °, and the spin directions of the free magnetic layer 21 and the pinned magnetic layer 23 are reversed. In this case, rightward spin electrons emitted from the pinned magnetic layer 23 are strongly scattered when entering the free magnetic layer 21 because there are many reverse spin electrons. As a result, electrons from the pinned magnetic layer 23 move in the GMR element through a locus such as an electron locus 820. Thus, since the electron scattering increases when Δθ = 180 °, the electrical resistance increases as compared with the case where Δθ = 0 °.

  When Δθ is between 0 ° and 180 °, the state is intermediate between FIG. 17 (a) and FIG. 17 (b). The resistance value at this time decreases as Δθ approaches 0 ° and increases as it approaches 180 °.

When the resistance value of the GMR element is R, the magnitude of R is expressed by the following equation (3). In the formula (3), R ′ ₀ , R ₀ and G are all predetermined constants.

（３）

(3)

The rate of change of the resistance value R expressed by Equation (3), that is, G / R ₀ is called a GMR coefficient. A GMR element having a GMR coefficient of about several percent to several tens of percent is generally known as a GMR element.

  As described above, in the GMR element, the current flow (electrical resistance) can be controlled by the direction of electron spin. For this reason, the GMR element is also called a spin valve element.

  In general, a thin magnetic film (thin film magnetic film) has an extremely large demagnetizing factor in the normal direction of the surface, so the magnetization vector cannot rise in the normal direction (thickness direction). , Lying in the plane. Since both the free magnetic layer 21 and the pinned magnetic layer 23 constituting the GMR element are sufficiently thin, the respective magnetization vectors lie in the in-plane direction.

  The magnetic field sensor 201 has two sets of Wheatstone bridge circuits each configured by using four GMR elements as described above. The configuration of the Wheatstone bridge circuit is shown in FIG. In FIG. 18, the left bridge circuit 30 is called a COS bridge, and the right bridge circuit 40 is called a SIN bridge. The COS bridge 30 and the SIN bridge 40 correspond to the cosine sensor 201A and the sine sensor 201B in FIG. 14, respectively.

  The COS bridge 30 includes GMR elements R1 to R4 indicated by reference numerals 31 to 34, respectively. The magnetization directions of the pinned magnetic layers of the elements R1 and R3 are respectively set to θp = 0, and the magnetization directions of the pinned magnetic layers of the elements R2 and R4 are respectively set to θp = 180 °. On the other hand, since the magnetization direction θf of the free magnetic layer is determined by the direction of the external magnetic field as described above, it is the same in the elements R1 to R4 when a magnetic field is applied from the outside. That is, if the rotation angle θ of the rotating shaft 7 is the direction of the external magnetic field, the magnetization direction of the free magnetic layer in the elements R1 to R4 can be expressed as θf = θ.

Therefore, the difference between the magnetization direction θp of the pinned magnetic layer and the magnetization direction θf of the free magnetic layer in the elements R1 and R3 is Δθ1, and the magnetization direction θp of the pinned magnetic layer and the magnetization direction θf of the free magnetic layer in the elements R2 and R4 are If the difference between the two is Δθ2, the following equation (4) holds.
Δθ1 = θf−θp = θ, Δθ2 = θf−θp = θ + π (4)

  Based on the above formulas (3) and (4), the resistance values of the elements R1 to R4 are expressed by the following formula (5).

（ｎ＝１，３）

（ｎ＝２，４） （５）

(N = 1, 3)

(N = 2, 4) (5)

  When the difference between the voltage Vc1 at the terminal V1 and the voltage Vc2 at the terminal V2 when the excitation voltage e0 is applied to the COS bridge 30 in FIG. 18 is expressed as ΔVc = Vc2−Vc1, ΔVc is expressed based on the above equation (5). It can be expressed as the following formula (6).

（６）

(6)

  Assuming that Rn0 is equal for n = 1 to 4, it can be replaced with Rn0 = R0 in the above equation (5). Then, the equation (6) can be rewritten as the following equation (7). It can be seen from equation (7) that a signal proportional to cos θ can be obtained based on the voltage Vc1 and the voltage Vc2 output from the COS bridge 30.

（７）

(7)

  On the other hand, the SIN bridge 40 includes GMR elements R1 to R4 indicated by reference numerals 41 to 44, respectively. The magnetization directions of the pinned magnetic layers of the elements R1 and R3 are respectively set to θp = 90 °, and the magnetization directions of the pinned magnetic layers of the elements R2 and R4 are respectively set to θp = 270 °. When the difference between the voltage Vs1 at the terminal V1 and the voltage Vs2 at the terminal V2 when the excitation voltage e0 is applied to the SIN bridge 40 is expressed as ΔVs = Vs2−Vs1, ΔVs is expressed as the following equation (8). Can do. It can be seen from equation (8) that a signal proportional to sin θ can be obtained based on the voltage Vs1 and the voltage Vs2 output from the SIN bridge 40.

（８）

(8)

  From equations (7) and (8), it can be seen that the ratio of the output signals of the two bridges is tanθ. That is, the rotation angle θ can be obtained from the following equation (9).

（９）

(9)

  Note that the equation (9) has a problem that when the absolute value of ΔVc decreases, the denominator in the ArcTan function decreases, so that the influence of the error of ΔVc increases, and as a result, the calculation accuracy of θ decreases. In order to deal with such a problem, the magnitude relationship between the absolute values of ΔVc and ΔVs is determined. As a result, when | ΔVc | is larger than | ΔVs |, the rotation angle θ is obtained by the above equation (9).

  On the other hand, when | ΔVc | is smaller than | ΔVs |, the following equation (10) is used instead of equation (9) to determine the rotation angle θ.

（１０）

(10)

  As described above, by properly using the equations (9) and (10) according to the magnitude relationship between the absolute values of ΔVc and ΔVs, it is possible to prevent an increase in calculation error due to a decrease in the absolute value of ΔVc.

  In addition, since the ArcTan function is used in Equation (9), the rotation angle θ can be obtained only in the range of ± 90 °, and the rotation angle θ cannot be measured over the entire angle range of 0 to 360 °. There is a problem. In order to deal with such a problem, quadrant determination is performed to determine in which quadrant θ is based on the signs of ΔVs and ΔVc. By combining the result of this quadrant determination and the appropriate use of the above formulas (9) and (10), the rotation angle θ can be correctly obtained over the entire angle range of 0 to 360 °.

  FIG. 19 is a block diagram illustrating a configuration of a rotation angle detection device including the magnetic field sensor 201. The rotation angle detection device includes a magnetic field sensor 201, a microcomputer 10, detection circuits 14A and 14B, a positive output circuit 15, and a negative output circuit 16.

  As described with reference to FIG. 14, the magnetic field sensor 201 includes a cosine sensor 201 </ b> A and a sine sensor 201 </ b> B configured using the COS bridge 30 and the SIN bridge 40, respectively. Each of these sensors is supplied with the 5V DC voltage output from the positive output circuit 15 as the excitation voltage e0 and the ground voltage (earth potential) from the negative output circuit 16. Based on these supply voltages, voltage signals Vc1 and Vc2 corresponding to the cosine component cosθ of the rotation angle θ and voltage signals Vs1 and Vs2 corresponding to the sine component sinθ of the rotation angle θ are the cosine sensor 201A and the sine sensor 201B. To the microcomputer 10 via the detection circuits 14A and 14B. Note that the output voltages of the positive output circuit 15 and the negative output circuit 16 may be equal in a period in which the magnetic field sensor 201 is not excited, that is, in a period in which the rotation angle θ is not detected.

  The detection circuit 14A differentially detects the difference ΔVc expressed by the above equation (7) based on the voltage signals Vc1 and Vc2 from the cosine sensor 201A. The detection circuit 14B differentially detects the difference ΔVs expressed by the above equation (8) based on the voltage signals Vs1 and Vs2 from the sine sensor 201B. The detected differences ΔVc and ΔVs are each amplified by about 10 times and then output to the microcomputer 10. In the microcomputer 10, the input differences ΔVc and ΔVs are converted into digital signals by the A / D converters 11 </ b> A and 11 </ b> B and then input to the calculation unit 13. The A / D converters 11A and 11B correspond to the A / D converter 11 in FIG. Based on the differences ΔVc and ΔVs, the calculation unit 13 obtains the rotation angle θ using the equations (9) and (10) as described above. The rotation angle θ obtained by the calculation unit 13 is output from the microcomputer 10 to the outside.

  The calculation unit 13 also executes processing corresponding to the error detection function 12 of FIG. That is, based on the differences ΔVc and ΔVs, the values of sin θ and cos θ are obtained using the above-described equations (7) and (8). Then, based on the calculated values of sin θ and cos θ, it is determined whether or not the motor drive is prohibited by executing the Lissajous locus amplitude determination and the Lissajous locus quadrant determination as described above. This determination result is output to the outside from the microcomputer 10 as with the rotation angle θ.

  Next, the principle of failure detection using the GMR element of the magnetic field sensor 201 will be described. The GMR element constituting the magnetic field sensor 201 is constituted by a thin film having a thickness of about several nm. For this reason, if an excessive current flows, a part of the GMR element may burn out and cause a failure. When such a failure occurs, a local resistance increase occurs in the GMR element. By detecting this increase in resistance of the GMR element, failure detection of the magnetic field sensor 201 is performed.

  FIG. 20 is a diagram schematically showing a state in which a failure has occurred in one of the four GMR elements constituting the bridge circuit as described above in the magnetic field sensor 201. The wiring 210 is a wiring pattern formed by a GMR element, and a current flows in the direction indicated by the arrow. In FIG. 20, (a) shows the state of the wiring 210 in a normal state where no failure has occurred. On the other hand, (b) shows a state in which a part of the wiring 210 is burned down and becomes thin as indicated by reference numeral 211. As the burnout portion 211 is increased in resistance, a local increase in resistance occurs in the GMR element.

  As an example, consider the case where burnout occurs in the GMR element R1 in the SIN bridge 40 shown in FIG. In this case, the local resistance value (bulk resistance value) increases because the wiring pattern of the burned-out portion 211 becomes thinner as shown in FIG. 20 and the cross-sectional area decreases. As a result, the overall resistance value of the GMR element R1 increases. On the other hand, the magnetoresistive effect in the GMR element R1 is caused by scattering of electrons at the interface between the layers (the free magnetic layer 21, the spacer layer 22, and the fixed magnetic layer 23) in the entire wiring 210. Therefore, even if the local bulk resistance value increases in the burned portion 211 as described above, the amount of change in resistance value due to the magnetoresistive effect is not significantly affected. In other words, the resistance component that increases due to wiring burnout is a component that does not depend on the magnetic field direction. When the increase rate of the magnetic field direction independent term is b times, the resistance value of the GMR element R1 can be expressed as the following expression (11) from the above expression (5).

（１１）

(11)

  In this case, the difference ΔVs = Vs2−Vs1 between the terminals V1 and V2 in the SIN bridge 40 is expressed by the following equation (12), assuming that Rn0 = R0.

（１２）

(12)

  FIG. 21 shows a Lissajous locus by the output signal of the magnetic field sensor 201 in the present embodiment. This Lissajous locus is based on the difference ΔVs proportional to sin θ expressed by the equation (12) and the difference ΔVc proportional to cos θ expressed by the above equation (7).

The following (a) to (c) can be seen from the Lissajous locus in FIG.
(A) When b <5, the Lissajous waveform includes the origin. That is, the angle deviation is smaller than 90 degrees.
(B) When b = 5, the Lissajous waveform passes over the origin.
(C) When b> 5, the Lissajous waveform does not pass through the origin. That is, the angle deviation is larger than 90 degrees.

  Based on the above results, the microcomputer 10 monitors the Lissajous locus of (ΔVc, ΔVs), and when this does not include the origin, the Lissajous locus quadrant determination 122 determines that the magnetic field sensor 201 is abnormal. .

  In the embodiment described above, a signal upper limit / lower limit determination circuit for detecting an abnormality by detecting whether or not the difference ΔVc, ΔVs deviates from a predetermined upper limit and lower limit is provided in the rotation angle detection device. It may be. An example of this signal upper / lower limit determination circuit is shown in FIG. In the circuit shown in FIG. 22, the upper limit value and the lower limit value of the abnormality detection range set according to the resistance values of the divided resistors 391A to 391C, and the differential output ΔVc = Vc2−Vc1 from the COS bridge 30 of the magnetic field sensor 201. Are compared by the comparators 393A and 393B. The outputs of the comparators 393A and 393B are input to an OR circuit (OR circuit) 394. The OR circuit 394 outputs an abnormality detection signal when ΔVc exceeds either the upper limit value or the lower limit value. Thereby, it is determined whether or not ΔVc exceeds a certain range (abnormality detection range).

  In the above description, the output signal ΔVc of the COS bridge 30 has been described, but the output signal ΔVs of the SIN bridge 40 is similarly determined. At this time, the signals ΔVc and ΔVs may be amplified and then input to the comparators 393A and 393B. In this case, the upper and lower limit values are appropriately set according to the amplification factor.

  In the circuit of FIG. 22, for example, the abnormality detection range is set to a range of ± 1 V centered on the midpoint potential between the excitation voltage e0 and the g terminal voltage (for example, 0 V) of each bridge circuit. Further, the excitation voltage e0 is set to 5V, and en is set to -2V.

  The relationship between the upper / lower limit determination of the signal described above and the quadrant management criteria (criteria) based on the Lissajous locus shown in FIG. 21 will be described. As described above, the abnormal signal is notified based on the Lissajous quadrant management criterion when b> 5, that is, when the resistance value increases more than five times. On the other hand, when the resistance value becomes five times, since the change width of the difference ΔVc or ΔVs is +0.8 V or −0.8 V from the above-described midpoint potential, no abnormality is determined in the circuit of FIG. That is, the Lissajous waveform quadrant management standard determines the abnormality more strictly than the signal upper limit / lower limit determination standard.

  Thus, it is preferable to combine the signal upper limit / lower limit determination criterion and the Lissajous waveform quadrant management criterion because an abnormality of the rotation angle detection device can be detected at an early stage.

  In the signal upper limit / lower limit determination circuit shown in FIG. 22, as described above, the upper limit value and the lower limit value of the abnormality detection range are set by the dividing resistors 391A to 391C using the excitation voltage e0 of the magnetoresistive element bridge as a reference voltage. ing. As shown in the equation (7), the signal voltage ΔVc is proportional to the excitation voltage e0. Therefore, such a configuration is preferable because an abnormality can be detected in an abnormality detection range proportional to the fluctuation of the excitation voltage e0.

  In the present embodiment described above, the case where the GMR element is used as the magnetoresistive element has been described. However, other magnetoresistive elements such as anisotropic magnetoresistive elements (AMR elements), tunnel magnetoresistive elements (TMR elements) Needless to say, the present invention is also effective for magnetoresistive elements such as When TMR is used, the magnetic field sensor 201 is provided with a bridge configuration composed of TMR elements.

  As described above, since the magnetic field sensor 201 is used instead of the resolver 1 in this embodiment, the apparatus can be reduced in size and weight as compared with the resolver. In addition, high manufacturing accuracy and assembly accuracy required in the case of a resolver are not required, and the size of the magnetic field sensor 201 can be made constant regardless of the thickness of the rotor shaft, thereby reducing the cost of the apparatus. Can be realized.

  FIG. 23 is a diagram illustrating an example in which the rotation angle sensor abnormality detection device according to the fifth embodiment described above is applied to a motor control system. FIG. 24 is a diagram showing an example in which the rotation angle sensor abnormality detection device according to the fifth embodiment described above is applied to electric power steering. The examples shown in these drawings are the same as those in FIGS. 8 and 9 except that the resolver 1, the excitation signal generator 2 and the conversion trigger generator 3 in FIGS.

  According to the fifth embodiment described above, the same operational effects as those described in the first embodiment can be obtained. In addition, since a magnetic field sensor is used instead of the resolver, the apparatus can be reduced in size, weight, and cost.

  In the fifth embodiment described above, an example in which the magnetic field sensor 201 is provided instead of the resolver 1 in the first embodiment has been described. Similarly, the second to fourth embodiments are described. The resolver 1 in the form may be replaced with the magnetic field sensor 201. Also in this case, the same effects as those described in the second to fourth embodiments can be obtained. In addition, the apparatus can be reduced in size, weight, and cost as compared with the second to fourth embodiments using a resolver.

  Each embodiment and various modifications described above are merely examples. Therefore, the present invention is not limited to these contents and configurations.

DESCRIPTION OF SYMBOLS 1 Resolver 2 Excitation signal production | generation part 3 Conversion trigger production | generation part 4 Motor driver 5 Motor 6 Control object 7 Rotating shaft 8 Steering wheel 9 Torque sensor 10 Microcomputer 11 A / D converter 12 Error detection function 13 Calculation part 14A, 14B Detection circuit 15 Positive output circuit 16 Negative output circuit 20 External magnetic field direction 21 Free magnetic layer 22 Spacer layer 23 Fixed magnetic layer 30 COS bridge 40 SIN bridge 100 Motor unit 101 Frame 102 First bracket 103 Second bracket 106, 107 Bearing 110 Stator 111 Stator core 112 Stator coil 121 Lissajous locus amplitude determination 122 Lissajous locus quadrant determination 123 AND (logical product)
124 OR (logical sum)
130 rotor 200 rotation angle detector 201 magnetic field sensor 202 sensor magnet 203 housing

Claims (6)

A magnetic field sensor that detects a magnetic field change caused by rotation of the rotation shaft, and outputs a sine signal corresponding to a sine of the rotation angle of the rotation shaft and a cosine signal corresponding to a cosine of the rotation angle of the rotation shaft;
A rotation angle sensor abnormality detection device comprising abnormality detection means for detecting an abnormality of the magnetic field sensor based on a Lissajous waveform of the sine signal and the cosine signal output from the magnetic field sensor;
The abnormality detecting means detects the presence or absence of abnormality of the magnetic field sensor based on the amplitude of the Lissajous waveform and the quadrant through which the Lissajous waveform passes, and the amplitude of the Lissajous waveform is a predetermined first detection. A rotation angle sensor abnormality detection device that detects an abnormality of the magnetic field sensor when there is a quadrant in which the Lissajous waveform does not pass within a predetermined time. A magnetic field sensor that detects a magnetic field change caused by rotation of the rotation shaft, and outputs a sine signal corresponding to a sine of the rotation angle of the rotation shaft and a cosine signal corresponding to a cosine of the rotation angle of the rotation shaft;
A rotation angle sensor abnormality detection device comprising abnormality detection means for detecting an abnormality of the magnetic field sensor based on a Lissajous waveform of the sine signal and the cosine signal output from the magnetic field sensor;
The abnormality detecting means detects the presence or absence of abnormality of the magnetic field sensor based on the amplitude of the Lissajous waveform and the quadrant through which the Lissajous waveform passes, and the amplitude of the Lissajous waveform is a predetermined first detection. If the Lissajous waveform does not pass through any one quadrant within a predetermined time and interpolates the Lissajous waveform in the quadrant based on the Lissajous waveforms in the two quadrants adjacent to the quadrant, A rotation angle sensor abnormality detecting device, wherein an abnormality of the magnetic field sensor is detected when a Lissajous waveform does not surround the origin. In the rotation angle sensor abnormality detection device according to claim 2 ,
The abnormality detection means interpolates the Lissajous waveform of the quadrant by connecting the trajectories of the Lissajous waveforms in the two adjacent quadrants with a straight line,
A rotation angle sensor abnormality detecting device, wherein it is determined whether or not the interpolated Lissajous waveform surrounds the origin based on the positive or negative of the y-intercept of the straight line. In the rotation angle sensor abnormality detection device according to any one of claims 1 to 3 ,
When the amplitude of the Lissajous waveform is within a predetermined second detection range, the abnormality detecting means detects the abnormality of the magnetic field sensor regardless of whether there is a quadrant in which the Lissajous waveform does not pass within a predetermined time. A rotation angle sensor abnormality detection device characterized by detecting the rotation angle sensor. The rotation angle sensor abnormality detection device according to any one of claims 1 to 4 ,
A motor coupled to the rotating shaft;
A magnet installed on the rotating shaft;
A motor drive circuit for driving the motor;
A motor drive control circuit that controls driving of the motor by the motor drive circuit based on the detection result of the rotation angle;
The magnetic field sensor of the rotation angle sensor abnormality detection device detects a magnetic field change caused by a change in the rotational position of the magnet according to the rotation of the rotating shaft,
When an abnormality of the magnetic field sensor is detected by the abnormality detection means, the motor control system prohibits driving of the motor according to the detection result. The rotation angle sensor abnormality detection device according to any one of claims 1 to 5 ,
A motor coupled to the rotating shaft;
A magnet installed on the rotating shaft;
A motor drive circuit for driving the motor;
A wheel steering mechanism mechanically connected to the motor via a column shaft;
A steering wheel mechanically connected to the column shaft;
A torque sensor for detecting an operation input via the steering wheel;
A motor drive control circuit for controlling the drive of the motor by the motor drive circuit based on the detection result of the operation by the torque sensor and the detection result of the rotation angle;
The magnetic field sensor of the rotation angle sensor abnormality detection device detects a magnetic field change caused by a change in the rotational position of the magnet according to the rotation of the rotating shaft,
An electric power steering system characterized in that, when an abnormality of the magnetic field sensor is detected by the abnormality detection means, driving of the motor is prohibited according to the detection result.

Images