JP5195787B2 - Rotation sensor - Google Patents

Description

  The present invention relates to a rotation sensor in which a magnetoelectric conversion element is arranged in a magnetic field of a magnetic generation unit, and a relative rotation angle of the magnetic generation unit is calculated using a signal output from the magnetoelectric conversion element.

  Conventionally, as this type of rotation sensor, one shown in FIG. 33 is known (Patent Document 1). As shown in FIG. 33A, this is configured by arranging one magnetic sensor 100 and a Hall element 101 on the surface of the printed circuit board 104. The magnetic sensor 100 is disposed at a position facing the rotating surface of the permanent magnet 107 so that a magnetic field 106 generated from the permanent magnet 107 and parallel to the surface of the printed circuit board 104 can be detected. The Hall element 101 is arranged outside the permanent magnet 107 so that a magnetic field 105 generated from the permanent magnet 107 and perpendicular to the surface of the printed circuit board 104 can be detected. As shown in FIG. 33B, the Hall element 101 includes two horizontal Hall elements 102 and 103. The horizontal Hall elements 102 and 103 are arranged at an angle γ in the plane direction with respect to the magnetic sensor 100.

  When the permanent magnet 107 rotates once, the magnetic sensor 100 outputs a signal with one wavelength having an electrical angle of 180 °, and the horizontal Hall elements 102 and 103 each output a signal with one wavelength having an electrical angle of 360 °. . Further, the relative rotation angle α of the permanent magnet 107 is obtained in an angle range of 180 ° at the maximum based on the output signal of the magnetic sensor 100, and based on the magnitude relationship between the output values of the horizontal Hall elements 102 and 103, the permanent magnet It is determined to which range the relative rotation angle α of 107 belongs to 0 ° ≦ α ≦ 90 °, 90 ° ≦ α ≦ 180 °, 180 ° ≦ α ≦ 270 ° and 270 ° ≦ α ≦ 360 °. .

JP-A-11-94512 (paragraphs 22 to 24, FIGS. 4 and 5).

However, in the above-described conventional one, it is necessary to always determine which range the relative rotation angle α of the permanent magnet 107 belongs while the permanent magnet 107 is rotating.
Therefore, the above-described conventional device has a problem that it takes time to calculate the relative rotation angle α of the permanent magnet 107.

  Accordingly, the present invention has been made to solve the above-described problems, and an object of the present invention is to realize a rotation sensor that can shorten the calculation time of the relative rotation angle of the magnetism generator.

  In order to achieve the above object, the first feature of the present invention is that it is arranged in the magnetic field (B1) of the relatively rotating magnetic generator (2), and the magnetic field is rotated during one rotation of the magnetic generator. A plurality of magnetoelectric transducers (M1, M2) arranged so as to output signals (sin 2θ, cos 2θ) whose signal levels change in two cycles according to the intensity of the signals and to produce a phase difference between the signals. A rotation sensor (1) configured to obtain a relative rotation angle (θ) with respect to the magnetism generation unit using a signal output from each magnetoelectric conversion element, while the magnetism generation unit makes one rotation, A plurality of detection elements (H1, H2) arranged so as to output detection signals (sin θ, cos θ) whose signal levels change in one cycle according to the intensity of the magnetic field and to produce a phase difference between the detection signals. And each magnetoelectric transducer The angle at which the relative rotation angle is calculated by performing feedback control so that the deviation between the relative rotation angle with respect to the magnetism generating portion and the calculation angle (φ) obtained by calculation converges to a predetermined value Using the calculation unit (60) and each comparison result between each signal level and each threshold value of each detection signal output from each detection element, an angle range including the initial value of the relative rotation angle is determined, and the determined angle The initial value for determining the initial value of the calculation angle so that the absolute value of the difference between the initial value (θ0) of the relative rotation angle that can occur in the range and the initial value (φ0) of the calculation angle is less than 90 °. A value determination unit (53), and an output unit (70) that outputs a signal corresponding to the calculation angle calculated by the angle calculation unit in one cycle while the magnetism generation unit makes one rotation, The initial value determination unit includes the magnetic generation unit. The initial value of the calculation angle is determined only before starting relative rotation, and the angle calculation unit starts the feedback control using the initial value of the calculation angle determined by the initial value determination unit, and It is to calculate the rotation angle.

Since the initial value determination unit determines the initial value of the calculation angle only before the magnetism generation unit starts relative rotation, the calculation time of the relative rotation angle when the magnetism generation unit is relatively rotating can be shortened. it can.
That is, conventionally, during the relative rotation of the magnetism generating unit, it has been necessary to determine the angle range including the relative rotation angle using the signal level of each detection signal from each detection element. According to the above feature, the signal level of each detection signal from each detection element is used only when the initial value of the calculation angle is determined before the magnetism generator starts relative rotation. During the relative rotation, it is not necessary to compare the signal level of each detection signal from each detection element with the threshold value, so that the calculation time of the relative rotation angle can be shortened.

  The second feature of the present invention is that it is arranged in the magnetic field (B1) of the magnetism generator (2) that rotates relatively, and the signal level depends on the strength of the magnet field during one rotation of the magnetism generator. A plurality of magnetoelectric transducers (M1, M2) arranged so as to output signals (sin 2θ, cos 2θ) that change in two cycles and a phase difference between the signals are output from each magnetoelectric transducer. In the rotation sensor (1) configured to obtain a relative rotation angle (θ) with respect to the magnetic generation unit using the generated signal, a signal is generated according to the strength of the magnetic field during one rotation of the magnetic generation unit. A detection signal (sin θ, cos θ) whose level changes in one cycle is output, and a plurality of detection elements (H1, H2) arranged so as to produce a phase difference between the detection signals, and output from each magnetoelectric conversion element Used before and after An angle calculation unit (60) for calculating the relative rotation angle by performing feedback control so that a deviation between a relative rotation angle with respect to the magnetism generating unit and a calculation angle (φ) obtained by calculation converges to a predetermined value; Using each comparison result between each signal level of each detection signal output from the detection element and the threshold value, an angle range including the initial value (θ0) of the relative rotation angle is determined, and the occurrence occurs within the determined angle range. An initial value determination unit (53) for determining an initial value of the calculation angle so that an absolute value of a difference between an initial value of a possible relative rotation angle and an initial value (φ0) of the calculation angle is less than 90 °; An output unit (70) that outputs a signal corresponding to the calculation angle calculated by the angle calculation unit in one cycle while the magnetism generation unit makes one rotation, and the initial value determination unit includes: Before the magnetism generator starts relative rotation, An initial value of the calculation angle is determined at a predetermined time after starting the counter rotation, and the angle calculation unit uses the initial value of the calculation angle determined by the initial value determination unit to perform the feedback. The control is to calculate the relative rotation angle.

The initial value determination unit determines the initial value of the calculation angle before the magnetism generation unit starts relative rotation and at a predetermined time after the relative rotation starts. The calculation time of the relative rotation angle can be shortened.
That is, conventionally, during the relative rotation of the magnetism generating unit, it is necessary to always determine the angle range including the relative rotation angle by using the signal level of each detection signal from each detection element. According to the feature, the signal level of each detection signal from each detection element is the initial value of the calculation angle before the magnetism generator starts relative rotation and at a predetermined time after the relative rotation starts. It is only used when determining the value, and it is not necessary to compare the signal level of each detection signal from each detection element and the threshold each time the magnetism generator is relatively rotated. It can be shortened.

  According to a third feature of the present invention, in the first or second feature described above, when the initial value of the relative rotation angle is θ0 and the initial value of the calculation angle is φ0, the angle calculation unit (60 ) Is that the initial value θ0 of the relative rotation angle existing in the range of (φ0−90 °) <θ0 <(φ0 + 90 °) can be calculated.

  According to the above third feature, even if the initial value of the relative rotation angle does not exist in the angle range determined by the initial value determination unit and actually exists in a position beyond the angle range. The angle calculator can accurately calculate the relative rotation angle.

  For example, as shown in FIG. 22, when the angle range determined by the initial value determination unit is 0 ≦ θ0 <90 ° and the initial value φ0 of the calculation angle is determined to be 45 °, (45 ° − 90 °) <θ0 <(45 ° + 90 °), that is, the initial value θ0 of the relative rotation angle is calculated up to −45 ° <θ0 <135 ° (0 ° ≦ θ0 <135 ° and 315 ° <θ0 ≦ 360 °). The possible range can be expanded.

  For this reason, for example, the angle range determined by the initial value determination unit is 0 ≦ θ0 <90 °, but even if the actual initial value θ0 is 120 °, it corresponds to the initial value φ0 = 45 °. Since the angle range to include includes 120 °, the angle calculation unit can calculate the initial value θ0 = 120 ° of the relative rotation angle θ using the initial value φ0 = 45 ° of the calculation angle φ.

  Therefore, for example, even when the initial value θ0 or the initial value φ0 changes due to the influence of external noise or an external magnetic field, the relative rotation angle θ can be accurately calculated. That is, it is possible to realize a rotation sensor in which the detection accuracy is unlikely to deteriorate even under the influence of external noise or an external magnetic field.

  According to a fourth feature of the present invention, in any one of the first to third features described above, the plurality of detection elements (H1, H2) have a phase difference of 90 ° between the detection signals. It is in being arranged in.

According to the fourth feature, each time the relative rotation angle with respect to the magnetism generator changes by 90 °, detection signals having phases different by 90 ° can be output from the respective detection elements.
That is, each time the relative rotation angle changes by 90 °, the combination of the signal levels of the detection signals can be changed.
Therefore, the initial value determination unit can determine the angle range in which the initial value of the relative rotation angle exists in units of 90 ° by using the combination of the signal levels of the detection signals.

  According to a fifth feature of the present invention, in any one of the first to fourth features described above, the plurality of magnetoelectric transducers (M1, M2) have a 45 ° phase difference between the signals. It is in being arranged in.

According to the fifth feature, a sine wave signal (sin signal) is output from one of the magnetoresistive elements, and a cosine wave signal (cos signal) whose phase is delayed by 45 ° from the sine wave signal is output from the other. be able to.
Therefore, the relative rotation angle can be calculated using the above sine wave signal and cosine wave signal.

  A sixth feature of the present invention is that, in any one of the first to fifth features described above, the relative rotation angle (θ) of 0 to 360 ° is an output signal of each detection element (H1, H2). Each value of each detection signal is compared with a threshold value when n angle ranges are set by dividing the phase difference between them by n and dividing the range of 0 to 360 ° by n. The combination of the results is configured to be different in each angle range.

  According to the sixth feature, the combination of the comparison results between the signal levels of the detection signals and the threshold values is totally different in each angle range, so that the accuracy determination of the angle range in the initial value determination unit can be improved. it can.

  A seventh feature of the present invention is that, in any one of the first to sixth features described above, when the relative rotation angle is θ and the calculation angle is φ, the angle calculation unit (60) Performs feedback control using the signals (sin2θ, cos2θ) output from the magnetoelectric transducers (M1, M2) so that the deviation (2θ-2φ) becomes zero, and starts the feedback control. The initial value determined in the initial value determination unit (53) is used as the initial value (φ0) of the calculation angle φ.

  According to the seventh feature, since the relative rotation angle is calculated by performing feedback control so that the deviation (2θ−2φ) becomes zero, the relative rotation angle can be calculated with high accuracy.

  According to an eighth feature of the present invention, in the seventh feature described above, the plurality of magnetoelectric transducers (M1, M2) output a sin 2θ signal and a cos 2θ signal, and the angle calculator (60) is configured to output the sin 2θ. Sin (2θ-2φ) is created based on the signal and the cos 2θ signal, and a deviation (2θ-2φ) is calculated based on the created sin (2θ-2φ), and the deviation (2θ-2φ) is calculated The relative rotation angle is calculated by performing feedback control so as to be zero.

  According to the eighth feature, the deviation (2θ-2φ) can be calculated using the sin 2θ signal and the cos 2θ signal output from each magnetoelectric conversion element. That is, since sin (2θ-2φ) = sin2θcos2φ−cos2θsin2φ, a circuit for multiplying the sin2θ signal and the cos2θ signal, the sin2φ signal and the cos2φ signal output from each magnetoelectric transducer, and a circuit for subtracting the signal Can be used to calculate the deviation (2θ-2φ).

  According to a ninth feature of the present invention, in the eighth feature described above, the angle calculation unit (60) includes a counter that counts a count value corresponding to the calculation angle φ, and the deviation (2θ-2φ) The determination is positive or negative, and the count value of the counter is increased or decreased based on the determination result.

  According to the ninth feature, since the calculation angle φ can be made to correspond to the count value of the counter, the calculation angle φ can be calculated with high accuracy.

  According to a tenth feature of the present invention, in the ninth feature described above, the angle calculator (60) calculates the deviation (2θ-2φ) by performing an arc sine operation on the sin (2θ-2φ) signal. The positive / negative is determined based on the calculation result.

  According to the tenth feature, the deviation (2θ-2φ) is calculated by performing an inverse sine operation on the sin (2θ-2φ) signal, and the sign of the deviation (2θ-2φ) is determined based on the calculation result. can do.

  According to an eleventh feature of the present invention, in the ninth feature described above, when the sin (2θ-2φ) signal is greater than 0, the angle calculator (60) has a positive deviation (2θ-2φ). When the sin (2θ-2φ) signal is smaller than 0, it is determined that the deviation (2θ-2φ) is negative.

  According to the eleventh feature described above, since the sign (2θ-2φ) can be determined whether the sin (2θ-2φ) signal is greater than 0, the sign (2θ-2φ) can be determined. There is no need to perform an arc sine operation on the signal.

  A twelfth feature of the present invention is that, in any one of the first to eleventh features described above, the plurality of detection elements (H1, H2) are Hall elements.

  According to the twelfth feature, it is possible to obtain a detection signal whose signal level changes in one cycle according to the strength of the magnetic field while the magnetism generator rotates once. In addition, since the Hall element can be built in the semiconductor substrate, the rotation sensor can be downsized.

  A thirteenth feature of the present invention is that, in any one of the first to twelfth features described above, each of the plurality of magnetoelectric transducers (M1, M2) is a magnetoresistive element.

  According to the thirteenth feature described above, a signal whose signal level changes in two cycles according to the strength of the magnetic field can be obtained while the magnetism generator rotates once. In addition, since the magnetoresistive element can be built in the semiconductor substrate, the rotation sensor can be miniaturized.

  In addition, the code | symbol in each said parenthesis shows the correspondence with the specific means as described in embodiment mentioned later.

It is explanatory drawing which shows the main structures of the rotation sensor of 1st Embodiment with a block. It is explanatory drawing which shows an example of the use condition of the sensor chip shown in FIG. 1, (a) is a longitudinal cross-sectional view of a sensor chip and a permanent magnet, (b) is a top view of the permanent magnet shown to (a). It is a longitudinal cross-sectional view which shows the state which the permanent magnet shown to Fig.2 (a) rotated 180 degrees. It is explanatory drawing which shows the structure of a sensor chip typically, (a) is a top view, (b) is AA arrow sectional drawing of (a). (A) is a top view of the magnetoresistive element area | region E1 and Hall element area | region E2, (b) is explanatory drawing which shows the arrangement | positioning angle of Hall element H1, H2. It is a top view which shows typically the structure of AMR sensor M1. It is a top view which shows typically the structure of AMR sensor M2. It is an equivalent circuit of the AMR sensor M1. It is an equivalent circuit of the AMR sensor M2. It is explanatory drawing which shows each output signal of AMR sensor M1, M2 and Hall element H1, H2. It is explanatory drawing of Hall element H2, (a) is a top view which shows a part of Hall element H2 and its periphery, (b) is AA arrow sectional drawing of (a), (c) is (a). It is BB arrow sectional drawing of. It is explanatory drawing which shows the main electrical structures of the rotation sensor 1 with a block, and is a figure corresponding to FIG. It is explanatory drawing which shows the flow of the signal between each block shown in FIG. It is explanatory drawing which shows the structure of the initial value table 54b shown in FIG. It is explanatory drawing which shows output waveforms, such as a Hall element, (a) is an output waveform of Hall element H1, (b) is an output waveform of comparison circuit 53a, (c) is an output waveform of Hall element H2, (d) is The output waveform of the comparison circuit 53b, (e) is the output waveform of the output unit 70. It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the process in which the initial value (phi) 0 of the calculation angle (phi) follows the initial value (theta) 0 of relative rotation angle (theta). It is explanatory drawing which shows the margin of initial value (theta) 0 corresponding to initial value (phi) 0. It is explanatory drawing which shows typically the structure of the sensor chip with which the rotation sensor of 2nd Embodiment was equipped. It is explanatory drawing which shows each output signal of AMR sensor M1, M2 and Hall element H1, H2. It is explanatory drawing which shows the structure of the initial value table 54b. It is explanatory drawing which shows the margin of initial value (theta) 0 corresponding to initial value (phi) 0. It is explanatory drawing which shows typically the structure of the sensor chip with which the rotation sensor of 3rd Embodiment was equipped. It is explanatory drawing which shows each output signal of AMR sensor M1, M2 and Hall element H1-H3. It is explanatory drawing which shows the structure of the initial value table 54b. It is explanatory drawing which shows the margin of initial value (theta) 0 corresponding to initial value (phi) 0. It is explanatory drawing which shows the structure of the initial value table 54b in 4th Embodiment. It is explanatory drawing which shows the margin of initial value (theta) 0 corresponding to initial value (phi) 0. It is explanatory drawing of the conventional rotation sensor.

<First Embodiment>
Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is an explanatory diagram showing the main configuration of the rotation sensor of this embodiment in blocks. 2A and 2B are explanatory views showing an example of a usage state of the sensor chip shown in FIG. 1, wherein FIG. 2A is a longitudinal sectional view of the sensor chip and the permanent magnet, and FIG. 2B is a plan view of the permanent magnet shown in FIG. It is. FIG. 3 is a longitudinal sectional view showing a state in which the permanent magnet shown in FIG.

[Main configuration]
The main configuration of the rotation sensor of this embodiment will be described. As shown in FIG. 1, the rotation sensor 1 of this embodiment includes a sensor chip 5 and a detection circuit 50 electrically connected to the sensor chip 5. The sensor chip 5 includes two anisotropic magnetoresistive sensors (hereinafter referred to as AMR sensors) M1 and M2 made of magnetoresistive elements and two Hall elements H1 and H2.

As shown in FIG. 2A, the sensor chip 5 is arranged at a position facing the rotating surface 2c along the diameter of the permanent magnet (magnetism generating unit) 2 to be detected. The sensor chip 5 is supported by a support member (not shown), and is fixed so that the arrangement position does not change.
As shown in FIG. 2 (b), the permanent magnet 2 has a disk shape, one of which is divided into the same size in the radial direction is an N pole permanent magnet 2a and the other is an S pole. It is a permanent magnet 2b. As shown in FIG. 2A, the permanent magnet 2 is attached to the tip of the rotating shaft 3 and rotates in the direction indicated by the arrow F1.

  The permanent magnet 2 generates a magnetic field from the N-pole permanent magnet 2 a toward the S-pole permanent magnet 2 b, and generates a magnetic field B 1 parallel to the surface 5 a of the sensor chip 5. In the illustrated example, the magnetic field B1 passes through the Hall element H2 from the Hall element H1. When the permanent magnet 2 rotates 180 ° from the state shown in FIG. 2A as shown in FIG. 3, the direction of the magnetic field B1 changes by 180 °. In the illustrated example, the magnetic field B1 passes through the Hall elements H2 to H1.

As shown in FIG. 1, the detection circuit 50 includes amplification units 51 and 52, an initial value determination unit 53, an angle calculation unit 60, and an output unit 70. The amplifying unit 51 amplifies output signals output from the AMR sensors M1 and M2. The angle calculation unit 60 calculates the relative rotation angle θ of the permanent magnet 2 using the amplified signal output from the amplification unit 51.
The amplifying unit 52 amplifies output signals output from the Hall elements H1 and H2. The initial value determining unit 53 compares each amplified signal output from the amplifying unit 52 with a threshold value, and the initial value θ0 of the relative rotation angle θ of the permanent magnet 2 is determined from how many times to how many times based on each comparison result. The initial value φ0 of the calculation angle φ corresponding to the determined angle range is determined.

  The output unit 70 receives the calculation angle φ calculated by the angle calculation unit 60 and outputs a linear characteristic signal having a voltage Vo corresponding to the calculation angle φ in one cycle while the permanent magnet 2 makes one rotation. To do.

[Sensor chip structure]
The structure of the sensor chip 5 will be described. 4A and 4B are explanatory views schematically showing the structure of the sensor chip, in which FIG. 4A is a plan view and FIG. 4B is a cross-sectional view taken along line AA in FIG. FIG. 5A is a plan view of the magnetoresistive element region E1 and the Hall element region E2, and FIG. 5B is an explanatory diagram showing the arrangement angles of the Hall elements H1 and H2. In each drawing, the dimensions of the hall elements H1 and H2 are drawn larger than the actual dimensions for easy understanding. Also, the shape of the magnetoresistive element is drawn larger than the actual size in order to make the formation direction of the element easy to understand.

  As shown in FIG. 4, the sensor chip 5 includes a silicon substrate 10, an insulating film 90 formed on the surface of the silicon substrate 10, and AMR sensors M1 and M2 (magnetoelectric conversion) formed on the surface of the insulating film 90. Element) and Hall elements H1 and H2 (detection elements) formed in the silicon substrate 10. The AMR sensor M1 includes magnetoresistive elements R1 to R4, and the AMR sensor M2 includes magnetoresistive elements R5 to R8. The Hall elements H1 and H2 are arranged so as to overlap below the magnetoresistive elements R1 to R8 via the insulating film 90.

  As shown in FIG. 5B, the Hall elements H1 and H2 are arranged such that the angle formed by the magnetic detection surfaces HP1 and HP2 of each magnetic detection unit HP is 90 °. That is, the Hall elements H1 and H2 are arranged so that the phase difference between the output signals is 90 °. A line extending horizontally from the relative rotation center P1 of the sensor chip 5 toward the magnetoresistive element R2 is a reference line L3, a line parallel to the magnetic detection surface HP1 of the Hall element H1 is L4, and the position of the reference line L3 is a reference. When the angle is 0 °, the Hall element H1 is arranged such that an angle α formed by its magnetic detection surface HP1 and the reference line L3 is 90 °.

  In addition, the Hall element H2 is arranged so that its magnetic detection surface HP2 and the reference line L3 are parallel to each other. Further, the magnetic detection surface HP1 of the Hall element H1 and the easy axis of the magnetoresistive element R2 arranged at the reference angle of 0 ° form an angle of 90 °. That is, the Hall element H1 outputs a sin θ signal having the same phase with respect to the relative rotation angle θ, and the Hall element H2 outputs a cos θ signal whose phase is 90 ° different from that of the Hall element H1. The signal levels of the sin θ signal and the cos signal change in two periods according to the strength of the magnetic field while the permanent magnet 2 rotates once.

  Here, a region where the magnetoresistive elements R1 to R8 are disposed is referred to as a magnetoresistive element region E1, and a region where the Hall elements H1 and H2 are disposed is referred to as a Hall element region E2. FIG. 5A is created based on FIG. The magnetoresistive element region E1 has a quadrangular shape, and its area is substantially equal to the minimum area required for arranging the magnetoresistive elements R1 to R8. The hall element region E2 has a T-shape, and its area is substantially equal to the minimum area necessary for arranging the hall elements H1 and H2.

  As illustrated, the entire Hall element region E2 is overlapped below the magnetoresistive element region E1, and a part of the Hall element region E2 does not protrude from the end of the magnetoresistive element region E1. Further, the intersection of the diagonal lines L1 and L2 of the magnetoresistive element region E1 coincides with the relative rotation center P1 of the sensor chip 5.

  That is, the relative rotation center P1 of the sensor chip 5 is located on the extension line of the relative rotation axis C1 (FIG. 2A) of the permanent magnet 2, and the relative rotation center P1 of the sensor chip 5 and the permanent magnet 2 are relative to each other. The center of rotation exists on the same axis. For this reason, even when the sensor chip 5 is relatively rotated around the relative rotation center P1 in a state where the permanent magnet 2 is not rotating, the relative rotation angle of the sensor chip 5 with respect to the permanent magnet 2 can be detected.

Since the sensor chip 5 has the above-described structure, the sensor chip 5 has a size (horizontal width) in the substrate surface direction as compared with the conventional rotation sensor in which the AMR sensor and the Hall element are arranged in the substrate surface direction of the semiconductor substrate. Can be small.
The AMR sensors M1 and M2 and the hall element regions H1 and H2 are overlapped in a direction corresponding to the relative rotation axis C1 of the permanent magnet 2 (FIG. 2A).
Therefore, since the sensor chip 5 can be reduced in the direction of the relative rotation center of the permanent magnet 2, the space facing the rotation surface 2c of the permanent magnet 2 can be used effectively.

(Structure of AMR sensor)
Next, the structure of the AMR sensors M1 and M2 will be described. FIG. 6 is a plan view schematically showing the structure of the AMR sensor M1. FIG. 7 is a plan view schematically showing the structure of the AMR sensor M2. FIG. 8 is an equivalent circuit of the AMR sensor M1, and FIG. 9 is an equivalent circuit of the AMR sensor M2. FIG. 10 is an explanatory diagram showing output signals of the AMR sensors M1, M2 and the Hall elements H1, H2.

The magnetoresistive elements R <b> 1 to R <b> 8 are formed in a shape obtained by folding a plurality of band-shaped regions, that is, in a meander shape (meandering shape). The resistance values of the magnetoresistive elements R1 to R8 change mainly depending on the intensity and direction of the magnetic field parallel to the surface of the silicon substrate 10, and output a signal having a level corresponding to the resistance value. That is, the magnetoresistive elements R1 to R8 are elements that generate an anisotropic magnetoresistive effect.
In this embodiment, the magnetoresistive elements R1 to R8 are formed of a ferromagnetic metal thin film. As the ferromagnetic material, NiFe (permalloy), NiCo, or the like can be used. The ferromagnetic metal thin film can be formed by sputtering or vapor deposition.

  As shown in FIG. 6, the AMR sensor M1 includes four magnetoresistive elements R1 to R4. The magnetoresistive elements R <b> 1 to R <b> 4 are arranged so that the angle formed by the extending direction of the strip-shaped elements in the adjacent magnetoresistive elements is 90 °. In other words, the magnetoresistive elements R <b> 1 to R <b> 4 are arranged so that the current direction (magnetization easy axis) of the adjacent magnetoresistive elements forms an angle of 90 °. That is, each set of the magnetoresistive elements R1, R4 and R2, R3 is arranged such that the phase between the output signals is 90 ° different in each set.

  As shown in FIG. 8, the magnetoresistive elements R1 and R4 are electrically connected in series to form a half bridge circuit. An output terminal 31 for taking out the midpoint output Vout1 is electrically connected to the midpoint of the half bridge circuit. The magnetoresistive elements R2 and R3 are also electrically connected in series to form a half bridge circuit. An output terminal 32 for taking out the midpoint output Vout2 is electrically connected to the midpoint of the half bridge circuit.

  Both half bridge circuits are connected in parallel to form a full bridge circuit that outputs a cos 2θ signal. The full bridge circuit is electrically connected to a power supply terminal 30 for supplying power Vcc and a terminal 33 for electrically connecting to the ground G1. The magnetoresistive elements R1 and R2 facing each other in the full bridge circuit output (R0−ΔRcos2θ) signals, and the magnetoresistive elements R3 and R4 output (R0 + ΔRcos2θ) signals. Here, R0 is a resistance value of the magnetoresistive element in the absence of a magnetic field, and ΔR is a resistance value change amount.

Since each of the midpoint outputs Vout1 and Vout2 vibrates around Vcc / 2, it is possible to suppress an offset of the output waveform caused by a change in environmental temperature or the like.
The output terminals 31 and 32 are connected to a differential amplifier circuit (indicated by reference numeral 51a in FIG. 12), and the midpoint outputs Vout1 and Vout2 are differentially amplified. Therefore, the output amplitude of the AMR sensor M1 can be doubled compared to the case where the AMR sensor M1 is configured by one half bridge circuit, so that the magnetic detection sensitivity can be increased.

  As shown in FIG. 7, the AMR sensor M2 includes four magnetoresistive elements R5 to R8. The magnetoresistive elements R5 to R8 are arranged so that the angle formed by the extending direction of the strip-shaped elements in the adjacent magnetoresistive elements is 90 °. In other words, the magnetoresistive elements R5 to R8 are arranged so that the current direction (magnetization easy axis) of the adjacent magnetoresistive elements forms an angle of 90 °. That is, each set of magnetoresistive elements R5, R7 and R8, R6 is arranged such that the phase between the output signals is 90 ° different in each set.

  As shown in FIG. 9, the magnetoresistive elements R5 and R7 are electrically connected in series to form a half bridge circuit. An output terminal 37 for taking out the midpoint output Vout3 is electrically connected to the midpoint of the half bridge circuit. The magnetoresistive elements R8 and R6 are also electrically connected in series to form a half bridge circuit. An output terminal 38 for taking out the midpoint output Vout4 is electrically connected to the midpoint of the half bridge circuit.

  Both half-bridge circuits are connected in parallel to form a full-bridge circuit that outputs a sin 2θ signal. The full bridge circuit is electrically connected to a power supply terminal 36 for supplying power Vcc and a terminal 39 for electrically connecting to the ground G2. In the full bridge circuit, the magnetoresistive elements R5 and R6 facing each other output (R0 + ΔRsin2θ) signal, and the magnetoresistive elements R7 and R8 output (R0−ΔRsin2θ) signal.

Since each of the midpoint outputs Vout3 and Vout4 vibrates around Vcc / 2, it is possible to suppress an offset of the output waveform caused by a change in environmental temperature or the like.
The output terminals 37 and 38 are connected to a differential amplifier circuit (indicated by reference numeral 51b in FIG. 12), and the midpoint outputs Vout3 and Vout4 are differentially amplified. For this reason, the output amplitude of the AMR sensor M2 can be doubled as compared with the case where the AMR sensor M2 is configured by one half bridge circuit, so that the magnetic detection sensitivity can be increased.

  As shown in FIG. 4A, the magnetoresistive elements of the AMR sensors M1 and M2 are alternately arranged concentrically, and the magnetoresistive elements R1 to R4 of the adjacent AMR sensor M1 and the magnetoresistive elements of the AMR sensor M2. The resistance elements R5 to R8 are arranged so that the direction of current (magnetization easy axis) forms an angle of 45 °. The amount of change ΔR in the electrical resistance of the anisotropic magnetoresistive element is maximized when the angle between the direction of the current flowing in its own metal thin film (magnetization easy axis) and the direction of the magnetic field is 90 ° and 270 °. , 0 ° and 180 °.

  Therefore, as shown in FIG. 10, the AMR sensor M1 outputs a sin signal whose one wavelength has an electrical angle of 180 °, and the AMR sensor M2 has a phase difference of 45 ° with respect to the AMR sensor M1 and one wavelength has an electrical angle. A 180 ° cos signal is output.

  As shown in FIG. 4B, the magnetoresistive elements R1 to R8 constituting the AMR sensors M1 and M2 are arranged on the surface layer portion of the silicon substrate 10 through the insulating film 90. The AMR sensors M1 and M2 mainly detect a change in magnetic flux density of the magnetic field B1 parallel to the magnetoresistive element region E1, that is, the silicon substrate 10. The Hall elements H1 and H2 are built in the silicon substrate 10, and are disposed below the magnetoresistive elements R1 to R8 through the insulating film 90. In this embodiment, each of the Hall elements H1 and H2 is a vertical Hall element having a CMOS (Complementary Metal Oxide Semiconductor) structure. The insulating film 90 is a silicon oxide film.

  The Hall elements H1 and H2 are arranged so that the phase difference between the output signals is 90 °. For this reason, as shown in FIG. 9, when the permanent magnet 2 rotates 360 °, the Hall element H1 outputs a sin signal having an electrical angle of 360 °, and the Hall element H2 has an electrical angle of 360 °. The cos signal is output. In addition, when the permanent magnet 2 carries out 2 rotations or more, it treats that 360 degrees and 0 degrees are continuous.

(Hall element structure)
Next, the structure of the Hall elements H1 and H2 will be described. Since the Hall elements H1 and H2 have the same structure, the Hall element H2 will be described as an example here. 11A and 11B are explanatory diagrams of the Hall element H2, in which FIG. 11A is a plan view showing the Hall element H2 and a part of the periphery thereof, FIG. 11B is a cross-sectional view taken along the line AA in FIG. [FIG. 2] is a sectional view taken along line BB in FIG.

  The Hall element H2 has a high breakdown voltage CMOS transistor (HVCMOS) structure. The Hall element H2 includes a P-type (first conductivity type) silicon substrate (P-sub) 10 and an N-type (second conductivity type) semiconductor region formed in the depth direction from the surface layer portion of the silicon substrate 10. (Nwell) 91, a P-type (first conductivity type) diffusion layer (Pwell) 92 that surrounds the entire circumference of the semiconductor region 91, and a depth direction from the surface layer portion of the silicon substrate 10. P-type (first conductivity type) diffusion layers (Pwell) 93, 99 that divide a region from the surface layer portion to a predetermined depth into three semiconductor regions 91a, 91b, 91c, and respective surface layers of the semiconductor regions 91a, 91b, 91c Contact regions (N + diffusion layers (impurity diffusion regions)) 94 to 98 formed in the portion.

  Terminals S, V1, V2, G3, and G4 are electrically connected to the contact regions 94 to 98 through wiring. Terminals S, G3, and G4 are terminals for supplying a driving current, and terminals V1 and V2 are terminals for taking out a Hall voltage signal. That is, the contact regions 97 and 98 are current supply pairs, and the contact regions 95 and 96 are voltage output pairs. Therefore, the Hall elements H1 and H2 shown in FIG. 4A are arranged so that the lines connecting the current supply pairs are orthogonal to each other. Further, the lines connecting the voltage output pairs are arranged so as to be orthogonal to each other.

As shown in FIG. 11C, a region sandwiched between the contact regions 95 and 96 is a magnetic detection unit (hole plate) HP. Further, in the magnetic detection part HP, both surfaces parallel to the line connecting the contact regions 95 and 96 become the magnetic detection surface HP2, respectively. That is, the Hall element H2 outputs the Hall voltage signal corresponding to the magnetic field applied from the magnetic detection surface HP2 to the magnetic detection unit HP from the terminals V1 and V2.
As shown in FIG. 11B, when a constant drive current i is supplied from the terminal S to the terminal G3 and from the terminal S to the terminal G4, the drive current i is supplied from the contact region 94 to the magnetic detection unit HP, Then, it flows to the contact regions 97 and 98 through the semiconductor region 91 below the diffusion layers 93 and 98, respectively.

  That is, a drive current including a component perpendicular to the substrate surface (sensor chip surface) flows through the magnetic detection unit HP. For this reason, a magnetic field (for example, a magnetic field indicated by an arrow B1 in FIG. 11C) including a component parallel to the substrate surface (sensor chip surface) is applied to the magnetic detection unit HP in a state where the drive current flows. The Hall voltage VH corresponding to the magnetic field is generated between the terminals V1 and V2 by the Hall effect. The Hall voltage VH changes according to the angle formed by the magnetic detection surface HP2 and the direction of the magnetic field, that is, the incident angle of the magnetic field with respect to the magnetic detection surface HP2.

  As shown in FIGS. 2A and 3, the Hall elements H <b> 1 and H <b> 2 are arranged so that the magnetic detection surfaces HP <b> 1 and HP <b> 2 are perpendicular to the surface of the silicon substrate 10. 2 and a magnetic field B1 parallel to the surface of the silicon substrate 10 penetrates the magnetic detection surfaces HP1 and HP2 vertically. In the illustrated state, the magnetic field B1 penetrates perpendicularly to the magnetic detection surface HP2 of the Hall element H2, but when the permanent magnet 2 rotates 90 ° from the illustrated position, the magnetic field B1 changes to the magnetic detection surface of the Hall element H1. It penetrates perpendicularly to HP1. That is, the Hall elements H1 and H2 mainly detect changes in the magnetic flux density of the magnetic field B1 parallel to the surface of the silicon substrate 10.

  The N-type semiconductor region 91 is formed deeper than the N-type semiconductor region in the low breakdown voltage CMOS transistor structure, and accordingly, the P-type diffusion layers 92, 93, 98 are also P-type in the low breakdown voltage CMOS transistor structure. It is formed deeper than the diffusion layer. In this embodiment, the P-type diffusion layers 92, 93 and 98 are each formed to a depth approximately half that of the N-type semiconductor region 91.

Thus, since the Hall element H1 has the N-type semiconductor region 91 formed deeply, the carrier mobility is increased and the Hall effect can be increased, so that the Hall voltage VH can be increased. The detection sensitivity with respect to the magnetic field can be increased.
In addition, since the Hall element H1 is manufactured by a CMOS process, it is more cost-effective than a vertical Hall element manufactured by a bipolar process.

[Electrical configuration]
Next, the main electrical configuration of the rotation sensor 1 will be described. FIG. 12 is an explanatory diagram showing the main electrical configuration of the rotation sensor 1 in blocks, and corresponds to FIG. FIG. 13 is an explanatory diagram showing a signal flow between the blocks shown in FIG. FIG. 14 is an explanatory diagram showing the configuration of the initial value table 54b shown in FIG. FIGS. 15A and 15B are explanatory diagrams showing output waveforms of the Hall element, where FIG. 15A shows the output waveform of the Hall element H1, FIG. 15B shows the output waveform of the comparison circuit 53a, and FIG. 15C shows the output waveform of the Hall element H2. (D) is an output waveform of the comparison circuit 53b, and (e) is an output waveform of the output unit.

(Amplifier 52 and initial value determiner 53)
The amplification unit 52 includes amplification circuits 52a and 52b. The amplifier circuit 52a amplifies the detection signal sin θ output from the Hall element H1, and the amplifier circuit 52b amplifies the detection signal cos θ output from the Hall element H2. The initial value determining unit 53 includes comparison circuits 53a and 53b, an initial value reading unit 53c, and an initial value table 53d.
The comparison circuit 53a compares the signal level VH1 of the detection signal (FIG. 15 (a)) output from the amplification circuit 52a with a threshold value (0V), and a pulse signal (FIG. 15 (b)) corresponding to the comparison result. Is output. The comparison circuit 53b compares the signal level VH2 of the detection signal (FIG. 15C) output from the amplification circuit 52b with a threshold value (0V), and a pulse signal corresponding to the comparison result (FIG. 15D). Is output.

  The initial value determination unit 53 uses the comparison results between the signal levels VH1 and VH2 of the detection signals output from the Hall elements H1 and H2 and the threshold value (0 V), that is, the outputs of the comparison circuits 53a and 53b. An angle range including the initial value θ0 of the relative rotation angle θ is determined. The absolute value of the difference between the initial value θ0 of the relative rotation angle θ and the initial value φ0 of the calculation angle φ that can occur within the determined angle range is less than 90 ° (| θ0−φ0 | <90 °). The initial value φ0 of the calculation angle φ is determined using the initial value table 53d.

  As shown in FIG. 15B, the comparison circuit 53a maintains a high level (H) when the input angle θ is 0 to 180 °, and a low level (when the input angle θ is 180 to 360 °). A pulse signal maintaining L) is output. Further, as shown in FIG. 15D, the comparison circuit 53b maintains a high level (H) when the input angle θ is 90 to 270 ° and is low when the input angle θ is 270 to 90 °. A pulse signal that maintains the level (L) is output.

  Accordingly, when the signal level of the pulse signal output from the comparison circuit 53a is H and the signal level of the pulse signal output from the comparison circuit 53b is L in the initial state where the permanent magnet 2 is not rotating. The initial value θ0 of the relative rotation angle θ of the permanent magnet 2 can be determined to exist in the first quadrant (0 ° ≦ θ <90 °). When the signal level of the pulse signal output from the comparison circuit 53a is H and the signal level of the pulse signal output from the comparison circuit 53b is also H, the initial value of the relative rotation angle θ of the permanent magnet 2 is set. It can be determined that θ0 exists in the second quadrant (90 ° ≦ θ <180 °).

When the signal level of the pulse signal output from the comparison circuit 53a is L and the signal level of the pulse signal output from the comparison circuit 53b is H, the initial value of the relative rotation angle θ of the permanent magnet 2 is set. It can be determined that θ0 exists in the third quadrant (180 ° ≦ θ <270 °). Further, when the signal level of the pulse signal output from the comparison circuit 53a is L and the signal level of the pulse signal output from the comparison circuit 53b is also L, the initial value of the relative rotation angle θ of the permanent magnet 2 is set. It can be determined that θ0 exists in the fourth quadrant (270 ° ≦ θ <360 °).
That is, a quadrant (angle range) where the initial value θ0 of the relative rotation angle θ of the permanent magnet 2 exists can be determined by using a combination of signal levels of the pulse signals output from the comparison circuits 53a and 53b. .

  In this embodiment, the angle range including the initial value θ0 is divided by a value 4 obtained by dividing 0 to 360 ° of the relative rotation angle θ by the phase difference 90 ° between the output signals of the Hall elements H1 and H2. Four angle ranges are set. That is, as shown in FIG. 14, the first quadrant (0 ° ≦ θ0 <90 °), the second quadrant (90 ° ≦ θ0 <180 °), the third quadrant (180 ° ≦ θ0 <270 °), and Four quadrants composed of the fourth quadrant (270 ° ≦ θ0 <360 °) are set as the angle range.

  As shown in FIG. 14, the initial value table 53d includes the signal level VH1 (H or L) of the pulse signal output from the comparison circuit 53a and the signal level VH2 (H of the pulse signal output from the comparison circuit 53b. Or L) and the initial value φ0 of the calculation angle φ are associated with each other. In this embodiment, the combinations H and L of the signal levels VH1 and VH2 and the initial value 45 ° are the combinations H and H of the signal levels VH1 and VH2 and the initial value 135 ° are the combinations L and H of the signal levels VH1 and VH2. The initial value 225 ° is associated with the combinations L and L of the signal levels VH1 and VH2 and the initial value 315 °, respectively, and the combinations of the signal levels VH1 and VH2 are all different in each quadrant.

  Each initial value is a digital angle, which is a count value by an up / down counter 64 constituting an angle calculation unit 60 described later, and the count value is stored in the initial value table 53d as an initial value. The initial value table 53d can be stored in a storage medium such as a ROM or a flash ROM.

  The initial value reading unit 53c (FIG. 12) reads the initial value φ0 associated with the combination of the signal levels VH1 and VH2 of each pulse signal output from the comparison circuits 53a and 53b with reference to the initial value table 53d. put out. For example, when the combination of the signal levels VH1 and VH2 of the pulse signals output from the comparison circuits 53a and 53b is H and L, the initial value reading unit 53c sets 45 ° as the initial value φ0 from the initial value table 53d. Read out.

(Amplifier 51 and angle calculator 60)
The amplifier 51 includes differential amplifier circuits 51a and 51b. The differential amplifier circuit 51a differentially amplifies the output signal sin2θ of the AMR sensor M1, and the differential amplifier circuit 51b differentially amplifies the output signal cos2θ of the AMR sensor M2. The angle calculation unit 60 is a tracking loop type digital angle conversion circuit, and includes a signal creation unit 61, a deviation calculation unit 62, a positive / negative determination unit 63, and an up / down counter (U / D counter) 64.

  The angle calculator 60 uses the signals output from the AMR sensors M1 and M2 to perform feedback control so that the deviation between the relative rotation angle θ with respect to the permanent magnet 2 and the calculated angle φ obtained by the calculation converges to a predetermined value. To calculate the relative rotation angle θ. Further, the angle calculation unit 60 uses the initial value φ0 determined by the initial value determination unit 53 as the initial value φ0 of the calculation angle φ when the calculation of the relative rotation angle θ is started.

  The signal generator 61 uses the signal Asin (2θ + α) output from the differential amplifier circuit 51a and the signal Acos (2θ−α) output from the differential amplifier circuit 51b, and uses the signal 2Asin (2θ−2φ). create. Here, A is the amplitude and α is the phase difference. In this embodiment, the amplitude A = 1 and the phase difference α = 45 °. The deviation calculating unit 62 calculates the deviation (2θ-2φ) using the signal 2Asin (2θ-2φ) output from the signal creating unit 61. The positive / negative determining unit 63 determines whether the deviation (2θ−2φ) calculated by the deviation calculating unit is a positive value or a negative value. The up / down counter 64 adds (counts up) or subtracts (counts down) the count value according to the determination result of the positive / negative determination unit 63.

(Processing content executed by the signal generator 61)
The processing contents executed by the signal creation unit 61 will be described with reference to FIG. Each block denoted by reference numerals 61a to 61k in FIG. 13 indicates the contents of processing executed by the signal generating unit 61, or signals or data generated by the processing.

  The signal creation unit 61 adds the signal Asin (2θ + α) and the signal Acos (2θ−α) to create the signal 2Asin2θcosα (61c). This addition can be performed using a known addition circuit. Further, the signal creation unit 61 creates the signal 2Acos2θsinα by subtracting the signal Acos (2θ−α) from the signal Asin (2θ + α) (61d). This subtraction can be performed using a known subtraction circuit.

  Subsequently, the signal creation unit 61 multiplies the signal Asin2θcosα by the signals cos2φ and (1 / cosα) to create the signal 2Asin2θcos2φ (61c, 61i, 61g). The signal creation unit 61 multiplies the signal 2Acos2θsinα by the signals sin2φ and (1 / sinα) to create the signal 2Acos2θsin2φ (61d, 61j, 61h). Each of these multiplications can be performed using a known multiplication circuit.

  (1 / cos α) and (1 / sin α) are coefficients that do not change. Each φ of cos 2 φ (61 i) and sin 2 φ (61 j) is a variable that varies depending on the count value of the up / down counter 64. Before the permanent magnet 2 starts rotating, that is, before the rotation sensor 1 detects the relative rotation angle θ, the initial value reading unit 53c uses the initial value φ0 read from the initial value table 53d as the calculation angle φ.

  Subsequently, the signal generator 61 subtracts the signal 2Acos2θsin2φ from the signal 2Asin2θcos2φ to generate a sin signal having the signal 2Asin (2θ-2φ), that is, the deviation (2θ-2φ) as a variable (61k). This subtraction can be performed using a known subtraction circuit.

  Next, the deviation calculation unit 62 performs an arc sine operation (arcsine operation) on the signal 2Asin (2θ-2φ) created by the signal creation unit 61 to obtain a deviation (2θ-2φ) (62). Next, the positive / negative determining unit 63 determines whether the deviation (2θ−2φ) obtained by the deviation calculating unit 62 is a positive value or a negative value. The deviation (2θ-2φ) may be determined to be positive when the signal 2Asin (2θ-2φ) is greater than 0, and may be determined to be negative when the signal 2Asin (2θ-2φ) is less than 0. If this method is used, it is not necessary to perform an inverse sine operation on the signal 2Asin (2θ-2φ).

  Next, when the determination result of the positive / negative determination unit 63 is positive, the up / down counter 64 adds 1 to the least significant bit (LSB) of the counter and adds the count value. If the result is negative, 1 is subtracted from the least significant bit of the counter. The count value of the up / down counter 64 is a digital angle, that is, a calculation angle φ (65).

  Further, the signal creation unit 61 creates signals cos2φ and sin2φ using the calculation angle φ (count value) output from the up / down counter 64 (61i, 61j). The generation of these signals uses, for example, a table in which the calculation angle φ (count value) is associated with the data cos2φ and sin2φ, and the data cos2φ and sin2φ associated with the calculation angle φ are read and read. This can be done by converting data into analog signals.

  Then, the signal creating unit 61 again multiplies the signal 2Asin2θcosα by the signals cos2φ and (1 / cosα) to create the signal 2Asin2θcos2φ. Further, the signal 2Acos2θsinα is again multiplied by the signals sin2φ and (1 / sinα) to generate the signal 2Acos2θsin2φ. That is, the deviation (2θ-2φ) is fed back to the signals cos2φ and sin2φ, and the signal 2Asin (2θ-2φ) changes. This feedback is repeated until the deviation (2θ-2φ) converges to zero.

(Output unit 70)
Next, the output unit 70 outputs a signal obtained by converting the calculation angle φ output from the up / down counter 64 into an analog value. Specifically, the output unit 70 latches the calculation angle φ output from the up / down counter 64, converts the latched calculation angle φ when the deviation (2θ−2φ) becomes zero, and converts the calculation angle φ into an analog voltage Vo. An angle signal having a characteristic that the voltage (Vo) rises linearly corresponding to 0 to 360 ° of the angle φ is generated and output (FIG. 15E).

[Problem when initial value φ0 is not determined]
Here, as in this embodiment, instead of using the initial value φ0 of the calculation angle φ determined based on the angle range to which the initial value θ0 of the relative rotation angle θ belongs, the initial value φ0 of the calculation angle φ is not determined. The problem when 0 ° is used as the initial value φ0 will be described with reference to the drawings.

  16 and 17 are explanatory diagrams illustrating a process in which the initial value φ0 of the calculation angle φ follows the initial value θ0 of the relative rotation angle θ. In FIGS. 16 and 17, an arrow indicated by “up” indicates that the up / down counter 64 adds (counts up), and an arrow indicated by “down” indicates that the counter is subtracted (count down). In each figure, it is assumed that the initial value φ0 of the calculation angle φ is 0 °. Moreover, in each figure, in order to make it easy to understand, a deviation is represented not by a count value but by an angle.

(Example 1) When the initial value θ0 of the relative rotation angle θ is 45 ° (FIG. 16A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 45 ° −2 × 0 °) = 90 °. The direction in which the calculated angle φ follows the relative rotation angle θ, that is, whether the calculated angle φ is counted up or counted down is determined by the sign of the deviation (2θ−2φ). The calculation angle φ is counted up when the deviation is positive, and is counted down when the deviation is negative.

  In this example, since the deviation of the initial value (2θ0-2φ0) = 90 °, 2sin (2θ0-2φ0) = 2, so sin (2θ0-2φ0) = 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 90 °. Since the deviation is 90 °> 0, the calculation angle φ is counted up. Therefore, the calculation angle φ is counted up by 1 ° from the initial value 0 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 0 ° → 1 ° →... → 44 ° → 45 °, and the deviation (2θ−2φ) is 90 ° → 88 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 45 °, which is equal to the initial value θ0 = 45 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 2) When the initial value θ0 of the relative rotation angle θ is 89 ° (FIG. 16B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 89 ° −2 × 0 °) = 178 °. Therefore, since 2sin (2θ0-2φ0) ≈0.07, sin (2θ0-2φ0) = 0.035. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 2 °. Since the deviation is 2 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value 0 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 0 ° → 1 ° →... → 88 ° → 89 °, and the deviation (2θ−2φ) increases from 178 ° → 176 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 89 °, which is equal to the initial value θ0 = 89 ° of the relative rotation angle θ, and therefore the calculation angle φ correctly follows the relative rotation angle θ.

(Example 3) When the initial value θ0 of the relative rotation angle θ is 91 ° (FIG. 16C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 91 ° −2 × 0 °) = 182 °. Therefore, since 2sin (2θ0-2φ0) ≈−0.07, sin (2θ0-2φ0) = − 0.035. When this is arcsine calculated, the initial value deviation (2θ0-2φ0) = − 2 °. Since the deviation is −2 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 0 ° (360 °), and the deviation (2θ−2φ) is 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. Subtracted. That is, the calculation angle φ decreases as 0 ° (360 °) → 359 ° →... → 272 ° → 271 °, and the deviation (2θ−2φ) is −178 ° → −176 ° →. It decreases from -2 ° to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 271 °, and the calculation angle φ selects tracking in the direction of 271 ° close to the initial value φ0 = 0 °. For this reason, the deviation of the calculation angle φ from the relative rotation angle θ is always 180 ° (= 271 ° -91 °), and the calculation angle φ does not correctly follow the relative rotation angle θ.

(Example 4) When the initial value θ0 of the relative rotation angle θ is 135 ° (FIG. 17A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 135 ° −2 × 0 °) = 270 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 0 ° (360 °), and the deviation (2θ−2φ) is 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. Subtracted. That is, the calculation angle φ decreases as 0 ° (360 °) → 359 ° →... → 316 ° → 315 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases from -2 ° to 0 °. When the deviation converges to 0 °, the calculation angle φ is 315 °, and the calculation angle φ selects tracking in the direction of 315 ° close to the initial value φ0 = 0 °. For this reason, the deviation of the calculation angle φ from the relative rotation angle θ is always 180 ° (= 315 ° -135 °), and the calculation angle φ does not correctly follow the relative rotation angle θ.

(Example 5) When the initial value θ0 of the relative rotation angle θ is 225 ° (FIG. 17B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 225 ° −2 × 0 °) = 450 °. Therefore, since 2sin (2θ0-2φ0) = 2, sin (2θ0-2φ0) = 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 90 °. Since the deviation is 90 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value 0 ° (360 °), and the deviation (2θ−2φ) is 2φ (= 2 × 1 ° every time the calculation angle φ is counted up by 1 °. ) Is subtracted. That is, the calculation angle φ increases from 0 ° → 1 ° →... → 44 ° → 45 °, and the deviation (2θ−2φ) is 90 ° → 88 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ is 45 °, and the calculation angle φ is selected to follow in the direction of 45 ° close to the initial value φ0 = 0 °. For this reason, there is always 180 ° (= 225 ° −45 °) deviation of the calculation angle φ from the relative rotation angle θ, and the calculation angle φ does not follow the relative rotation angle θ correctly.

(Example 6) When the initial value θ0 of the relative rotation angle θ is 315 ° (FIG. 17C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 315 ° −2 × 0 °) = 630 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 0 ° (360 °), and the deviation (2θ−2φ) is 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. Subtracted. That is, the calculation angle φ increases from 0 ° (360 °) → 359 ° →... → 316 ° → 315 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases from -2 ° to 0 °. When the deviation converges to 0 °, the calculation angle φ is 315 °, which is equal to the initial value θ0 = 315 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

  As described above, when the initial value φ0 of the calculation angle φ is 0 °, the calculation angle φ may not correctly follow the relative rotation angle θ depending on the value of the initial value θ0 of the relative rotation angle θ. This is because the solution angle X of the calculation angle φ becomes two (X, X + 180 °) in order to obtain the calculation angle φ of the single value from the deviation (2θ−2φ) of the double value. That is, if 0 ° is used as the initial value φ0 without determining the initial value φ0 of the calculation angle φ, the relative rotation angle θ may not be detected accurately.

[Effect of determining initial value φ0]
Here, the effect of determining the initial value φ0 as in this embodiment will be described with reference to the drawings. 18 to 21 are explanatory diagrams illustrating a process in which the initial value φ0 of the calculation angle φ follows the initial value θ0 of the relative rotation angle θ. As described above, the initial value φ0 is determined based on the angle range to which the initial value θ0 of the relative rotation angle θ belongs.

(When φ0 = 45 °)
(Example 1) When the initial value θ0 of the relative rotation angle θ is 0 ° (FIG. 18A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 0 ° −2 × 45 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Therefore, the calculation angle φ is counted down by 1 ° from the initial value of 45 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases from 45 ° → 44 ° →... → 1 ° → 0 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ also becomes 0 °, which is equal to the initial value θ0 = 0 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 2) When the initial value θ0 of the relative rotation angle θ is 30 ° (FIG. 18B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 30 ° −2 × 45 °) = − 30 °. Therefore, since 2sin (2θ0-2φ0) = − 1, sin (2θ0-2φ0) = − 0.5. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 30 °. Since the deviation is −30 ° <0, the calculation angle φ is counted down.

  Therefore, the calculation angle φ is counted down by 1 ° from the initial value of 45 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases from 45 ° → 44 ° →... → 31 ° → 30 °, and the deviation (2θ−2φ) is −30 ° → −28 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 30 °, which is equal to the initial value θ0 = 30 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 3) When the initial value θ0 of the relative rotation angle θ is 80 ° (FIG. 18C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 80 ° −2 × 45 °) = 70 °. Therefore, since 2sin (2θ0-2φ0) ≈1.88, sin (2θ0-2φ0) ≈0.94. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 70 °. Since the deviation is 70 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value of 45 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 45 ° → 46 ° →... → 79 ° → 80 °, and the deviation (2θ−2φ) increases from 70 ° → 68 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 80 °, which is equal to the initial value θ0 = 80 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(When φ0 = 135 °)
(Example 1) When the initial value θ0 of the relative rotation angle θ is 90 ° (FIG. 19A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 90 ° −2 × 135 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Therefore, the calculation angle φ is counted down by 1 ° from the initial value of 135 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases as 135 ° → 134 ° →... → 89 ° → 90 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 90 °, which is equal to the initial value θ0 = 90 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 2) When the initial value θ0 of the relative rotation angle θ is 150 ° (FIG. 19B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 150 ° −2 × 135 °) = 30 °. Therefore, since 2sin (2θ0-2φ0) = 1, sin (2θ0-2φ0) = 0.5. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 30 °. Since the deviation is 30 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is incremented by 1 ° from the initial value of 135 °, and the deviation (2θ−2φ) is incremented by 2φ (= 2 × 1 °) every time the calculation angle φ is incremented by 1 °. The That is, the calculation angle φ increases from 135 ° → 136 ° →... → 149 ° → 150 °, and the deviation (2θ−2φ) is 30 ° → 28 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 150 °, which is equal to the initial value θ0 = 150 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 3) When the initial value θ0 of the relative rotation angle θ is 180 ° (FIG. 19C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 180 ° −2 × 135 °) = 90 °. Therefore, since 2sin (2θ0-2φ0) = 2, sin (2θ0-2φ0) = 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 90 °. Since the deviation is 90 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value of 135 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 135 ° → 136 ° → ・ ・ ・ → 179 ° → 180 °, and the deviation (2θ−2φ) is 90 ° → 88 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 180 °, which is equal to the initial value θ0 = 180 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(When φ0 = 225 °)
(Example 1) When the initial value θ0 of the relative rotation angle θ is 180 ° (FIG. 20A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 180 ° −2 × 225 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 225 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ increases from 225 ° → 224 ° →... → 181 ° → 180 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 180 °, which is equal to the initial value θ0 = 180 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 2) When the initial value θ0 of the relative rotation angle θ is 240 ° (FIG. 20B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 240 ° −2 × 225 °) = 30 °. Therefore, since 2sin (2θ0-2φ0) = 1, sin (2θ0-2φ0) = 0.5. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 30 °. Since the deviation is 30 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value 225 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 225 ° → 226 ° →... → 239 ° → 240 °, and the deviation (2θ−2φ) is 30 ° → 28 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ is 240 °, which is equal to the initial value θ0 = 240 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 3) When the initial value θ0 of the relative rotation angle θ is 270 ° (FIG. 20C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 270 ° −2 × 225 °) = 90 °. Therefore, since 2sin (2θ0-2φ0) = 2, sin (2θ0-2φ0) = 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 90 °. Since the deviation is 90 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value 225 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. The That is, the calculation angle φ increases from 225 ° → 226 ° →... → 269 ° → 270 °, and the deviation (2θ−2φ) is 90 ° → 88 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 270 °, which is equal to the initial value θ0 = 270 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(When φ0 = 315 °)
(Example 1) When the initial value θ0 of the relative rotation angle θ is 270 ° (FIG. 21A)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 270 ° −2 × 315 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Therefore, the calculation angle φ is counted down by 1 ° from the initial value 315 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ increases from 315 ° → 314 ° →... → 271 ° → 270 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 270 °, which is equal to the initial value θ0 = 270 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 2) When the initial value θ0 of the relative rotation angle θ is 330 ° (FIG. 21B)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 330 ° −2 × 315 °) = 30 °. Therefore, since 2sin (2θ0-2φ0) = 1, sin (2θ0-2φ0) = 0.5. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 30 °. Since the deviation is 30 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is incremented by 1 ° from the initial value of 315 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is incremented by 1 °. The That is, the calculation angle φ increases from 315 ° → 316 ° →... → 329 ° → 330 °, and the deviation (2θ−2φ) is 30 ° → 28 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ is 240 °, which is equal to the initial value θ0 = 330 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(Example 3) When the initial value θ0 of the relative rotation angle θ is 360 ° (FIG. 21C)
The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 360 ° −2 × 315 °) = 90 °. Therefore, since 2sin (2θ0-2φ0) = 2, sin (2θ0-2φ0) = 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = 90 °. Since the deviation is 90 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is incremented by 1 ° from the initial value of 315 °, and the deviation (2θ−2φ) is subtracted by 2φ (= 2 × 1 °) every time the calculation angle φ is incremented by 1 °. The That is, the calculation angle φ increases from 315 ° → 316 ° →... → 359 ° → 360 °, and the deviation (2θ−2φ) is 90 ° → 88 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ is 270 °, which is equal to the initial value θ0 = 360 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

[Effect of the first embodiment]
(1) As described above, the rotation sensor 1 of this embodiment satisfies the condition of | θ0−φ0 | <90 ° between the initial value θ0 of the relative rotation angle θ and the initial value φ0 of the calculation angle φ. If satisfied, the calculation angle φ can accurately follow the relative rotation angle θ. In addition, the initial value φ0 of the calculation angle φ becomes equal to the initial value θ0 of the relative rotation angle θ before the permanent magnet 2 relatively rotates, and thereafter the calculation angle φ is relatively rotated while the permanent magnet 2 is rotating. It is possible to accurately follow the angle θ.

Therefore, it is sufficient to determine the quadrant where the initial value θ0 of the relative rotation angle θ exists before the permanent magnet 2 relatively rotates, and it is not necessary to determine the quadrant during the relative rotation.
That is, the calculation time of the relative rotation angle θ when the permanent magnet 2 is relatively rotating can be shortened, and the processing load and power consumption of the detection circuit 50 can be reduced.

(2) Since each output signal of the Hall elements H1 and H2 is used only for determining an angle range in which the initial value θ0 of the relative rotation angle θ exists, the Hall elements H1 and H2 and the AMR sensors M1 and M2 There is no need to match each output phase. Therefore, it is not necessary to increase the accuracy of the relative arrangement positions of the Hall elements H1 and H2 and the AMR sensors M1 and M2, so that the manufacturing yield of the rotation sensor 1 can be increased.

(3) Since the Hall elements H1 and H2 are arranged so that a phase difference of 90 ° is generated between the detection signals, the signal level of each Hall element is changed every time the relative rotation angle θ changes by 90 °. The combination of can be changed.
Therefore, the initial value determination unit 53 can determine the angle range where the initial value θ0 of the relative rotation angle θ exists in units of 90 ° by using the combination of the signal levels of the detection signals.

(4) Since the AMR sensors M1 and M2 are arranged such that a phase difference of 45 ° is generated between the signals, a sine wave signal (sin signal) is output from one of the AMR sensors, and the sine wave A cosine wave signal (cos signal) whose phase is delayed by 45 ° from the signal can be output from the other.
Therefore, the relative rotation angle can be calculated using the above sine wave signal and cosine wave signal.

(5) Since the combinations of the comparison results between the signal levels of the detection signals of the Hall elements H1 and H2 and the threshold values are all different in the first to fourth quadrants, the quadrant determination accuracy by the initial value determination unit 53 Can be increased.

(6) Since the relative rotation angle θ is calculated by performing feedback control using the sin 2θ signal and the cos 2θ signal output from the AMR sensors M1, M2 so that the deviation (2θ-2φ) becomes 0, the relative rotation angle The angle θ can be calculated with high accuracy.

(7) Further, by using the sin2θ signal and the cos2θ signal output from the AMR sensors M1 and M2, the sin2φ signal and the cos2φ signal, a circuit for multiplying the signal, and a circuit for subtracting the signal, the deviation (2θ− 2φ) can be calculated.

(8) Furthermore, since the Hall elements H1 and H2 and the AMR sensors M1 and M2 can be built in the silicon substrate 10, the rotation sensor 1 can be downsized.

[Degree of margin of initial value θ0 corresponding to initial value φ0]
Next, the margin of the initial value θ0 of the relative rotation angle θ corresponding to the initial value φ0 of the calculation angle φ will be described with reference to the drawings. FIG. 22 is an explanatory diagram showing the margin of the initial value θ0 corresponding to the initial value φ0.

  If the condition of | θ0−φ0 | <90 ° is satisfied between the initial value φ0 of the calculation angle φ and the initial value θ0 of the relative rotation angle θ, the calculation angle φ accurately determines the relative rotation angle θ. Can follow. In FIG. 22, the range indicated by “positive” indicates the range of the initial value θ0 in which the initial value φ0 can accurately follow the initial value θ0, and the range indicated by “false” indicates that the initial value φ0 accurately follows. The range of the initial value θ0 that cannot be performed is shown. As shown in the figure, the range of the initial value θ0 that each initial value φ0 can follow is wider than the corresponding quadrant (see FIG. 14) of each initial value φ0. Hereinafter, the range of the initial value θ0 that can be followed by each initial value φ0 will be described for each initial value φ0.

(Initial value φ0 = 45 °)
When the initial value φ0 is 45 °, the range of the initial value θ0 that can be followed is (45 ° −90 °) <θ0 <(45 ° + 90 °), that is, −45 ° <θ0 <135 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <135 ° and 315 ° <θ0 ≦ 360 °.

  For example, assume that 45 ° is selected as the initial value φ0 when the initial value θ0 is 134 °. The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 134 ° −2 × 45 °) = 178 °. Therefore, since 2sin (2θ0-2φ0) ≈0.07, sin (2θ0-2φ0) ≈0.035. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) ≈2 °. Since the deviation is 2 °> 0, the calculation angle φ is counted up.

  Therefore, the calculation angle φ is counted up by 1 ° from the initial value 45 °, and the deviation (2θ−2φ) decreases by 2φ (= 2 × 1 °) every time the calculation angle φ is counted up by 1 °. . That is, the calculation angle φ increases from 45 ° → 46 ° →... → 133 ° → 134 °, and the deviation (2θ−2φ) increases from 178 ° → 176 ° →... → 2 ° → 0 °. Decrease. When the deviation converges to 0 °, the calculation angle φ becomes 134 °, which is equal to the initial value θ0 = 134 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

  As described above, even if the initial value θ0 of the relative rotation angle θ is outside the first quadrant range (0 ° ≦ θ0 <90 °) corresponding to the initial value φ0 = 45 ° of the calculation angle φ, | θ0 If the condition of −φ0 | <90 ° is satisfied, the calculation angle φ can accurately follow the relative rotation angle θ.

(In case of initial value φ0 = 135 °)
When the initial value φ0 is 135 °, the range of the initial value θ0 that can be followed is (135 ° −90 °) <θ0 <(135 ° + 90 °), that is, 45 ° <θ0 <225 °.

  For example, it is assumed that 135 ° is selected as the initial value φ0 when the initial value θ0 is 90 °. The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 90 ° −2 × 135 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Therefore, the calculation angle φ is counted down by 1 ° from the initial value of 135 °, and the deviation (2θ−2φ) decreases by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases as 135 ° → 134 ° →... → 91 ° → 90 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 90 °, which is equal to the initial value θ0 = 90 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(In case of initial value φ0 = 225 °)
When the initial value φ0 is 225 °, the range of the initial value θ0 that can be followed is (225 ° −90 °) <θ0 <(225 ° + 90 °), that is, 135 ° <θ0 <315 °.

  For example, assume that 225 ° is selected as the initial value φ0 when the initial value θ0 is 150 °. The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 150 ° −2 × 225 °) = − 150 °. Therefore, since 2sin (2θ0-2φ0) = − 1, sin (2θ0-2φ0) = − 0.5. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 150 °. Since the deviation is −150 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 225 °, and the deviation (2θ−2φ) decreases by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases from 225 ° → 224 ° →... → 151 ° → 150 °, and the deviation (2θ−2φ) is −150 ° → −148 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 150 °, which is equal to the initial value θ0 = 150 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

(In case of initial value φ0 = 315 °)
When the initial value φ0 is 315 °, the range of the initial value θ0 that can be followed is (315 ° −90 °) <θ0 <(315 ° + 90 °), that is, 225 ° <θ0 <405 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <45 ° and 225 ° <θ0 ≦ 360 °.

  For example, assume that 315 ° is selected as the initial value φ0 when the initial value θ0 is 270 °. The initial value of the deviation (2θ-2φ) is (2θ0-2φ0) = (2 × 270 ° −2 × 315 °) = − 90 °. Therefore, since 2sin (2θ0-2φ0) = − 2, sin (2θ0-2φ0) = − 1. When arcsine calculation is performed on this, the initial value deviation (2θ0-2φ0) = − 90 °. Since the deviation is −90 ° <0, the calculation angle φ is counted down.

  Accordingly, the calculation angle φ is counted down by 1 ° from the initial value 315 °, and the deviation (2θ−2φ) decreases by 2φ (= 2 × 1 °) every time the calculation angle φ is counted down by 1 °. That is, the calculation angle φ decreases as 315 ° → 314 ° →... → 271 ° → 270 °, and the deviation (2θ−2φ) is −90 ° → −88 ° →. It decreases to 0 °. When the deviation converges to 0 °, the calculation angle φ becomes 270 °, which is equal to the initial value θ0 = 270 ° of the relative rotation angle θ, so that the calculation angle φ correctly follows the relative rotation angle θ.

  As described above, the rotation sensor 1 of this embodiment satisfies the condition | θ0−φ0 | <90 ° between the initial value θ0 of the relative rotation angle θ and the initial value φ0 of the calculation angle φ. Even if the initial value θ0 selected from the initial value table 53d is outside the range of the quadrant corresponding to the initial value φ0, the calculation angle φ can accurately follow the relative rotation angle θ.

For example, when 45 ° is selected as the initial value φ0, the initial value θ0 is actually 45 °, but even if it is erroneously detected as 90 °, the initial value φ0 = 45 ° is equal to the initial value θ0 = 90 °. Since it is covered, the calculation angle φ can accurately follow the relative rotation angle θ.
On the other hand, in FIG. 22, the initial value θ0 = 90 ° corresponds to both the initial value φ0 = 45 ° and 135 °. Therefore, when the initial value θ0 is 90 °, the initial value φ0 is 135. Even if 45 ° is selected by mistake when selecting °, the calculation angle φ can accurately follow the relative rotation angle θ.

  Therefore, for example, even when the initial value θ0 or the initial value φ0 changes due to the influence of external noise or an external magnetic field, the calculation angle φ can accurately follow the relative rotation angle θ. That is, it is possible to realize a rotation sensor in which the detection accuracy is unlikely to deteriorate even under the influence of external noise or an external magnetic field.

[Example of change]
The initial value φ0 of the calculation angle φ can be determined so as to be determined before the permanent magnet 2 relatively rotates and at a predetermined time after the relative rotation starts. If this configuration is used, even if the follow-up route of the calculation angle φ with respect to the relative rotation angle θ deviates during the relative rotation of the permanent magnet 2 and an error occurs in the calculation angle φ, the initial value φ0 is newly set. Since the relative rotation angle θ can be calculated using the determined initial value φ0, the tracking route can be returned to the original accurate tracking route, so that the error of the calculation angle φ can be corrected.

Second Embodiment
Next explained is the second embodiment of the invention. FIG. 23 is an explanatory view schematically showing the structure of a sensor chip provided in the rotation sensor of this embodiment. FIG. 24 is an explanatory diagram showing output signals of the AMR sensors M1 and M2 and the Hall elements H1 and H2. FIG. 25 is an explanatory diagram showing the configuration of the initial value table 54b. FIG. 26 is an explanatory diagram showing a margin of the initial value θ0 corresponding to the initial value φ0.

  The rotation sensor of this embodiment is characterized in that the phase difference between the output signals of the Hall elements H1 and H2 is 45 °. As shown in FIG. 23, the Hall elements H1 and H2 are arranged such that each magnetic detection surface (Hall plate surface) forms 45 °. For this reason, as shown in FIG. 24, there is a 45 ° phase difference between the output signals of the Hall elements H1 and H2, and in this example, the output signal of the Hall element H1 changes from a low level (L) to a high level ( The output signal of the Hall element H2 changes from the low level (L) to the high level (H) with a delay of 45 ° from the timing when the change to H).

  When the 0 to 360 ° of the initial value θ0 of the relative rotation angle θ is divided based on the level of the output signal of the Hall elements H1 and H2 shown in FIG. 24, the total of the first to fourth ranges is obtained as shown in FIG. It can be divided into four. As shown, the first range is 0 ° ≦ θ0 <45 °, the second range is 45 ° ≦ θ0 <180 °, the third range is 180 ° ≦ θ0 <225 °, The fourth range is 225 ° ≦ θ0 <360 °. In addition, the initial value φ0 of the calculation angle φ is set to the median value of each range in the initial value table 53d, 22.5 ° with respect to the first range, 120 ° with respect to the second range, 202.5 ° is set for the third range, and 300 ° is set for the fourth range.

  For example, if the signal level VH1 of the Hall element H1 is high (H) and the signal level VH2 of the Hall element H2 is low (L), the initial value φ0 is 22.5 ° from the initial value table 53d. Is selected.

Further, as shown in FIG. 26, the range of the initial value θ0 that can be accurately followed by the initial value φ0 exceeds the first to fourth ranges, and there is a margin. Compared with the first embodiment, the initial value φ0 is different, but the range of the initial value θ0 that can be accurately followed by the initial value φ0 is the same as the first embodiment under the condition of | θ0−φ0 | <90 °. You can decide to meet.
The range of the initial value θ0 that the initial value φ0 = 22.5 ° can follow is (22.5 ° −90 °) <θ0 <(22.5 ° + 90 °), that is, −67.5 ° <θ0 <. 112.5 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <112.5 ° and 292.5 ° <θ0 ≦ 360 °.

  The range of the initial value θ0 that the initial value φ0 = 120 ° can follow is (120 ° −90 °) <θ0 <(120 ° + 90 °), that is, 30 ° <θ0 <210 °. The range of the initial value θ0 that can be followed by the initial value φ0 = 202.5 ° is (202.5 ° −90 °) <θ0 <(202.5 ° + 90 °), that is, 112.5 ° <θ0. <292.5 °. Further, the range of the initial value θ0 that the initial value φ0 = 300 ° can follow is (300 ° −90 °) <θ0 <(300 ° + 90 °), that is, 210 ° <θ0 <390 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <30 ° and 210 ° <θ0 ≦ 360 °.

  As described above, the rotation sensor of the second embodiment has the same configuration as the rotation sensor of the first embodiment except that a 45 ° phase difference exists between the output signals of the Hall elements H1 and H2. The same effect as the rotation sensor of one embodiment can be produced.

<Third Embodiment>
Next explained is the third embodiment of the invention. FIG. 27 is an explanatory view schematically showing the structure of a sensor chip provided in the rotation sensor of this embodiment. FIG. 28 is an explanatory diagram showing output signals of the AMR sensors M1 and M2 and the Hall elements H1 to H3. FIG. 29 is an explanatory diagram showing the configuration of the initial value table 54b. FIG. 30 is an explanatory diagram showing a margin of the initial value θ0 corresponding to the initial value φ0.

  The rotation sensor of this embodiment includes three Hall elements H1, H2, and H3, and is characterized in that the phase difference between the output signals of the Hall elements is 60 °. As shown in FIG. 27, the Hall elements H1, H2, and H3 are arranged such that each magnetic detection surface (Hall plate surface) forms 60 °. For this reason, as shown in FIG. 28, there is a phase difference of 60 ° between the output signals of the Hall elements H1, H2, and H3. In this example, the output signal of the Hall element H1 changes from low level (L) to high. The output signal of the Hall element H2 changes from the low level (L) to the high level (H) with a delay of 60 ° from the timing of changing to the level (H), and further, the output signal of the Hall element H2 changes from the low level (L). The output signal of the Hall element H3 changes from the low level (L) to the high level (H) with a delay of 60 ° from the timing when the level changes to the high level (H).

  When the 0 to 360 ° of the initial value θ0 of the relative rotation angle θ is divided based on the level of the output signal of the Hall elements H1, H2, and H3 shown in FIG. 28, the first 60 ° unit first is obtained as shown in FIG. Or it can be divided into six in the sixth range. The initial value φ0 of the calculation angle φ is set to the median value of each range in the initial value table 53d.

  For example, the signal level VH1 of the Hall element H1 is high level (H), the signal level VH2 of the Hall element H2 is low level (L), and the signal level VH3 of the Hall element H3 is low level (L). In this case, 30 ° is selected as the initial value φ0 from the initial value table 53d.

Further, as shown in FIG. 30, the range of the initial value θ0 that can be accurately followed by the initial value φ0 exceeds the first to sixth ranges, and there is a margin. Compared with the first embodiment, the initial value φ0 is different, but the range of the initial value θ0 that can be accurately followed by the initial value φ0 is the same as the first embodiment under the condition of | θ0−φ0 | <90 °. You can decide to meet.
The range of the initial value θ0 that can follow the initial value φ0 = 30 ° is (30 ° −90 °) <θ0 <(30 ° + 90 °), that is, −60 ° <θ0 <120 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <120 ° and 300 ° <θ0 ≦ 360 °.

  The range of the initial value θ0 that the initial value φ0 = 90 ° can follow is (90 ° −90 °) <θ0 <(90 ° + 90 °), that is, 0 ° <θ0 <180 °. The range of the initial value θ0 that the initial value φ0 = 150 ° can follow is (150 ° −90 °) <θ0 <(150 ° + 90 °), that is, 60 ° <θ0 <240 °. The range of the initial value θ0 that the initial value φ0 = 210 ° can follow is (210 ° −90 °) <θ0 <(210 ° + 90 °), that is, 120 ° <θ0 <300 °. The range of the initial value θ0 that can be followed by the initial value φ0 = 270 ° is (270 ° −90 °) <θ0 <(270 ° + 90 °), that is, 180 ° <θ0 <360 °. Further, the range of the initial value θ0 that the initial value φ0 = 330 ° can follow is (330 ° −90 °) <θ0 <(330 ° + 90 °), that is, 240 ° <θ0 <420 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <60 ° and <θ0 ≦ 360 °.

  As described above, the rotation sensor of the third embodiment has the same configuration as the rotation sensor of the first embodiment except that a phase difference of 60 ° exists between the output signals of the Hall elements H1, H2, and H3. The same effect as the rotation sensor of the first embodiment can be obtained. Moreover, the number of angle ranges in which the initial value θ0 exists is six in the first to sixth ranges, which is two more than four in the first embodiment, and accordingly, the angle range in which the initial value θ0 exists. Can be narrowed down to a narrow range, the time required for the calculation angle φ to follow the initial value θ0 and converge to the initial value θ0 can be shortened.

<Fourth embodiment>
Next explained is the fourth embodiment of the invention. FIG. 31 is an explanatory diagram showing the configuration of the initial value table 54b. FIG. 32 is an explanatory diagram showing a margin of the initial value θ0 corresponding to the initial value φ0.

  The rotation sensor of this embodiment includes four Hall elements H1 to H4, and is characterized in that the phase difference between output signals of the Hall elements is 45 °. The Hall elements H1 to H4 are arranged such that each magnetic detection surface (hole plate surface) forms 45 °. Therefore, there is a 45 ° phase difference between the output signals of the Hall elements H1 to H4. In this example, the timing at which the output signal changes from the low level (L) to the high level (H) is the Hall element H1. From H4 to H4 in order.

  When 0 to 360 ° of the initial value θ0 of the relative rotation angle θ is divided based on the level of the output signal of the Hall elements H1 to H4, as shown in FIG. 31, the first to eighth ranges in units of 45 ° are calculated. It can be divided into eight. The initial value φ0 of the calculation angle φ is set to the median value of each range in the initial value table 53d.

  For example, if each of the signal levels VH1 to VH4 of the Hall elements H1 to H4 is high level (H), low level (L), low level (L), and low level (L) in order, the initial value table 53d. Therefore, 22.5 ° is selected as the initial value φ0.

Also, as shown in FIG. 32, the range of the initial value θ0 that the initial value φ0 can accurately follow exceeds the first to eighth ranges, and there is a margin. Compared with the first embodiment, the initial value φ0 is different, but the range of the initial value θ0 that can be accurately followed by the initial value φ0 is the same as the first embodiment under the condition of | θ0−φ0 | <90 °. You can decide to meet.
The range of the initial value θ0 that the initial value φ0 = 22.5 ° can follow is (22.5 ° −90 °) <θ0 <(22.5 ° + 90 °), that is, −67.5 ° <θ0 <. 112.5 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <112.5 ° and 292.5 ° <θ0 ≦ 360 °.

  The range of the initial value θ0 that can be followed by the initial value φ0 = 67.5 ° is (67.5 ° −90 °) <θ0 <(67.5 ° + 90 °), that is, −22.5 ° < θ0 <157.5 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <157.5 ° and 337.5 ° <θ0 ≦ 360 °. The range of the initial value θ0 that can be followed by the initial value φ0 = 112.5 ° is (112.5 ° −90 °) <θ0 <(112.5 ° + 90 °), that is, 22.5 ° <θ0. <202.5 °. Further, the range of the initial value θ0 that the initial value φ0 = 157.5 ° can follow is (157.5 ° −90 °) <θ0 <(157.5 ° + 90 °), that is, 67.5 ° <θ0. <247.5 °.

  The range of the initial value θ0 that can be followed by the initial value φ0 = 202.5 ° is (202.5 ° −90 °) <θ0 <(202.5 ° + 90 °), that is, 112.5 ° <θ0. <292.5 °. The range of the initial value θ0 that can be followed by the initial value φ0 = 247.5 ° is (247.5 ° −90 °) <θ0 <(247.5 ° + 90 °), that is, 157.5 ° <θ0. <337.5 °. The range of the initial value θ0 that can be followed by the initial value φ0 = 292.5 ° is (292.5 ° −90 °) <θ0 <(292.5 ° + 90 °), that is, 202.5 ° <θ0. <382.5 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <22.5 ° and 202.5 ° <θ0 ≦ 360 °. Further, the range of the initial value θ0 that the initial value φ0 = 337.5 ° can follow is (337.5 ° −90 °) <θ0 <(337.5 ° + 90 °), that is, 247.5 ° <θ0. <427.5 °. When this is based on φ0 = 0 ° (360 °), 0 ° ≦ θ0 <67.5 ° and 247.5 ° <θ0 ≦ 360 °.

  As described above, the rotation sensor of the fourth embodiment has the same configuration as the rotation sensor of the first embodiment, except that a 45 ° phase difference exists between the output signals of the Hall elements H1 to H4. The same effect as the rotation sensor of the first embodiment can be obtained. In addition, the number of angle ranges in which the initial value θ0 exists is eight in the first to eighth ranges, which is two more than six in the third embodiment, and accordingly, the angle range in which the initial value θ0 exists. Can be narrowed down to a narrower range, the time required for the calculation angle φ to follow the initial value θ0 and converge to the initial value θ0 can be further reduced.

<Other embodiments>
(1) The phase difference between the output signals of the Hall elements H1 and H2 may be other than 90 °, 60 ° and 45 °. That is, if the phase difference between the output signals of the Hall elements H1 and H2 exceeds 0 °, the calculation angle φ can accurately follow the relative rotation angle θ. Therefore, since the selection range of the angle formed between the magnetic detection surfaces of the Hall element is very wide, the degree of freedom regarding the arrangement position of the Hall element can be increased.

(2) An element other than the Hall element can be used as long as the signal level of the output signal changes by one period while the permanent magnet 2 makes one rotation, that is, while the relative rotation angle θ changes by one period. . For example, a magnetic detection element such as a GMR (Giant Magneto-Resistive effect) type element or a TMR (Tunnel Magneto-Resistive effect) type element or a switch such as a coil can be used.

(3) A sensor other than the AMR sensor can be used as long as the signal level of the output signal changes for two cycles while the permanent magnet 2 rotates once, that is, while the relative rotation angle θ changes for one cycle. .

(4) A correction unit that corrects the amplitude difference, offset, and initial phase error between the sin 2θ signal and the cos 2θ signal output from the amplification unit 51 may be provided. If the rotation sensor provided with this configuration is used, the detection accuracy of the relative rotation angle θ can be further enhanced.

(5) A correction unit that corrects the amplitude difference, offset, and initial phase error between the sin θ signal and the cos θ signal output from the amplifying unit 52 may be provided. If the rotation sensor provided with this configuration is used, the detection accuracy of the relative rotation angle θ can be further enhanced.

(6) The angle calculation unit 60, the initial value determination unit 53, and the output unit 70 can be realized by hardware such as a discrete circuit, or can be realized by software using a microcomputer.

(7) The number of pulse outputs of the Hall element H1 or H2 can be counted to detect multiple rotations (360 ° or more).
(8) The content performed by the initial value determination unit 53 may be performed by the output unit 70.

(9) The output unit 70 may be configured to output a digital value without converting the calculation angle into an analog signal.
(9) The detection circuit 50 can be formed on the silicon substrate 10 and the sensor chip 5 and the detection circuit 50 can be integrated.

(10) Instead of the silicon substrate 10, a substrate formed of a compound semiconductor such as GaAs, InAs, or InSb can also be used.
(11) A member coated with magnetic ink can be used instead of the permanent magnet. A member magnetized on the surface of the conductive member can also be used.
(12) A horizontal Hall element can be used instead of the vertical Hall element. Further, it is also possible to use the horizontal Hall element so as to overlap the magnetoresistive element region so that the magnetic detection portion is perpendicular to the magnetoresistive element.

  Application of the rotation sensor 1 according to the present invention is not limited as long as the detection target rotates relatively. For example, a crank angle sensor that detects a crank angle of a crankshaft provided in an internal combustion engine, a cam angle sensor that detects a cam angle of a camshaft, a steering angle sensor that detects a steering angle of a steering device provided in a vehicle, etc. Can be applied. The present invention can also be applied to a sensor that detects the angle of a joint provided in a robot.

1 .... rotation sensor, 2 .... permanent magnet (magnetic generator), 5 .... sensor chip,
10 .. Silicon substrate, H1, H2, .. Hall element, M1, M2, .. AMR sensor,
R1 to R8 .. magnetoresistive element, .theta .. relative rotation angle, .theta.0.
φ ·· Calculation angle, φ0 ·· Initial value.

Claims (13)

Each of the signals is arranged in a magnetic field of a relatively rotating magnetism generating unit, and outputs a signal whose signal level changes in two cycles according to the strength of the magnetism field while the magnetism generating unit makes one rotation. A plurality of magnetoelectric transducers arranged so that a phase difference appears between them,
In a rotation sensor configured to obtain a relative rotation angle with respect to the magnetism generation unit using a signal output from each magnetoelectric conversion element,
A plurality of detections are arranged so that a detection signal whose signal level changes in one cycle according to the intensity of the magnetic field is output and a phase difference is generated between the detection signals during one rotation of the magnetism generator. Elements,
Using the signal output from each magnetoelectric conversion element, the relative rotation angle is calculated by performing feedback control so that the deviation between the relative rotation angle with respect to the magnetism generating unit and the calculation angle obtained by calculation converges to a predetermined value. An angle calculator,
Using each comparison result between each signal level and each threshold value of each detection signal output from each detection element, an angle range including the initial value of the relative rotation angle is determined, and the detection can occur within the determined angle range. An initial value determination unit that determines the initial value of the calculation angle so that the absolute value of the difference between the initial value of the relative rotation angle and the initial value of the calculation angle is less than 90 °;
An output unit that outputs a signal corresponding to a calculation angle calculated by the angle calculation unit in one cycle while the magnetism generation unit makes one rotation;
The initial value determination unit determines an initial value of the calculation angle only before the magnetism generation unit starts relative rotation,
The angle calculation unit starts the feedback control using the initial value of the calculation angle determined by the initial value determination unit, and calculates the relative rotation angle. Each of the signals is arranged in a magnetic field of a relatively rotating magnetism generating unit, and outputs a signal whose signal level changes in two cycles according to the strength of the magnetism field while the magnetism generating unit makes one rotation. A plurality of magnetoelectric transducers arranged so that a phase difference appears between them,
In a rotation sensor configured to obtain a relative rotation angle with respect to the magnetism generation unit using a signal output from each magnetoelectric conversion element,
A plurality of detections are arranged so that a detection signal whose signal level changes in one cycle according to the intensity of the magnetic field is output and a phase difference is generated between the detection signals during one rotation of the magnetism generator. Elements,
Using the signal output from each magnetoelectric conversion element, the relative rotation angle is calculated by performing feedback control so that the deviation between the relative rotation angle with respect to the magnetism generating unit and the calculation angle obtained by calculation converges to a predetermined value. An angle calculator,
Using each comparison result between each signal level and each threshold value of each detection signal output from each detection element, an angle range including the initial value of the relative rotation angle is determined, and the detection can occur within the determined angle range. An initial value determination unit that determines the initial value of the calculation angle so that the absolute value of the difference between the initial value of the relative rotation angle and the initial value of the calculation angle is less than 90 °;
An output unit that outputs a signal corresponding to a calculation angle calculated by the angle calculation unit in one cycle while the magnetism generation unit makes one rotation;
The initial value determining unit determines an initial value of the calculation angle before the magnetism generating unit starts relative rotation and at a predetermined time after the relative rotation starts,
The angle calculation unit performs the feedback control using the initial value of the calculation angle determined by the initial value determination unit, and calculates the relative rotation angle.   When the initial value of the relative rotation angle is θ0 and the initial value of the calculation angle is φ0, the angle calculation unit has a relative rotation existing in the range of (φ0−90 °) <θ0 <(φ0 + 90 °). The rotation sensor according to claim 1, wherein an initial value θ0 of the angle can be calculated.   The rotation sensor according to any one of claims 1 to 3, wherein the plurality of detection elements are arranged such that a phase difference of 90 ° is generated between the detection signals.   The rotation sensor according to any one of claims 1 to 4, wherein the plurality of magnetoelectric transducers are arranged so that a phase difference of 45 ° is generated between the signals.   A value obtained by dividing 0 to 360 ° of the relative rotation angle by a phase difference between output signals of the detection elements is n, and n angle ranges are set by dividing a range of 0 to 360 ° by n. In this case, the combinations of the comparison results between the signal levels of the detection signals and the threshold values are all different in each angle range. The rotation sensor described in 1. When the relative rotation angle is θ and the calculation angle is φ,
The angle calculation unit performs feedback control using a signal output from each magnetoelectric conversion element so that the deviation (2θ−2φ) becomes 0, and the calculation angle φ when the feedback control is started The rotation sensor according to any one of claims 1 to 6, wherein an initial value determined by the initial value determination unit is used as an initial value. The plurality of magnetoelectric transducers output a sin 2θ signal and a cos 2θ signal,
The angle calculator is
A sin (2θ-2φ) signal is created based on the sin2θ signal and the cos2θ signal, and a deviation (2θ-2φ) is calculated based on the created sin (2θ-2φ), and the deviation (2θ− The rotation sensor according to claim 7, wherein the relative rotation angle is calculated by performing feedback control so that 2Φ) becomes zero. The angle calculator is
A counter for counting a count value corresponding to the calculation angle φ is provided, the sign of the deviation (2θ-2φ) is determined, and the count value of the counter is increased or decreased based on the determination result. The rotation sensor according to claim 8. The angle calculator is
The rotation sensor according to claim 9, wherein the deviation (2θ-2φ) is calculated by performing an inverse sine operation on the sin (2θ-2φ) signal, and the positive / negative is determined based on the calculation result. . The angle calculator is
When the sin (2θ-2φ) signal is greater than 0, it is determined that the deviation (2θ-2φ) is positive, and when the sin (2θ-2φ) signal is less than 0, the deviation (2θ− The rotation sensor according to claim 9, wherein it is determined that 2φ) is negative.   The rotation sensor according to claim 1, wherein each of the plurality of detection elements is a Hall element.   The rotation sensor according to any one of claims 1 to 12, wherein each of the plurality of magnetoelectric conversion elements is a magnetoresistive element.

Images