JP2012189350A - Encoder device and correction method for encoder device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect, with high accuracy, a rotating angle or a moving position of a detection target while suppressing reduction of interpolation accuracy as further as possible.SOLUTION: A sensor unit outputs a phase-A output signal and a phase-B output signal which are sinusoidal and of which the phases are different at 90° from each other in accordance with the rotation or move of a detection target. Four sampling values VA1-VA4 are acquired correspondingly to a position of +90° and a position of -90° with respect to a maximum value Vand a minimum value Vof the phase-A output signal, and an average value of these sampling values is defined as an offset value. Regarding the phase-B output signal, the offset value is also similarly computed. By using these offset values, the rotating angle or moving distance of the detection target for which offset correction has been performed on the sampling values of the phase-A output signal and the phase-B output signal, is computed.

Description

  The present invention relates to an encoder device that detects a rotation angle or a movement position of a detection object, and in particular, an encoder device that corrects a sinusoidal detection signal output from a sensor unit in accordance with the rotation angle or movement position of a detection object. And a correction method for the encoder device.

Conventionally, a sensor unit that outputs first and second analog detection signals having sinusoidal phases different from each other by 90 degrees in accordance with rotation or movement of a detection target, and first and second analog detection signals from the sensor unit And a calculation processing unit that performs digital conversion into the first and second digital detection signals, respectively, and calculates the rotation angle or movement position of the detection target using the converted first and second digital detection signals. Encoder devices are well known. In this case, the first and second digital detection signals V _A (θ), V _B (θ) (hereinafter referred to as A-phase output signal V _A (θ) and B-phase output signal V _B (θ)) are expressed by the following equation (1). , 2 respectively, the rotation angle (movement position) ω is expressed by the following equation (3).

In this case, ideally, both the amplitude values A and B in the equations 1 and 2 are always equal, the offset values ΔA and ΔB are both “0”, and the A-phase output signal V _A (θ) and It is desirable that the B-phase output signal V _B (θ) is a complete sine wave signal. However, actually, the offset values ΔA and ΔB and the amplitude values A and B include an error latent in the sensor unit itself and an installation error between the sensor unit and the detection target (for example, a magnet), and the offset value ΔA, ΔB and amplitude values A and B fluctuate, and the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) do not become perfect sinusoidal signals. As a result, the rotation angle ω that is actually calculated by the arithmetic processing of Equation 3 includes an error. This error is called the interpolation accuracy φ, and the interpolation accuracy φ is usually expressed by the following equation 4 obtained by subtracting an angle (mechanical angle) that does not include an error from the rotation angle ω (electrical angle). The present inventors confirmed the influence of the offset values ΔA and ΔB and the amplitude values A and B on the interpolation accuracy φ by the calculation under the following first and second conditions.

(First condition)
Both the amplitude values A and B of the A phase output signal V _A (θ) and the B phase output signal V _B (θ) are equal, both amplitude widths are set to “2”, and the offset value ΔB is maintained at “0”. Each time the offset value ΔA is changed by 0.5% (that is, 0.01) within the range of 0 to 5% (that is, 0 to 0.1) of the amplitude width, and the offset value ΔA is changed by 0.5%. In addition, the interpolation accuracy φ was sequentially calculated by changing θ from 0 to 360 degrees. A value φ _Max −φ _Min (Error) obtained by subtracting the minimum value φ _Min from the maximum value φ _Max of the interpolation accuracy φ over 0 to 360 degrees of θ is indicated by black triangles in the graph of FIG.

(Second condition)
Both offsets ΔA and ΔB of the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) are both kept at “0”, and the ratio (1−A / B) of the amplitude values A and B is set to 0 to 0. The amplitude value is changed by 0.05% within a range of 5% (for example, if the amplitude value B is “1”, the amplitude value A is changed by 0.005 from 1 to 0.95), the amplitude values A and B Each time the ratio (1-A / B) was changed by 0.05%, the interpolation accuracy φ was sequentially calculated by changing θ from 0 to 360 degrees. A value φ _Max −φ _Min (Error) obtained by subtracting the minimum value φ _Min from the maximum value φ _Max of the interpolation accuracy φ over 0 to 360 degrees of θ is indicated by black square marks in the graph of FIG.

  As can be understood from this calculation result, the influence of the offset values ΔA and ΔB on the interpolation accuracy φ is very large and is slightly more than four times the amplitude ratio A / B. Therefore, it can be understood that the correction regarding the offset values ΔA and ΔB is much more important than the amplitude values A and B. In this sense, in the present invention, as will be described in detail later, correction relating to the offset values ΔA and ΔB is essential, and correction relating to the amplitude values A and B is selected.

  This type of correction has been conventionally performed, and is disclosed in, for example, Patent Document 1 below. Note that the encoder device disclosed in Patent Document 1 is arranged so as to face the side surface of a multipolar magnet that is magnetized so as to alternately divide the circumference of the N pole and S pole alternately on the side surface of the cylinder. It has a structure in which two magnetic sensors are arranged so as to output sinusoidal signals that are 90 degrees out of phase.

As a correction method, the following method is adopted. First, regarding the offset value, the maximum values V _Amax and V _Bmax and the minimum value V _Amin are respectively obtained for the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) ranging from 0 to 360 degrees. , V _Bmin are extracted, and the average value (amplitude center value) of the extracted maximum values V _Amax , V _Bmax and the minimum values V _Amin , V _Bmin is expressed as the offset value V _Aoff , Set as V _Boff respectively.

Then, by subtracting the set offset values V _Aoff and V _Boff from the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ), respectively, The phase output signal V _A (θ) and the B phase output signal V _B (θ) are offset-corrected.

そして、Ａ相出力信号Ｖ_A(θ)に関してはそのまま用いるが、Ｂ相出力信号Ｖ_B(θ)に関しては下記数１０に示すように振幅補正値Ａamを乗算することにより両振幅巾が同じになるように振幅補正するようにしている。

As for the amplitude values A and B, the minimum values V _Amin and V _Bmin are obtained from the maximum values V _Amax and V _Bmax in the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ), respectively. By subtracting, the amplitude widths V _Amax −V _Amin and V _Bmax −V _Bmin are respectively calculated, and the ratio of the amplitude width V _Amax −V _Amin to the amplitude width V _Bmax −V _Bmin is corrected as shown in _Equation 9 below. Calculate as the value Aam.

The A-phase output signal V _A (θ) is used as it is, but the B-phase output signal V _B (θ) is multiplied by the amplitude correction value Aam as shown in the following equation 10 so that both amplitude widths are the same. The amplitude is corrected so that

Then, using the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) that have been subjected to the offset correction of Equations 7 and 8 and the amplitude correction of Equation 10, the following Equation 11 is applied. The rotation angle ω is calculated by executing this calculation.

Although not directly related to the offset correction and the amplitude correction, for example, Patent Document 2 below shows an encoder device in which eight AMR magnetoresistive effect elements as detection elements are arranged in a point-symmetric manner in a sensor unit. ing. In this encoder device, the sensor unit is opposed to a dipole magnet that is an object to be detected, thereby minimizing the amplitude ratio latent in the sensor unit, and having a uniform sinusoidal A phase that is different in phase by 90 degrees from each other. The output signal V _A (θ) and the B-phase output signal V _B (θ) are output for two cycles by one rotation of the magnet. Patent Documents 3 and 4 below show encoder devices of the same type as described above, in which an SV-GMR magnetoresistive element and a Hall element are used as detection elements in the sensor section. Also in these encoder devices, in order to minimize the amplitude ratio latent in the sensor unit, a plurality of detection elements are arranged in a point-symmetric manner. Further, in these encoder devices, the sinusoidal A-phase output signal V _A (θ) and B-phase output signal V _B (θ) having a phase difference of 90 degrees from each other are equivalent to one cycle by one rotation of the magnet. Is output. In addition, in order to increase the mechanical resolution, the encoder devices described in Patent Documents 2 to 4 may be arranged on the side surfaces of linear magnets and ring magnets that are magnetized with multiple poles.

Japanese Patent Laid-Open No. 04-118513 JP-A-59-41822 Japanese Patent Laid-Open No. 10-70325 JP 2002-71381 A

  However, in this type of encoder device, as described above, an error that is latent in the sensor unit itself and the following installation error between the detection target and the sensor unit occur. That is, (1) an error due to a tilt with a detection target caused by eccentricity during rotation, (2) a deviation between the center of the detection target and the rotation axis, (3) a tilt of the sensor unit with respect to the detection target, and ( 4) An installation error occurs due to the displacement of the sensor from the line of the rotation axis. In addition, these installation errors (1) to (4) occur one or more.

  And even if it performs offset correction and amplitude like patent document 1 mentioned above with respect to the encoder apparatus shown by the said patent documents 2-4, the fall of an interpolation precision cannot fully be suppressed. This is because the error related to the offset value and amplitude ratio latent in the sensor unit itself and the influence of the installation error of the sensor unit on the detection target are corrected together, so the installation error affects the error latent in the sensor unit itself. This is considered to be because a decrease in interpolation accuracy cannot be suppressed.

  The present invention has been made to address the above problems, and an object of the present invention is to provide an encoder device capable of detecting a rotation angle or a moving position of a detection target with high accuracy while minimizing a decrease in interpolation accuracy. It is to provide a correction method for an encoder device. In addition, in the description of each constituent element of the present invention below, in order to facilitate understanding of the present invention, reference numerals of corresponding portions of the embodiment are described in parentheses, but each constituent element of the present invention is The present invention should not be construed as being limited to the configurations of the corresponding portions indicated by the reference numerals of the embodiments.

  In order to achieve the above object, the constitutional feature of the present invention is that the first and second analog detection signals in the form of sinusoids whose phases are different from each other by 90 degrees in accordance with the rotation or movement of the detection object (21, 22). The first and second analog detection signals from the sensor unit are digitally converted into first and second digital detection signals, respectively, and the converted first and second digital detection signals are converted into the first and second digital detection signals. An encoder device that is applied to an encoder device that includes an arithmetic processing unit (31 to 33) that performs arithmetic processing on a rotation angle or a moving position of a detection target using the encoder and corrects the converted first and second digital detection signals. The first and second analog detection signals from the sensor unit are sampled at predetermined intervals, respectively. (2) A data acquisition procedure (S14 to S18) for acquiring digital data, and a position detection procedure (S20, S26) for detecting the positions of the maximum value and the minimum value from the acquired first and second digital data, respectively. And from the acquired first digital data, the respective digital data at four positions that are 90 degrees apart from the respective detected positions of the maximum value and the minimum value of the detected first digital data are respectively extracted. And four correction digital data, and four of the acquired second digital data, each 90 degrees away from each detected position of the maximum value and the minimum value of the detected second digital data. Digital data for correction by extracting each digital data of position and making four digital data for correction An output procedure (S22, S28) and a value obtained by averaging four digital data for correction related to the extracted first digital data are set as a first offset value, and four corrections related to the extracted second digital data are set. Offset value setting procedure (S24, S30) for setting a value obtained by averaging the digital data for use as the second offset value, and the rotation angle or movement position of the detection object using the first and second digital detection signals in the arithmetic processing unit The offset value correction procedure (S44, S46) for correcting the first and second digital detection signals input from the sensor unit and converted by using the first and second offset values before the arithmetic processing is performed. There is.

  In the present invention configured as described above, in the correction digital data extraction procedure, each of the detection positions of the maximum value and the minimum value of the first digital detection signal corresponding to the first analog detection signal output from the sensor unit is determined. Each of the digital data at four positions separated by 90 degrees is extracted as four correction digital data, and is output from the sensor unit to be a second signal having a phase difference of 90 degrees from the first analog detection signal. The digital data at four positions that are 90 degrees apart from the detection positions of the maximum value and the minimum value of the second digital detection signal corresponding to the analog detection signal are extracted and used as four correction digital data. In the offset value setting procedure, a value obtained by averaging the four correction digital data related to the extracted first digital data is set as the first offset value, and the four correction digital data related to the extracted second digital data are set. A value obtained by averaging the data is set as the second offset value. Then, according to the offset value correction procedure, the arithmetic processing unit uses the first and second digital detection signals to calculate the rotation angle or the moving position of the detection target, and the first converted and input from the sensor unit. And the second digital detection signal are corrected using the first and second offset values, respectively. As a result, the interpolation accuracy caused by the installation error between the sensor unit and the detection object becomes good, and the rotation angle or movement position of the detection object can be detected with high accuracy.

  According to another feature of the present invention, the sensor unit may have a plurality of detection elements arranged symmetrically on a single substrate. According to this, even in an encoder apparatus having a high tolerance for installation errors, the rotation angle or the movement position of the detection target can be detected with higher accuracy.

  In the present invention, the correction related to the amplitude described in the section of the background art may be performed together. According to this, the rotation angle or the movement position of the detection target can be detected with higher accuracy.

  Furthermore, the present invention provides a correction method for the encoder device, wherein the sensor unit outputs sinusoidal first and second analog detection signals having phases different from each other by 90 degrees in accordance with the rotation or movement of the detection object. The first and second analog detection signals from the sensor unit are digitally converted into first and second digital detection signals, respectively, and the converted first and second digital detection signals are converted into the first and second analog detection signals, respectively. And an arithmetic processing unit that performs arithmetic processing on the rotation angle or moving position of the detection target using the corrected first and second digital detection signals. It may be. Also in this manner, as described above, the interpolation accuracy caused by the installation error between the sensor unit and the detection object becomes good, and the rotation angle or movement position of the detection object can be detected with high accuracy.

It is a graph which shows the magnitude | size of the error of the interpolation accuracy for demonstrating the influence which an offset value and an amplitude ratio have on interpolation accuracy. It is a block diagram which shows roughly the correction apparatus applied to the encoder apparatus which concerns on one Embodiment of this invention. It is a flowchart which shows the correction value acquisition program performed with the arithmetic processing apparatus of FIG. It is explanatory drawing for demonstrating the extraction operation | movement of the sampling value in the position of plus 90 degree | times and minus 90 degree | times from the maximum value and the minimum value. It is explanatory drawing for demonstrating the other extraction example of the said sampling value. It is explanatory drawing for demonstrating the other extraction example of the said sampling value. 1 is a block diagram schematically showing an encoder device according to an embodiment of the present invention. It is a flowchart which shows the rotation angle calculation program performed with the arithmetic processing apparatus of FIG. It is a schematic block diagram of the magnetic sensor which concerns on 1st Example of this invention. It is a figure which shows the equivalent circuit of the magnetic sensor of the said 1st Example. It is a schematic layout for demonstrating arrangement | positioning relationship between the magnetic sensor of the said 1st Example, and a magnet. It is a graph which shows the interpolation precision calculated using the conventional correction method in the magnetic sensor of the said 1st Example. It is a graph which shows the interpolation precision calculated using the correction method of this invention in the magnetic sensor of the said 1st Example. It is a schematic block diagram of the magnetic sensor which concerns on 2nd Example of this invention. It is a graph which shows the interpolation precision calculated using the conventional correction method in the magnetic sensor of the said 2nd Example. It is a graph which shows the interpolation precision calculated using the correction method of this invention in the magnetic sensor of the said 2nd Example. It is a schematic block diagram of the magnetic sensor which concerns on 3rd Example of this invention. It is a wave form diagram which shows the relationship of the A position output signal and B phase output signal which are output from the magnetic sensor of the said 3rd Example, and a sampling position. It is a graph which shows the interpolation precision calculated using the conventional correction method in the magnetic sensor of the said 3rd Example. It is a graph which shows the interpolation precision calculated using the correction method of this invention in the magnetic sensor of the said 3rd Example. It is a figure for demonstrating the magnetic field change in a state which concerns on the said 1st and 2nd Example, and there is no installation error. It is a figure for demonstrating the magnetic field change in a state with installation error concerning the said 1st and 2nd Example. It is a figure for demonstrating the magnetic field change in a state with an installation error concerning the said 3rd Example. It is a schematic block diagram of the magnetic sensor which concerns on the modification of this invention. It is a schematic layout for demonstrating the arrangement | positioning relationship with respect to the magnetic sensor which concerns on the said modification, and a multipolar ring magnet in a state without installation error. It is a schematic layout for demonstrating the arrangement | positioning relationship with respect to the magnetic sensor which concerns on the said modification, and a multipolar linear magnet in a state without installation error. It is a schematic layout for demonstrating arrangement | positioning relationship with respect to the magnetic sensor which concerns on the said modification, and a multipolar ring magnet in the state which has an installation error. It is a schematic layout for demonstrating the arrangement | positioning relationship with respect to the magnetic sensor which concerns on the said modification, and a multipolar linear magnet in the state which has an installation error. It is a graph which shows the interpolation accuracy calculated using the correction method by the conventional correction method and this invention in the magnetic sensor arrange | positioned like the said FIG. 24A. It is a graph which shows the interpolation precision calculated using the correction method by the conventional correction method and this invention in the magnetic sensor arrange | positioned like the said FIG. 24B. It is the schematic which shows the magnetic sensor and multipolar ring magnet which concern on the other modification of this invention.

a. Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 2 is a block diagram schematically showing a correction device applied to the encoder device according to the present invention. The magnetic sensor 10 constitutes a sensor unit of the present invention, and is disposed opposite to a magnet 22 fixed to a rotating shaft 21, and has a sinusoidal shape whose phases are different from each other by 90 degrees in accordance with the rotation of a detection object. An A-phase output signal V _A and a B-phase output signal V _B which are analog signals are output. The rotating shaft 21 and the magnet 22 constitute the detection object of the present invention.

In this case, as shown in FIG. 2, the magnetic sensor 10 may be disposed on the upper surface or the bottom surface of the disk-shaped magnet 22 that is divided in the circumferential direction and has the N-pole and the S-pole, which will be described later. As will be described in a specific embodiment, a magnetic sensor may be provided on the side surface of the cylinder so as to face the side surface of the magnet magnetized so as to alternately divide the circumference equally between the north and south poles. In addition, a magnetic sensor is arranged opposite to the side surface of the linear magnet magnetized in multiple poles, and the magnetic sensor has a sinusoidal A-phase output signal V having a phase difference of 90 degrees according to the linear movement of the magnet. _It may be configured to output the _A and B phase output signals V _B. Also for the magnetic sensor 10, two packages each including a magnetoresistive effect element are arranged at positions shifted by 90 degrees, and each package outputs an A-phase output signal V _A and a B-phase output signal V _B , respectively. Alternatively, the magnetic sensor 10 may be configured to include a plurality of magnetoresistive elements in one package so that the A phase output signal V _A and the B phase output signal V _B are output from one package. May be. Further, it may be a magnetic sensor 10 of a half bridge type, a full bridge type, a double bridge type or the like in which a large number of magnetoresistive elements are arranged symmetrically in one package.

A sampling circuit 11 and A / D converters 12 a and 12 b are connected to the output terminal of the magnetic sensor 10. The sampling circuit 11 is controlled by an arithmetic processing unit 13 which will be described later, and samples the A phase output signal V _A and the B phase output signal V _B that are analog signals from the magnetic sensor 10 respectively in a predetermined cycle, and performs A / D conversion. Output to the units 12a and 12b, respectively. The A / D converters 12 a and 12 b A / D convert the sampled values of the sampled A-phase output signal V _A and B-phase output signal V _B , respectively, and output them to the arithmetic processing unit 13. The arithmetic processing unit 13 is composed of a computer device including a CPU, ROM, RAM, and other memory devices, and stores and executes the correction value acquisition program shown in FIG. 3 in this embodiment.

  An input device 14, a display device 15, and an input / output circuit 16 are connected to the arithmetic processing device 13. The input device 14 includes operation switches and the like, and is used for operating instructions of the arithmetic processing device 13. The display device 15 displays an operation instruction, operation content, operation result, and the like of the arithmetic processing device 13. The input / output circuit 16 exchanges data with other devices. The arithmetic processing unit 13 is connected to a memory 17 for storing data generated by the arithmetic processing unit 13.

A driving device 23 including a motor and the like is also connected to the arithmetic processing device 13. The driving device 23 is controlled by the arithmetic processing device 13 to rotate the rotating shaft 21 and the magnet 22. In addition, as mentioned above, when the magnet 22 moves linearly, the drive device 23 drives the magnet 22 linearly. The rotating shaft 21, the magnet 22, and the magnetic sensor 10 are included in a device that is used for detecting a rotation angle or a moving position using the A-phase output signal V _A and the B-phase output signal V _B output from the magnetic sensor 10. It is already built in. The other devices and circuits 11 to 17 and 23 may be different from the device using the detection of the rotation angle or the movement position, but may be incorporated in the device.

Next, the operation of the embodiment configured as described above will be described. In this case, the arithmetic processing device 13 starts executing the correction value acquisition program of FIG. 3 according to an instruction using the input device 14 or the like of the user. This correction value acquisition program is started in step S10, and the arithmetic processing unit 13 starts driving control of the driving unit 23 in step S12 and starts rotating the rotary shaft 21 and the magnet 22 at a predetermined constant speed. As described above, when moving the magnet 22 linearly, the magnet 22 starts to move linearly at a constant speed. As a result, the magnet 22 starts rotating (moving) at a constant speed, and the magnetic sensor 10 starts outputting the analog A-phase output signal V _A and the B-phase output signal V _B to the sampling circuit 11 respectively. After the processing in step S12, the arithmetic processing unit 13 instructs the sampling circuit 11 to start sampling in step S14. In this case, the number of samplings per cycle of the analog A-phase output signal V _A and B-phase output signal V _B is inversely proportional to the rotational speed of the magnet 22 and proportional to the sampling rate of the sampling circuit 11. In the present embodiment, since the number of samplings per cycle of the analog A-phase output signal V _A and B-phase output signal V _B is determined to be a fixed number (for example, 1024), the arithmetic processing unit 13 At least one of the sampling rate by the sampling circuit 11 and the rotational speed of the magnet 22 by the driving device 23 is controlled.

After the processing in step S14, the arithmetic processing unit 13 in step S16 samples the digital A-phase output signals V _A and B sampled by the sampling circuit 11 and digitally converted by the A / D converters 12a and 12b, respectively. The sampling value of the phase output signal V _B is taken in and stored. And the arithmetic processing unit 13 confirms a drive end point in step S18. Here, it is determined whether the magnet 22 has rotated one or more times. This determination is performed by the rotation speed of the magnet 22 and time measurement by a timer. Until the magnet 22 rotates one or more times, the arithmetic processing unit 13 determines “No” in step S18, and takes in the sampling values of the A-phase output signal V _A and the B-phase output signal V _B in step S16 and Continue to execute the storage process repeatedly. When the magnet 22 rotates one or more times, the arithmetic processing unit 13 determines “Yes” in step S18 and proceeds to step S20.

In step S20, the maximum value V _Amax and the minimum value V _Amin are extracted from the sampling values for at least one period of the A-phase output signal V _A. Next, in step S22, the arithmetic processing unit 13 has four sampling values VA1, VA2, corresponding to the positions of plus 90 degrees and minus 90 degrees with respect to the positions of the maximum value V _Amax and the minimum value V _Amin . VA3 and VA4 are extracted from sampling values for at least one period of the A-phase output signal V _A.

The extraction of the four sampling values VA1, VA2, VA3, VA4 will be described with reference to the drawings. FIG. 4 shows an A-phase output signal V _A and a B-phase output signal V _B when the SV-GMR magnetoresistive effect element or Hall element is used as the magnetic sensor 10 as in the above-mentioned Patent Documents 3 and 4, that is, magnets. In this example, the A phase output signal V _A and the B phase output signal V _B for one cycle are output by one rotation of 22. In the A-phase output signal V _A of FIG. 4, as indicated by a circle, the maximum value V _Amax and the minimum value V _Amin is extracted, then, as indicated by triangles, maximum V _Amax and the minimum value V _Amin Four sampling values VA1, VA2, VA3, and VA4 corresponding to a position of plus 90 degrees and a position of minus 90 degrees with respect to the position are extracted. In this case, the number of samplings corresponding to 90 degrees is 256, and 256 sampling values that are separated from the sampling values of the maximum value V _Amax and the minimum value V _Amin before and after are the four sampling values VA1, VA2, VA3, VA4. Will be extracted as

As shown in FIG. 5, four sampling values VA1, VA2, VA3, and VA4 corresponding to positions of plus 90 degrees and minus 90 degrees with respect to the positions of the maximum value V _Amax and the minimum value V _Amin are directly obtained. In the case where the sampling value VA2 corresponding to the position of minus 90 degrees with respect to the position of the maximum value V _Amax cannot be directly extracted, for example, as shown in FIG. Sampling values (sampling value VA2) at the phase position corresponding to the later period are extracted as sampling values VA1, VA2, VA3, and VA4.

FIG. 6 shows the A-phase output signal V _A and the B-phase output signal V _B when the AMR magnetoresistive effect element is used as the magnetic sensor 22 in the same manner as in Patent Document 2 described above, that is, one rotation of the magnet 22. An example is shown in which the A-phase output signal V _A and the B-phase output signal V _B for two cycles are output. Therefore, in this case, since 90 degrees corresponds to the sampling number 128, the sampling values that are 128 back and forth from the sampling values of the maximum value V _Amax and the minimum value V _Amin are the four sampling values VA1, VA2, and VA3. , VA4. Also in this case, if the four sampling values VA1, VA2, VA3, and VA4 are not directly extracted, the sampling values at the corresponding phase positions in the previous or subsequent cycle are the sampling values VA1, VA2, and VA3. , VA4. Further, in this case, the sampling values may be taken in and processed for half the time in the case of FIG. 4 by the processing in steps S16 and S18.

After the process of step S22, the arithmetic processing unit 13 performs the four sampling values VA1, 1 by executing the following equation 12 using the extracted four sampling values VA1, VA2, VA3, VA4 in step S24. The average value of VA2, VA3, _VA4 is calculated as the offset value V _Aoff of the A-phase output signal V _A.

After processing at step S24, the processing unit 13, the sampling values of one period of the B-phase output signal V _B, by performing the processing similar steps S26~S30 as in step S20 to S24, B The maximum value V _Bmax and the minimum value V _Bmin are extracted from the sampling values of at least one period of the phase output signal V _B , and the position of plus 90 degrees with respect to the position of the maximum value V _Bmax and the minimum value V _Bmin Four sampling values VB1, VB2, VB3, and VB4 corresponding to the position of minus 90 degrees are extracted, and the offset value V _Boff of the B-phase output signal V _B is calculated by the following arithmetic processing.

Then, the processing unit 13, at step S32, the amplitude width V _Amax -V _Amin of the A-phase output signal V _A by subtracting the minimum value V _Amin from the maximum value V _Amax of the A-phase output signal V _A calculated, the amplitude width V _Bmax -V _Bmin of B-phase output signal V _B calculated by subtracting the minimum value V _Bmin from the maximum value V _Bmax of B-phase output signal V _B, these amplitudes width V _Amax -V _The amplitude correction value Aam is calculated by executing the following equation (14) for calculating the ratio of _Amin , V _Bmax -V _Bmin . The amplitude correction value Aam is the same as the conventional method.

Next, the arithmetic processing unit 13 stores the offset values V _Aoff , V _Boff and the amplitude correction value Aam calculated by the arithmetic processing of the formulas 12 to 14 in the memory 17 or other devices via the input / output circuit 16. For example, it outputs to the encoder apparatus which concerns on one Embodiment of this invention mentioned later. The encoder apparatus inputs and stores these offset values V _Aoff and V _Boff and the amplitude correction value Aam. After the process of step S34, the arithmetic processing unit 13 ends the execution of the correction value acquisition program in step S36.

Next, the correction process of the encoder device using the offset values V _Aoff and V _Boff and the amplitude correction value Aam calculated in this way will be described. As shown in FIG. 7, the encoder device includes a magnetic sensor 10 opposed to a rotation shaft 21 and a magnet 22 that are detection objects. In the correction device of FIG. 2, as described above, the offset values V _Aoff and V _Boff and the amplitude correction value Aam are calculated in a state where the magnetic sensor 10 is disposed opposite to the rotation shaft 21 and the magnet 22 which are detection objects. The magnetic sensor 10, the rotating shaft 21, and the magnet 22 shown in FIG. 7 are the same as those shown in FIG. The encoder device also includes a sampling circuit 31, A / D converters 32 a and 32 b, an arithmetic processing device 33, an input / output circuit 34, and a memory 35. These sampling circuit 31, A / D converters 32a and 32b, arithmetic processing unit 33, input / output circuit 34 and memory 35 are the same as the sampling circuit 11, A / D converters 12a and 12b, and arithmetic processing unit shown in FIG. 13, the same configuration as the input / output circuit 16 and the memory 17. The sampling circuit 31, the A / D converters 32a and 32b, the arithmetic processing unit 33, the input / output circuit 34, and the memory 35 are the same as the sampling circuit 11, the A / D converters 12a and 12b, and the arithmetic processing unit shown in FIG. 13, the input / output circuit 16 and the memory 17 may be provided separately or in common. In particular, the sampling circuit 31 and the A / D converters 32a and 32b may be common to the sampling circuit 11 and the A / D converters 12a and 12b in FIG. However, the arithmetic processing unit 33 stores a rotation angle calculation program and executes the rotation angle calculation program.

Next, the operation of the encoder apparatus configured as described above will be described. Also in this encoder apparatus, the arithmetic processing unit 33 instructs the sampling circuit 31 to sample the A-phase output signal V _A and the B-phase output signal V _B from the magnetic sensor 10 at a predetermined cycle. The sampling circuit 31 samples the A-phase output signal V _A and the B-phase output signal V _B from the magnetic sensor 10 at the predetermined period, and outputs the sampling values VA and VB to the A / D converters 32a and 32b, respectively. To do. The A / D converters 32 a and 32 b A / D convert the sampling values VA and VB, respectively, and output the digitally converted sampling values VA and VB to the arithmetic processing unit 33. As a result, the sampling values VA and VB of the A-phase output signal V _A and the B-phase output signal V _B from the magnetic sensor 10 that change according to the rotation angle ω of the rotating shaft 21 and the magnet 22 are given to the arithmetic processing unit 33. It will be supplied in a cycle.

The arithmetic processing unit 33 executes a rotation angle calculation program for each input of the sampling values VA and VB. Execution of the rotation angle calculation program is started in step S40 of FIG. 8, and the arithmetic processing unit 33 performs sampling values VA and VB which are A / D converted by the A / D converters 32a and 32b, respectively, in step S42. Enter each. Next, the arithmetic processing unit 33 performs an offset correction on the input sampling value VA in step S44 by the arithmetic processing of the following equation 15 using the stored offset value V _Aoff, and the input sampling value in step S46. VB is offset-corrected by the arithmetic processing of the following equation 16 using the stored offset value V _Boff .

Next, in step S48, the arithmetic processing unit 33 corrects the amplitude of the input sampling value VB by the arithmetic processing of the following Expression 17 using the stored amplitude correction value Aam. This amplitude correction is the same as the conventional method.

In step S50, the arithmetic processing unit 33 calculates the rotation angle ω of the rotating shaft 21 and the magnet 22 by the arithmetic processing of the following equation 18 using the corrected sampling values VA and VB.

  After the calculation of the rotation angle ω, the arithmetic processing unit 13 stores the calculated rotation angle ω in the memory 35 in step S52 or other devices such as the rotation shaft 21 and the magnet via the input / output circuit 34. Or output to a utilization device that uses the rotation angle ω of 22. Then, the arithmetic processing unit 33 ends the execution of the rotation angle calculation program in step S54. Thereafter, when the sampling values VA and VB from the A / D converters 32a and 32b are input again to the arithmetic processing unit 33, the arithmetic processing unit 33 executes the rotation angle calculation program consisting of the aforementioned steps S40 to S54 again. Thus, the sampling values VA and VB are corrected, and the rotation angle ω is calculated using the corrected sampling values VA and VB. By executing the rotation angle calculation program, the arithmetic processing unit 33 calculates and outputs the rotation angle ω of the rotating shaft 21 and the magnet 22.

Therefore, according to the present embodiment, the maximum values V _Amax and V _Bmax and the minimum values V _Amin and V _Bmin of the A-phase output signal V _A and the B-phase output signal V _B are obtained by the arithmetic processing of Equations 12-14. The average values of the four sampling values VA1 to VA4 and VB1 to VB4 corresponding to the plus 90 degree position and the minus 90 degree position are calculated as offset values V _Aoff and V _Boff respectively, and the A phase output signal The ratio between the amplitude width V _Amax −V _{Amin of} V _A and the amplitude width V _Bmax −V _Bmin of the B-phase output signal V _B is calculated as the amplitude correction value Aam. The sampling values VA and VB are offset and amplitude corrected using the offset values V _Aoff and V _Boff and the amplitude correction value Aam by the arithmetic processing of the _{formulas 15-17} . Then, the rotational angle ω is calculated by using the corrected sampling values VA and VB by the arithmetic processing of Equation 18.

b. Verification by Experiment of the Present Invention Next, the effect of the correction method according to the above embodiment will be verified by using an experimental result.
(1) First Example A magnetic sensor 10 according to a first example is a double full-bridge type magnetic sensor, and includes two sets of full-bridge magnetoresistive elements (in this case, AMR magnetoresistive elements). They are formed on a single substrate inclined at 45 degrees from each other. Specifically, as shown in FIG. 9A, the extending directions are shifted from each other by 45 degrees, and the magnetoresistive elements 1 to 8 are arranged along the circumferential direction so that the magnetism in the same extending direction with respect to the center of rotation. Resistive effect elements are arranged at point-symmetric positions. Then, as shown in the equivalent circuit of FIG. 9B, two sets of full bridge circuits are formed. In FIG. 9B, the specified voltage is applied between the terminals Vcc and Gnd, the phase A output signal V _A is output from the terminals VoutA + and VoutA−, and the phase B output signal V _B is output from the terminals VoutB + and VoutB−. The

A disk-shaped dipole magnet 22 having a diameter of 9 mm and arranged in the circumferential direction is fixed to one end of the rotating shaft 21 so as to face the magnetic sensor 10. The axis (rotation center) of the rotation shaft 21 and the center line (rotation center) of the magnet 22 are coincident. Using the rotating shaft 21 and the magnet 22 that are detection objects configured as described above and the magnetic sensor 10, the magnet 22 is rotated once under the following five conditions (a) to (e), and the magnetic sensor 10 is rotated. The A-phase output signal V _A and the B-phase output signal V _B in the first cycle output from the above are sampled at a sampling number of 2000 points per 360 degrees (per rotation of the magnet 22). Then, the sampled values of the sampled A-phase output signal V _A and B-phase output signal V _B are corrected using a conventional correction method, and the corrected sampling values of the A-phase output signal V _A and the B-phase output signal V _B are obtained. The interpolation accuracy φ (electrical angle−mechanical angle) was calculated.

That is, the A-phase output signal V _A and the B-phase output signal V _B are offset-corrected using the above formulas 5 to 8, and the amplitude is corrected using the formulas 9 and 10, and the corrected A-phase output signal is corrected. The interpolation accuracy φ was calculated for each of the following conditions (a) to (d) by executing the calculation of Equation 4 using the V _A and B phase output signals V _B.
(a) The center line of the magnetic sensor 10 is made to coincide with the axis of the rotary shaft 21 and the magnet 22.
(b) The center line of the magnetic sensor 10 is shifted +0.5 mm in the X direction with respect to the axis of the rotary shaft 21 and the magnet 22.
(c) The center line of the magnetic sensor 10 is shifted by −0.5 mm in the X direction with respect to the axis of the rotary shaft 21 and the magnet 22.
(d) The center line of the magnetic sensor 10 is shifted +0.5 mm in the Y direction with respect to the axis of the rotary shaft 21 and the magnet 22.
(e) The center line of the magnetic sensor 10 is shifted by −0.5 mm in the Y direction with respect to the axis of the rotary shaft 21 and the magnet 22.

  The interpolation accuracy φ based on the experimental results for each of the above conditions (a), (b), (c), (d), and (e) is expressed as (a), (b), (c), and (d) in FIG. , (E) are shown as errors in the graph. According to this, it is understood that the interpolation accuracy φ (error) hardly appears when “the center line of the magnetic sensor 10 is made coincident with the axis of the rotating shaft 21 and the magnet 22” in the condition (a). it can. On the other hand, as in the conditions (b), (c), (d), (e), “the center line of the magnetic sensor 10 is set in the X direction and the Y direction with respect to the axis of the rotary shaft 21 and the magnet 22. It can be understood that the interpolation accuracy φ (error) increases, that is, deteriorates.

Next, using the magnetic sensor 10 and the magnet 22 shown in FIGS. 9A to 9C, the magnet 22 is rotated once under the five conditions (a) to (e) in the same manner as in the above case. The output A phase output signal V _A and B phase output signal V _B in the first cycle were sampled at a sampling number of 2000 points per 360 degrees. Then, the sampled values of the sampled A-phase output signal V _A and B-phase output signal V _B are corrected using the correction method of the present invention, and the corrected sampling values of the A-phase output signal V _A and B-phase output signal V _B are corrected. Was used to calculate the interpolation accuracy φ (electrical angle−mechanical angle). That is, the A-phase output signal V _A and the B-phase output signal V _B are corrected by performing offset correction using the above equations 12, 13, 15, and 16, and performing amplitude correction using the above equations 14 and 17, respectively. The interpolation accuracy φ was calculated for each of the conditions (a) to (d) by executing the calculation of Equation 4 using the A-phase output signal V _A and the B-phase output signal V _B.

(A), (b), (c), (d), and (e) in FIG. 11 correspond to the above-mentioned conditions (a), (b), (c), (d), and (e), respectively. The interpolation accuracy φ when using this correction method is shown as an error in the graph. According to this, the above condition (a)
, (B), (c), (d), and (e), when the center line of the magnetic sensor 10 is shifted in the X and Y directions with respect to the axis of the rotary shaft 21 and the magnet 22, respectively, Even if the magnetic sensor 10 is installed with an installation error in the X direction and the Y direction with respect to the rotating shaft 21 and the magnet 22, it is understood that if the offset correction according to the present invention is performed, the interpolation accuracy φ (error) can be kept small. it can.

(2) Second Example Next, a second example will be described. Also in this case, the A phase output signal V _A corrected using the same conventional magnetic sensor 10 and the correction method of the present invention as in the first embodiment, using the same magnetic sensor 10 and magnet 22 as in the first embodiment. The interpolation accuracy φ (electrical angle−mechanical angle) was calculated using the sampling value of the B phase output signal V _B. However, in this case, in order to confirm the influence of the composite installation error of the magnetic sensor 10 on the rotating shaft 21 and the magnet 22, the schematic perspective view of FIG. 12A and the schematic front view of FIG. As shown, the magnet 22 is attached with an eccentricity of 0.5 mm with respect to the axis of the rotating shaft 21. The magnetic sensor 10 is arranged with a shift of +0.5 mm in the X direction with respect to the magnet 22.

  13 shows the interpolation accuracy φ when the conventional correction method is used as an error in the graph, and FIG. 14 shows the interpolation accuracy φ when the correction method of the present invention is used as an error in the graph. According to this, it can be understood that a large error occurs when the conventional correction method is used even if the composite magnetic sensor 10 is installed. On the other hand, it can be understood that the error is suppressed when the correction method of the present invention is used.

(3) Third Example Next, a third example will be described. In this case, unlike the first and second embodiments, as shown in FIG. 15, a ring magnet that is divided into 10 in the circumferential direction and magnetized with 10 poles is used as the magnet 22. The magnet 22 is also rotated by a rotating means (not shown) around the central axis in the direction perpendicular to the paper surface of FIG. In this case, the diameter of the magnet 22 is 7 mm. In this third embodiment, the magnetic sensor 10 is configured in the same manner as in the first and second embodiments, but the extending surface of the magnetoresistive effect element is parallel to the radial direction of the magnet 22. It is arranged on the normal line on the side of the magnet 22.

Also in this case, the magnet 22 is rotated once around the central axis, and 2000 sampling values are obtained from the A-phase output signal V _A and the B-phase output signal V _B from the magnetic sensor 10 during this one rotation. To do. FIG. 16 is a waveform diagram showing the relationship between the A-phase output signal V _A and the B-phase output signal V _B from these magnetic sensors 10 and the sampling position. In this case, since the magnet 22 is magnetized 10 poles in the circumferential direction, a sine-wave A-phase output signal V _A and a B-phase output signal V _{B of} 10 waves per rotation of the magnet 22 are output.

Then, in this case, the first period from the sinusoidal 10 waves A phase output signal V _A and the B phase output signal V _B, 2-th cycle and the third cycle of the A-phase output signal V _A and the B phase output signal The sampling values of V _B were extracted, and the interpolation accuracy φ was calculated by applying the conventional correction method to the sampling values of the first period, the second period, and the third period, respectively. (A), (b), and (c) of FIG. 17 are respectively related to the sampling values for the first period, the second period, and the third period of the A-phase output signal V _A and the B-phase output signal V _B. The interpolation accuracy φ when using this correction method is shown as an error in the graph. According to this, in the magnetic sensor 10 of this type, when the conventional correction method is used, the A-phase output signal V _A and the B-phase output signal V _B in the first cycle, the second cycle, and the third cycle are used. It can be understood that a large error (interpolation accuracy φ) is included. This is because the magnetic sensor 10 is made to face in the rotation (or movement) direction of the magnet 22 having a plurality of magnetic poles as in the third embodiment, and the A-phase output signal V _A and the B-phase output signal from the magnetic sensor 10 are used. When the rotation angle of the magnet 22 (detection target) is detected based on V _B , this means that a large error is included in the detection of the rotation angle.

On the other hand, (a), (b), and (c) of FIG. 18 show the sampling values for the first period, the second period, and the third period of the A phase output signal V _A and the B phase output signal V _B. The interpolation accuracy φ when using the correction method of the present invention is shown as an error in the graph. In this case, since the sampling number for 10 cycles is “2000”, the sampling number for one cycle of the A-phase output signal V _A and the B-phase output signal V _B is “200”. Therefore, the number of samplings corresponding to 90 degrees is “50”. The digital data for correction of the A phase output signal V _A and the B phase output signal V _{B in} the first cycle will be described with reference to the example shown in FIG. The positions of the maximum value V _Amax and the minimum value V _Amin of the A-phase output signal V _{A in} the first cycle are “83” and “183”. Accordingly, the four positions that are 90 degrees apart from the positions “83” and “183” are “33”, “133”, “133”, and “233”, respectively. The sampling values at these positions become four correction digital data for the A-phase output signal V _{A in} the first period. The positions of the maximum value V _Bmax and the minimum value V _Bmin of the B-phase output signal V _{B in} the first cycle are “32” and “134”. Therefore, the four positions 90 degrees apart from the positions “32” and “134” are “1982”, “82”, “84”, and “184”, respectively. Then, the sampling values at these positions become four correction digital data for the B-phase output signal V _{B in} the first cycle.

According to the graphs shown in FIGS. 18A, 18B and 18C showing the interpolation accuracy φ (error) using the correction method according to the present invention, this kind of magnetic sensor 10 also has the present invention. When the correction method is used, it can be understood that errors included in the A-phase output signal V _A and the B-phase output signal V _B in the first period, the second period, and the third period can be suppressed. As a result, the A-phase output signal V _A and the B-phase output signal from the magnetic sensor 10 are made to face the magnetic sensor 10 in the rotation (or movement) direction of the magnet 22 having a plurality of magnetic poles as in the third embodiment. Even when the rotation angle of the magnet 22 (detection target) is detected based on V _B , it can be understood that the rotation angle error can be suppressed by using the correction method according to the present invention.

c. Theoretical description of the present invention The following can be confirmed in various examples. Interpolation accuracy φ (FIG. 10 (b) to FIG. 10 (e), FIG. 13 and FIG. 17 (a) to FIG. 17 (b) when correction is performed with the amplitude center value which is the conventional method in the first to third embodiments. c)) and a change in the error in the mechanical angle in the interpolation accuracy φ (FIGS. 25A (a) and 25B (a)) when corrected with the amplitude center value which is the conventional method in the first and second modified examples 1 and 2 described later. Indicates a uniform gradient. 10 (b), FIG. 13, FIG. 17 (a) to FIG. 17 (c), FIG. 25A (a) and FIG. 25B (a) approximate to the −cos waveform, and FIG. 10 (c) and FIG. ) Approximates a cos waveform, and FIG. 10D approximates a sin waveform, each having a gradient of one period. The change in error in the mechanical angle of the interpolation accuracy φ is defined as a gradient, and the correction method according to the above embodiment will be theoretically described with respect to the above first to third examples. The first to third embodiments can be understood by assuming a cardioid curve of polar coordinates (r, θ).

  First, in the case of the first embodiment, when there is no installation error of the magnetic sensor 10 with respect to the magnet 22, that is, in the state (a) described above, the magnetic field change on the black circle shown in FIG. X = cos θ and y = sin θ. On the other hand, a state in which the installation error of the magnetic sensor 10 with respect to the magnet 22 is shifted in any of the above-described directions (b) to (e) is indicated by a black circle in FIG. The magnetic field change at the black circle is as shown in FIG. In this way, the state in which the installation error of the magnetic sensor 10 with respect to the magnet 22 is shifted in one of the above-described (b) to (e) is represented by the equation r = 1 + a ′ · cos θ, and orthogonal coordinates (x, y ), X = r · cos θ, y = r · sin θ, and x and y represent magnetic field changes caused by the motion of the detection target. The change in the magnetic field when there is no displacement is a ′ = 0, that is, r = 1. When a positional deviation occurs, a 'increases in proportion to the magnitude of the positional deviation.

When r = 1 + a ′ · cos θ is substituted into x = r · cos θ and y = r · sin θ, x and y are expressed as in the following equations 19 and 20.

これにより、設置誤差による磁界Ｈ_A，Ｈ_B の変化には、2次高調波が一定の関係で重畳していることが分かる。 The installation errors (b) to (e) shown in the first embodiment are actually not only in these four directions but also in the full-angle (α = 0 to 360 °) range, and therefore the detection target object at that position. The magnetic fields H _A and H _B respectively corresponding to the A-phase output signal V _A and the B-phase output signal V _B can be expressed as the following formulas 21 and 22.

Thus, it can be seen that the second harmonic is superimposed on the change of the magnetic fields H _A and H _B due to the installation error in a fixed relationship.

前記数２３，２４を整理すると、下記数２５，２６のように変形される。

Next, the second embodiment will be described. In this case, it can be considered that there are two kinds of positional shifts and different second harmonics are superimposed. Therefore, the magnetic fields H _A and H _B are expressed by the following equations 23 and 24.

When the equations 23 and 24 are arranged, the following equations 25 and 26 are transformed.

これにより、この場合も、設置誤差による磁界Ｈ_A，Ｈ_B の変化には、2次高調波が一定の関係で重畳していることが分かる。 Here, when the above formulas 25 and 26 are modified using c and d defined by the following formulas 27 to 30, the following formulas 31 and 32 are obtained.

Thus, in this case as well, it can be seen that the second harmonic is superimposed on the change of the magnetic fields H _A and H _B due to the installation error in a fixed relationship.

  Next, the third embodiment will be described. In the third embodiment, the magnetic sensor 10 is installed on the normal line of the magnet 22, and each magnetoresistive element of the magnetic sensor 10 has a point-symmetric shape and has different directions (for example, a normal direction and a tangent line). Direction) magnetic field. The magnetic field rotates on the distorted base line at the magnetic field changes in the different directions and the side surfaces of the ring. Therefore, the change of the magnetic field draws a cardioid curve similar to that of FIG. 20 of the first embodiment, as shown in FIG.

仮に、磁気センサ１０内に、感度違いとオフセット電圧がない状態（すなわち、Ａ＝ＢかつΔＡ＝ΔＢ＝０）で、設置誤差が生じた場合の各出力信号は下記数３５，３６で表される。

Next, an output signal from the magnetic sensor 10 is considered. A magnetic sensor 10, when the output signal from the B phase _{V A,} and V _B, the output signal is expressed by the following Expression 33. However, A and B are sensitivities, and ΔA and ΔB are offset voltages.

Assuming that there is no difference in sensitivity and no offset voltage in the magnetic sensor 10 (ie, A = B and ΔA = ΔB = 0), each output signal when an installation error occurs is expressed by the following equations 35 and 36. The

ここで、下記数３８，３９のようｘ，ｙを定義すると、前記数３７は下記数４０のように表される。

On the other hand, the interpolation accuracy φ (electrical angle−mechanical angle) is defined by the following equation 37 (same as the equation 4) as described above.

Here, when x and y are defined as the following formulas 38 and 39, the formula 37 is expressed as the following formula 40.

前記数３８，３９は下記数４２，４３のように変形され、公式を用いた変形により前記数４１は下記数４４のように変形される。

Here, when the tangents of both sides of the equation 40 are taken, the 40 is transformed as the following equation 41.

The formulas 38 and 39 are transformed into the following formulas 42 and 43, and the formula 41 is transformed into the following formula 44 by transformation using a formula.

この数４５に前記数３５，３６で表されるＶ_A(θ)，Ｖ_B(θ)を代入すると、前記数４５は下記数４６のように表される。

そして、この数４６をさらに変形すると、下記数４７，４８のようになる。

When the equation 44 is returned to the arctangent, the equation 44 becomes the following equation 45.

By substituting V _A (θ) and V _B (θ) represented by the equations 35 and 36 into the equation 45, the equation 45 is expressed as the following equation 46.

Further, when this equation 46 is further modified, the following equations 47 and 48 are obtained.

Here, considering that the value A is extremely larger than the value a (in other words, the ratio value a / A is extremely small), the denominator of the equation 48 is a constant, and the interpolation accuracy φ is as follows: The slope of is one cycle.
(A) When α = 0 °, the gradient of the interpolation accuracy φ is a sin waveform.
(B) When α = 90 °, the gradient of the interpolation accuracy φ is a cosine waveform.
(C) When α = 180 °, the gradient of the interpolation accuracy φ is a −sin waveform.
(D) When α = 270 °, the gradient of the interpolation accuracy φ is a −cos waveform.

Next, not the difference in sensitivity latent in the magnetic sensor 10 itself (A = B), but the A phase output signal V _A (θ) and the B phase output signal V _B (θ) when the offset voltage is latent are as follows. 49, 50.

そして、この数５１をさらに変形すると、下記数５２，５３のようになる。

Interpolation accuracy φ (electrical angle−mechanical angle) is expressed by the equation 45, and the A phase output signal V _A (θ) and the B phase output signal V _B (θ) of the equations 49 and 50 are added to the equation 45. When substituted, the following formula 51 is obtained.

Further, when this number 51 is further modified, the following numbers 52 and 53 are obtained.

Here, considering that the value A is extremely larger than the values ΔA and ΔB (in other words, the ratio values ΔA / A and ΔB / A are extremely small), the denominator of the equation 53 is a constant, and Thus, the gradient of the interpolation accuracy φ is one cycle.
(A) When ΔB = 0 and ΔA> 0, the gradient of the interpolation accuracy φ is a cos waveform.
(B) When ΔB = 0 and ΔA <0, the gradient of the interpolation accuracy φ is a −cos waveform.
(C) When ΔA = 0 and ΔB> 0, the gradient of the interpolation accuracy φ is a −sin waveform.
(D) When ΔA = 0 and ΔB <0, the gradient of the interpolation accuracy (φ) is a sin waveform.
According to Equation 48 and Equation 53 representing the interpolation accuracy φ, the influence of the output signal in which the second harmonic is superimposed due to the installation error and the influence of the offset voltage latent in the magnetic sensor 10 itself are the same phenomenon. I understand that

Next, an optimum correction value for suppressing the gradient of the interpolation accuracy φ of the output signal in which the offset voltage latent in the magnetic sensor 10 itself and the second harmonic due to the installation error are superimposed is calculated. In this case, the signals shown in the following equations 54 and 55 are defined as the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ).

そして、前記数５６を変形すると、下記数５７のようになる。

Substituting the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) represented by the equations 54 and 55 into the expression of the interpolation accuracy φ defined by the equation 45, the interpolation is performed. The accuracy φ is given by the following equation 56.

Then, when the equation 56 is modified, the following equation 57 is obtained.

そして、前記数５８を変形すると、下記数５９のようになる。

The improvement of the interpolation accuracy φ is given by setting φ = 0 in the equation 57. Any value of θ may be used as long as the coefficient of each term in the molecular formula is “0”. When the offset voltage values at that time are ΔA ₀ and ΔB ₀ , the following _equation 58 is established.

When the above equation 58 is transformed, the following equation 59 is obtained.

これにより、最適補正値Ｘ，Ｙは、下記数６２，６３によって表されることになる。

Thus, the term of sin θ is a · cos α = ΔB ₀ , and the term of cos θ is −a · sin α = ΔA ₀ . Accordingly, if the optimum correction values of the offset of the output signal on which the second harmonic is superimposed are X and Y, the values ΔA ₀ and ΔB ₀ are expressed by the following formulas 61 and 62.

As a result, the optimum correction values X and Y are expressed by the following formulas 62 and 63.

Next, a description will be given of whether the amplitude center value can be the optimum correction value of the offset value as in the conventional offset correction in a waveform in which such second harmonics are superimposed. The amplitude center values X _VPP and Y _VPP of the A phase output signal V _A (θ) and the B phase output signal V _B (θ) are expressed by the following equations 64 and 65. In the equations 64 and 65, θmax and θmin indicate the positions (angles) of the maximum value and the minimum value of each output signal.

Here, it is considered that the angles θmax and θmin in the A-phase output signal V _A (θ) are approximately 90 ° and 270 °, respectively, and that the A-phase output signal V _A (θ) is expressed by the following equation 66. Then, V _A (90) and V _A (270) are as shown in the following formulas 67 and 68.

Therefore, the amplitude center value X _VPP is expressed by the following equation 69.

Considering that the angles θmax and θmin in the B-phase output signal V _B (θ) are approximately 0 ° and 180 °, respectively, and that the B-phase output signal V _B (θ) is expressed by the following equation 70. For example, V _B (0) and V _B (180) are represented by the following formulas 71 and 72.

Therefore, the amplitude center value Y _VPP is as shown in _Equation 73 below.

That is, the amplitude center values X _VPP and Y _VPP of the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) are expressed by the following equations 74 and 75.

However, these amplitude center values X _VPP and Y _VPP do not coincide with the optimum correction values X and Y expressed by the above-described equations 62 and 63. Rather, using the amplitude center values X _VPP and Y _VPP as the offset voltage correction value increases the slope of the interpolation accuracy φ on the contrary. Thus, it is understood that the conventional offset correction method is not preferable.

Next, a description will be given of the fact that an optimum correction value can be derived by the correction method according to the present invention described above with a waveform in which such second harmonics are superimposed. In the offset correction according to the present invention, the A phase output signal V _A (θ) and the B phase output signal V _B (θ) are shifted by a certain amount ε from the maximum detection position (angle) θmax to the plus and minus sides 2. An average value of output signals at four positions consisting of two positions θmax + ε, θmax−ε and two positions θmin + ε, θmin−ε shifted by a fixed amount ε from the minimum detection position (angle) θmin to the plus and minus sides. Are the offset voltage correction values X ₄ and Y ₄ . Also in this case, the angles θmax and θmin in the A-phase output signal V _A (θ) are approximately 90 ° and 270 °, respectively, and the A-phase output signal V _A (θ) is expressed by 66. , V _A (90 + ε), V _A (90−ε), V _A (270 + ε), and V _A (270−ε) are as shown in the following formulas 76, 77, 78, and 79.

Accordingly, the correction value X ₄ of the offset voltage is represented as the following equation 80, it is modified it is as the following equation 81.

Also, considering that the angles θmax and θmin in the B-phase output signal V _B (θ) are approximately 0 ° and 180 °, respectively, and that the B-phase output signal V _B (θ) is expressed by the equation 70. For example, V _B (0 + ε), V _B (0−ε), V _B (180 + ε), and V _B (180−ε) are expressed by the following equations 82, 83, 84, and 85.

Accordingly, the correction value Y ₄ of the offset voltage is represented as the following equation 86, it is modified it is as the following equation 87.

That is, the offset voltage correction values X ₄ and Y ₄ of the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) are expressed by the following equations 88 and 89, respectively.

Here, assuming that the constant amount ε to be shifted is 90 °, the correction values X ₄ and Y ₄ expressed by the equations 88 and 89 are as shown in the following equations 90 and 91. The following numbers 92 and 93 are obtained.

These correction values X ₄ and Y ₄ coincide with the optimum correction values X and Y represented by the above-described calculated numbers 62 and 63. That is, in order to obtain the optimum correction value for suppressing the gradient of the interpolation accuracy φ of the output signal in which the correction value of the offset voltage latent in the magnetic sensor 10 itself and the second harmonic due to the installation error are superimposed, The shift amount ε from the maximum value may be set to 90 °. As a result, it is understood that the deterioration of the interpolation accuracy φ can be satisfactorily suppressed by using the offset correction method of the present invention.

d. Modifications The embodiment of the present invention has been described above. However, the present invention is not limited to the above embodiment, and various modifications can be made without departing from the object of the present invention.

(1)
Modification 1
The magnetic sensor 10 described in the above embodiment and the prior art is a point-symmetric magnetic sensor. However, the magnetic sensor 10 is not limited to this, and the present invention is also applicable to various magnetic sensors. For example, as shown in FIG. 22, as the magnetic sensor 10, the magnetoresistive effect elements S1, S2, S3, and S4 as detection elements are arranged in the same direction (that is, the pitch direction) indicated by the illustrated arrows at intervals corresponding to the magnetization pitch. In addition, a magnetic sensor arranged in a line symmetry (axisymmetric with respect to a one-dot chain line in FIG. 22) can be used. In this case, the magnetoresistive effect elements S1, S2, S3, and S4 may be even numbers such as two other than four or eight.

  The magnetic sensor 10 is mainly used for detecting a multipolar magnet. As examples of the structure of the apparatus, suitable arrangements of the magnetic sensor 10 and the magnet 22 are shown in FIGS. 23A and 23B. FIG. 23A is arranged so that the tangential direction of the multipolar ring magnet 22 and the pitch direction of the magnetic sensor 10 (in the direction of the arrow in the figure) coincide, and the normal line of the magnet 22 and the symmetry line of the magnetic sensor 10 coincide. In this case as well, a rotary encoder device that outputs output signals that are 90 degrees out of phase with each other by detecting a magnetic field change only in the tangential direction as the magnet 22 rotates is provided. In FIG. 23B, the magnetic sensor 10 is arranged on the side surface of the multipolar linear magnet 22 so that the pitch direction (arrow direction in the drawing) of the magnet 22 and the magnetic sensor 10 is parallel. By linearly moving the magnet 22, a linear encoder device that outputs output signals that are 90 degrees out of phase with each other by detecting a change in the magnetic field only in that direction is obtained.

  In the encoder apparatus according to these modified examples, if the magnetic sensor 10 is appropriately installed with respect to the magnet 22, unlike the case of the third embodiment, the magnetic sensor 10 detects a magnetic field change in the same direction. The change in the magnetic field of the cardioid curve shown in FIG. 21 is not detected. However, as in the case of the first and second embodiments, as shown in FIGS. 24A and 24B, when the magnetic sensor 10 is arranged with respect to the magnet 22 in a state having an installation error, the second harmonic is output. Superimpose on. That is, as shown in FIG. 24A, in the case of the multipolar ring magnet 22, when there is an installation error between the magnetic sensor 10 symmetry line and the normal line of the magnet 22, the second harmonic is superimposed on the output signal. As shown in FIG. 24B, in the case of the multipolar linear magnet 22, when the magnetic sensor 10 is tilted with respect to the pitch direction of the magnet 22, the second harmonic is superimposed on the output signal.

  FIG. 25A shows the interpolation accuracy φ (= electric angle−mechanical angle) in the encoder device of FIG. 24A. In this case, (a) shows the interpolation accuracy φ when the conventional correction method is used as an error in the graph, and (b) shows the interpolation accuracy φ when the correction method of the present invention is used as an error in the graph. Show. FIG. 25B shows the interpolation accuracy φ (= electric angle−mechanical angle) in the encoder apparatus of FIG. 24B. Also in this case, (a) shows the interpolation accuracy φ when the conventional correction method is used as an error in the graph, and (b) shows the interpolation accuracy φ when the correction method of the present invention is used as an error in the graph. As shown. According to this, also in the encoder device according to these modified examples, by using the correction method according to the present invention, the deterioration of the interpolation accuracy φ (error) is improved compared to the case of using the conventional correction method. It is understood that it can be suppressed.

(2)
Modification 2
In the magnetic sensor 10 according to the above-described embodiment and the modification, a plurality of magnetoresistive effect elements as detection elements are arranged on one substrate. However, in the present invention, one piece is provided on one substrate. The present invention is also applied to an encoder device in which a magnetoresistive effect garment as a detection element is arranged. For example, as an encoder device of this type, as shown in FIG. 26, a plurality of magnetic sensors 10 each including only one magnetoresistive effect element are arranged on the side of the multipolar ring magnet 22, and According to the rotation, an output signal composed of an A-phase output signal V _A (θ) and a B-phase output signal V _B (θ) that are 90 ° out of phase with each other is obtained. In this case as well, a plurality of magnetic sensors 10 each including only one magnetoresistive element are arranged in parallel to the multipolar linear magnet as in FIG. It can also be modified to detect.

(3)
Other Modifications In the magnetic sensor 10 in the encoder device according to the above-described embodiment and the modification, the AMR magnetoresistive effect element is used. However, as described in the prior art, a GMR magnetoresistive effect element is used, An element can also be used.

In addition, a magnetic sensor is employed as a sensor in the encoder device according to the embodiment and the modification. However, as can be seen from the above-described theoretical formula, the phase A output signal V _A (θ) and the phase B output signal V _B (θ) are shifted from each other by 90 ° in accordance with the rotation or movement of the detection target. Since an output signal only needs to be obtained, a sensor other than a magnetic sensor may be used for the sensor unit. For example, the present invention is also applied to an optical encoder device in which a detection target is formed to have a slit and the sensor is an optical probe.

In the embodiment and the modification, both the offset correction and the amplitude correction are performed. However, the deterioration of the interpolation accuracy φ due to the amplitude ratio between the A-phase output signal V _A (θ) and the B-phase output signal V _B (θ) is small as described above. Therefore, when the deterioration of the interpolation accuracy φ related to the amplitude ratio does not matter, the amplitude correction may be omitted and only the offset correction may be performed.

  Furthermore, when the correction value once determined is not appropriate due to environmental changes during operation (for example, when temperature characteristics are added), the correction value once determined may be further corrected according to the environment. According to this, further improvement in accuracy can be expected.

DESCRIPTION OF SYMBOLS 10 ... Magnetic sensor 11, 31 ... Sampling circuit, 12a, 12b, 32a, 32b ... A / D converter, 13, 33 ... Arithmetic processing unit, 16, 34 ... Input / output circuit, 17, 35 ... Memory, 21 ... Rotating shaft, 22 ... magnet, 23 ... drive device

Claims (4)

A sensor unit that outputs sinusoidal first and second analog detection signals having phases different from each other by 90 degrees according to the rotation or movement of the detection target;
The first and second analog detection signals from the sensor unit are digitally converted into first and second digital detection signals, respectively, and the rotation angle of the detection object or the detection object is detected using the converted first and second digital detection signals. It is applied to an encoder device that includes an arithmetic processing unit that performs arithmetic processing on a moving position,
A correction method for an encoder device for correcting the converted first and second digital detection signals,
A data acquisition procedure for acquiring first and second digital data for at least one waveform period by sampling the first and second analog detection signals from the sensor unit at predetermined intervals, respectively.
A position detection procedure for detecting a position of a maximum value and a minimum value from the acquired first and second digital data, respectively;
From the acquired first digital data, each digital data at four positions respectively separated by 90 degrees from the respective detected positions of the maximum value and the minimum value of the detected first digital data is extracted to 4 Four correction digital data, and among the acquired second digital data, four positions separated by 90 degrees from the respective detection positions of the maximum value and the minimum value of the detected second digital data. A correction digital data extraction procedure for extracting each digital data into four correction digital data;
A value obtained by averaging the four correction digital data related to the extracted first digital data is set as a first offset value, and a value obtained by averaging the four correction digital data related to the extracted second digital data is set to the first value. 2 Offset value setting procedure to set as an offset value;
The first and second digital detection signals input from the sensor unit and converted before the calculation processing unit uses the first and second digital detection signals to calculate the rotation angle or movement position of the detection target. And an offset value correction procedure for correcting each of the values using the first and second offset values.   The correction method for an encoder apparatus according to claim 1, wherein the sensor unit includes a plurality of detection elements arranged symmetrically on a single substrate. A sensor unit that outputs sinusoidal first and second analog detection signals having phases different from each other by 90 degrees according to the rotation or movement of the detection target;
The first and second analog detection signals from the sensor unit are digitally converted into first and second digital detection signals, respectively, and the converted first and second digital detection signals are converted into the first and second analog detection signals, respectively. And an arithmetic processing unit that performs arithmetic processing on the rotation angle or movement position of the detection target using the corrected first and second digital detection signals.
Data acquisition means for acquiring digital data for at least one waveform period by sampling the first and second detection signals at predetermined position intervals;
Position detecting means for detecting a position of a maximum value and a minimum value from the acquired first and second digital data, respectively;
From the acquired first digital data, each digital data at four positions respectively separated by 90 degrees from the respective detected positions of the maximum value and the minimum value of the detected first digital data is extracted to 4 Four correction digital data, and among the acquired second digital data, four positions separated by 90 degrees from the respective detection positions of the maximum value and the minimum value of the detected second digital data. Correction digital data extraction means for extracting each digital data to obtain four correction digital data;
A value obtained by averaging the four correction digital data related to the extracted first digital data is set as a first offset value, and a value obtained by averaging the four correction digital data related to the extracted second digital data is set to the first value. An encoder device comprising offset value setting means for setting as two offset values.   The encoder device according to claim 3, wherein the sensor unit includes a plurality of detection elements arranged symmetrically on a single substrate.

Images