JP2000258186A - Self-correction angle detector and method for correcting detection accuracy - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an angle detector requiring no highly precise reference detector and to correct accuracy of angle conveniently with high accuracy. SOLUTION: A common machine shaft is fixed with first and second angle detectors 1, 2 and the fixing position of the second angle detector 2 with respect to the first angle detector 1 is shifted relatively by a specified angle at a time thus determining a plurality of relative fixing angle positions. Angle detection data of the first and second angle detectors 1, 2 is then determined at all rotary positions of the machine shaft for each position thus determined and error value concerning to each angle detection data of the first angle detector 1 is calculated based on the difference of respective angle detection data. Current angle detection data outputted from the first angle detector 1 depending on the current rotary positions of the machine shaft is corrected using the error value determined for the current angle detection data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a self-calibration type angle detection device and a detection accuracy calibration method using the same, and relates to a highly accurate machine axis rotation angle measurement and control in a machine tool, a semiconductor manufacturing device, various precision machines, and the like. The angle accuracy calibration process is performed while being mounted on various industrial machines such as a forging machine, a steel making machine, and the like.
This makes it possible to perform high-precision measurement and control of the machine rotation angle, and is also applicable when measuring and controlling the rotation position angle or speed with ultra-small size, high resolution, and high accuracy in various vehicles or vehicles. In addition, the present invention can be applied to calibration of the accuracy of a position detector other than the rotation angle.

[0002]

2. Description of the Related Art In various precision machines, various industrial machines, and the like, it is widely practiced to detect the rotation angle (ie, rotation position) of a machine shaft by a rotation angle detector (shaft encoder) and use the detected angle for a necessary control. ing. As this kind of rotation angle detector, an optical rotary encoder,
Various types are known, such as a magnetic induction type rotational position detector such as a resolver, or a variable magnetoresistive type rotational position detector based on the same phase detection principle as a resolver. Also in the detector, the detection accuracy is often insufficient due to various factors such as a manufacturing error of each detector, that is, a difference between individual devices, a misalignment when the detector is attached to a mechanical shaft, and a poor coupling. Therefore, when a rotational position detector is used for an application requiring high precision, it is absolutely necessary to calibrate the detection angle accuracy of the detector. Conventional angle accuracy calibration equipment requires a detector that itself requires high accuracy as a reference detector, and the calibration process is performed by attaching a calibration target detector and a reference detector to a dedicated angle indexing mechanism. In most cases, it is difficult to perform calibration with the detector attached to the machine axis at the site to be measured. Also, the structure of the calibration device was complicated. As a conventionally known angle accuracy calibration technique, for example, “Journal of Precision Engineering Society, Vol.
5 (published in 1978), pp. 539-544, a paper "Angular Standard and Calibration Method" (by Akira Toyoyama).

[0003]

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and does not require a high-precision reference detector, and performs angle calibration by a self-calibration method using a calibration target detector itself. It is an object of the present invention to provide a self-calibration-type angle detection device capable of performing a self-calibration, and to use any machine axis (preferably a machine axis to be measured at a site where a calibration object detector is mounted) without using a dedicated angle indexing mechanism. It is an object of the present invention to provide a detection accuracy calibration method that can easily perform calibration by attaching a calibration target detector to the detection accuracy calibration method.

[0004]

A self-calibration type angle detecting device according to the present invention is mounted on a common machine shaft and outputs angle detection data according to a rotation angle of the machine shaft at a predetermined resolution. An angle detector and a second angle detector;
The mounting position of the second angle detector with respect to the machine axis is relatively shifted by a predetermined angle with respect to the mounting position of the first angle detector with respect to the machine axis. And the angle detection data of the first and second angle detectors with respect to each rotational position of the machine shaft are determined for each of the determined relative mounting angle positions, and each angle of the first angle detector is determined. An error calculator for calculating at least an error value for each angle detection data of the first angle detector based on a difference between the two for each detection data; and at least the first angle corresponding to a current rotational position of a detection target machine axis. In accordance with the current angle detection data output from the detector, the error value obtained by the error calculator with respect to the current angle detection data is used, and the current angle is calculated in accordance with the error value. Comprising a calibration unit for calibrating the value of the detected data.

According to the present invention, the first angle detector and the second angle detector may both be calibration target detectors, that is, ordinary angle detectors, and special and dedicated high-precision reference detection. It does not need to be a container. During error data calculation processing for calibration, these first and second angle detectors are attached to a common machine axis. This machine axis is
The angle indexing axis of the special / dedicated angle indexing mechanism does not need to be the axis, and may be any mechanical axis. Most preferably, it is a machine axis at the site where the angle detector to be calibrated is actually attached and used (that is, the machine axis to be detected). In that case, the error data calculation process for calibration can be performed with the angle detector attached to the machine axis at the site (that is, the machine axis to be detected), and then the angle detector is attached. (I.e., without replacing the angle detector), the detection of the rotation angle of the machine shaft to be detected and the control based on the rotation angle can be performed. Since error data calculation processing for calibration is performed in a state where the apparatus is attached to a machine axis at the site of use (that is, a mechanical axis to be detected), the accuracy and reliability of angle accuracy calibration can be improved.

The error data calculation process for calibration is performed by an error calculator. The processing content includes a plurality of types of mounting positions of the second angle detector with respect to the machine axis which are relatively shifted by a predetermined angle with respect to the mounting position of the first angle detector with respect to the machine axis. Relative mounting angle positions are respectively determined, and angle detection data of the first and second angle detectors with respect to all rotational positions of the machine shaft are determined for each of the determined relative mounting angle positions, and the first angle is calculated. At least an error value relating to each angle detection data of the first angle detector is calculated based on a difference between the two for each angle detection data of the detector. The angle detection data of the first and second angle detectors each include a value indicating the true angle θ and an error ε. Since the error ε can have a unique error value depending on the value of each angle θ, this is represented by a function ε (θ). In particular, the error of the first angle detector is expressed by Aε (θ ), And the error of the second angle detector is represented by Bε (θ). Therefore, the difference between the angle detection data of the first and second angle detectors cancels out the value indicating the true angle θ and indicates a value related to the error Aε (θ) and Bε (θ) of both. Become. Therefore, by processing the difference between the angle detection data of the first and second angle detectors using a predetermined algorithm, at least the error value of each angle detection data of the first angle detector, that is, Aε (θ ) Can be calculated. This detailed algorithm example will be described later in an embodiment.

The error value Aε (θ) calculated in this way is stored in a memory or the like, and is used as calibration data when detecting the angle of the machine axis to be detected. That is, in the calibration unit, the error calculation unit obtains the current angle detection data in accordance with at least the current angle detection data output from the first angle detector according to the current rotational position of the detection target machine axis. Using the obtained error value Aε (θ), the value of the current angle detection data is calibrated according to the error value.

In this way, the angle can be calibrated by the self-calibration method using the calibration object detector itself, and any machine axis (preferably, calibration object detection) can be used without using a dedicated angle indexing mechanism. (It may be the machine axis itself to be measured at the site where the device is mounted.) It is possible to easily perform calibration by mounting the detector to be calibrated, and so on.

The present invention can be constructed and implemented not only as a device invention, but also as a method invention.
It can also be implemented. Further, the present invention can be implemented in the form of a computer program, or can be implemented in the form of a recording medium storing such a computer program.

[0010]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 is a diagram showing an example of the structure of an angle detector in a self-calibration type angle detection device according to the present invention, wherein (A) shows first and second angle detectors 1.
And FIG. 2B is a schematic front view of each of the angle detectors 1 and 2. The two angle detectors 1 and 2 are:
An absolute type angle detector (shaft encoder) having the same structure and the same resolution, and is attached to a common mechanical shaft 3. The principle of the structure of the detector is a non-contact type variable magnetoresistive type magnetic induction detector, and the position detection principle employs a phase measurement method according to the resolver principle as disclosed in, for example, JP-A-9-126809. It becomes.

An example of the first angle detector 1 will be described. The primary coil PW and the secondary coils SW1, SW
2, SW3, and SW4, and a rotor 1R formed of a ferromagnetic material such as iron in a predetermined shape (for example, a substantially heart shape as shown in FIG. 1B). It is composed of The material of the rotor 1R is not limited to a magnetic material such as iron, but may be a conductor such as copper or aluminum. In short, the material responds to magnetism and changes the induction coefficient for the coil. What kind of material and / or shape to use is a well-known technique. The secondary coils SW1, SW2, SW3, and SW4 are respectively wound around four-pole magnetic cores arranged at intervals of approximately 90 degrees in the circumferential direction on the substrate of the stator 1S. The end of the pole is non-contactly opposed to one surface of the rotor 1R via an air gap. Due to the substantially heart-shaped shape of the rotor 1R, the end of each secondary coil pole and the rotor 1R are adjusted in accordance with the relative rotation angle θ of the rotor 1R with respect to the stator 1S (the angle in the circumferential direction in FIG. 1B). And the area of the surface opposite to changes. In the stator 1S, the primary coil PW is wound along the outermost periphery, and its coil length sufficiently covers the locations of the secondary coils SW1 to SW4 and the rotor 1R as shown in FIG. It is a length that can be done. As a result, the AC magnetic field generated by the primary coil PW is
~ Sufficiently affect the location of SW4 and rotor 1R,
An effective induced voltage sufficiently reflecting the change in the area of the air gap where the end of each secondary coil pole faces the rotor 1R can be induced in each of the secondary coils SW1 to SW4.

Thus, the rotor 1 with respect to the stator 1S
In accordance with the relative rotation angle θ of R, the induced voltage level of the secondary coil SW1 shows a function characteristic of sin θ, and the induced voltage level of the secondary coil SW2 shifted by 90 degrees is co.
sθ, the induced voltage level of the secondary coil SW3 shifted by 90 degrees therefrom shows the function characteristic of −sin θ, and the induced voltage level of the secondary coil SW4 shifted by 90 degrees is −cos θ. It may be indicative of a function characteristic. Primary coil PW and secondary coils SW1 to SW
4 is as shown in FIG. 2 and the AC component is sinωt
, Sin θ · sin ωt can be obtained as a differential output of the secondary coils SW1 and SW3, and the secondary coil S
Cos θ · sin ωt can be obtained as a differential output between W2 and SW4.

Thus, the rotor 1 with respect to the stator 1S
A first output AC signal = sinθ · sinωt having a first function value sinθ corresponding to the rotation angle θ of R as an amplitude value, and a second function value cosθ corresponding to the same rotation angle θ as an amplitude value. 2, the output AC signal = cos θ · sin ωt. According to such a coil configuration, two output AC signals (a sine output and a cosine output) having an in-phase AC and a two-phase amplitude function similar to those obtained in a resolver conventionally known as a rotary position detecting device. Output) can be obtained from the secondary coil.

The two-phase output AC signals (sin θ · sin ωt and cos θ · sin ωt) output from the secondary coil can be used in the same manner as the output of a conventionally known resolver. For example, a two-phase output AC signal (sin
θ · sinωt and cosθ · sinωt) to an appropriate digital phase detection circuit, detects the sine function sinθ and the phase value θ of the cosine function cosθ by a digital phase detection method, and obtains the absolute value of the rotation angle θ at a predetermined resolution The digital data represented by the following can be obtained. As a digital phase detection method adopted in this digital phase detection circuit, a known RD (resolver-digital) converter may be applied.
A phase measurement method as disclosed in Japanese Patent No. 26809 may be employed.

In a phase measurement method as disclosed in Japanese Patent Application Laid-Open No. 9-126809, roughly speaking, one output AC signal sinθ · sinωt is electrically phase-shifted by 90 degrees to generate sinθ · cosωt. , And the other output AC signal cos θ · sin ωt are added or subtracted and combined to obtain a signal such as sin (ωt + θ) or sin (ωt−θ).
An output AC signal that includes the rotation angle θ as an initial phase component of the electrical phase angle is formed, and the electrical initial phase component θ in the output AC signal is digitally counted, so that the absolute value of the rotation angle θ can be determined by a predetermined resolution. The digital data represented by is obtained. Therefore, 1
The rotation angle θ of the rotation (in the range of 0 ° to 360 °) can be detected with extremely high resolution and high response. Further, the mechanical structure of the angle detector 1 including the stator 1S and the rotor 1R can be extremely reduced in size. As an example, the outer dimensions of the angle detector 1 composed of the stator 1S and the rotor 1R can be reduced to a few centimeters in diameter and several centimeters in the axial direction, and the absolute detection resolution per rotation is tens of thousands. Devices that can achieve high resolution such as division or hundreds of thousands of divisions have been manufactured.

The second angle detector 2 also has a stator 2
S and the rotor 2R, and may have the same structure as the first angle detector 1 described above.

FIG. 3 is a block diagram showing the overall configuration of the self-calibration type angle detecting device according to the present invention. The converter 1D is an electrical converter for the first angle detector 1, counts the high-speed clock pulse CP by the binary counter 11, and outputs the sine wave generator 1 based on the count output.
2, a sine wave signal sinωt is generated, converted into an analog waveform signal by the D / A converter 13, and applied to the primary coil PW of the first angle detector 1 as an excitation AC signal. One cycle of the count value of the counter 11 from 0 to the maximum value corresponds to one cycle of the excitation AC signal sinωt. Each secondary coil SW1 to SW4 of the first angle detector 1
2-phase output AC signal sinθ · sinωt output from
And cos θ · sin ωt are input via the input buffer amplifier 14, and the component of the rotation angle θ is converted from the amplitude component to the electrical phase angle by the electrical arithmetic processing in the amplitude / phase converter 15 as described above. (Ωt + θ) and / or sin (ωt−θ). This signal is input to the zero-cross comparator 16 to generate a zero-cross detection pulse at a timing corresponding to the initial phase θ, whereby the count value of the counter 11 at this time is latched by the latch circuit 17. That is, since the count value of the counter 11 starts counting from 0 when the instantaneous phase of sinωt is 0, sin (ωt + θ) or sin (ωt−
The count value at the zero-cross detection timing of θ) indicates an absolute value corresponding to the angle θ. The output of the latch circuit 17 is input to the arithmetic unit 18 and required arithmetic processing (for example, arithmetic processing as disclosed in Japanese Patent Application Laid-Open No. 9-126809) is performed as necessary. Then, the angle detection data by the first angle detector 1 (this is SA (θ)
Bus control and output buffer 1
9 to the data bus 20 of the host computer.

The converter 2D is an electrical converter for the second angle detector 2, and has the same configuration as the converter 1D. Therefore, detailed description is omitted. Thus, from the bus control and output buffer of the conversion unit 2D,
Angle detection data by the second angle detector 2 (this is referred to as SB
(Indicated by (θ)) is output to the data bus 20 of the host computer.

The angle detection data SA (θ) obtained by the first angle detector 1 and the angle detection data SB (θ) obtained by the second angle detector 2 are determined by various causes such as machine differences at the time of manufacture of each individual. It contains an error. Various processes for detecting this error and performing calibration are performed by the CPU 21 and the RA.
It is executed by a host computer including the M22, the ROM 23, the decoder 24 and the like. A processing program for calculating and calibrating the error is stored in the RAM 22 or the ROM 23 and is executed by the CPU 21. Also, R
The AM 22 is used as a temporary memory for various data and as a working memory. ROM23 is EEPRO
A writable ROM such as M is included, and the calculated error data is written here and stored in a nonvolatile manner.
The decoder 24 decodes an address given to the address bus 25 and an instruction given from the CPU 21. The output latch circuit 26 latches and outputs the calibrated angle data by the computer.

Next, the principle of the angle accuracy calibration method according to the present invention will be described. Both of the two angle detectors 1 and 2 used can be calibration target detectors. Two angle detectors 1 and 2 as calibration target detectors can be used without using a special / dedicated reference detector. In this method, the self-calibration type angular accuracy calibration using the calibration target detector itself is performed according to the angle accuracy calibration method according to the present invention.

An error included in the angle detection data SA (θ) by the first angle detector 1 is represented by Aε (θ). This error is not uniform, and the angular position θ (0 degree to 3 degrees)
The error is shown as Aε (θ) as a function of the angular position θ to indicate a specific value for each of the variable (60 degrees or a variable in the range of 0 to 2π radians). Similarly, an error included in the angle detection data SB (θ) by the second angle detector 2 is represented by Bε (θ). These errors Aε
(Θ) and Bε (θ) are unknown values, and are known values.
(Θ) and SB (θ), these errors Aε
(Θ) and Bε (θ) are calculated and used as calibration data. These errors Aε (θ) and Bε (θ)
Is a function of the angular position θ, and is also referred to as an angular position error.

For easy understanding, an example of each angular position error Aε (θ) and Bε (θ) is shown in a graph. FIG. 4A is a waveform diagram illustrating an example of an error Aε (θ) included in the angle detection data SA (θ) by the first angle detector 1, and FIG. 4B is a waveform diagram illustrating the second angle detector 2. FIG. 9 is a waveform diagram showing an example of an error Bε (θ) included in angle detection data SB (θ) based on the above. In the figure, the horizontal axis is the angular position θ, which corresponds to the mechanical angle.

For convenience, the angle detection data SA (θ) and S
If the data corresponding to the true angular position θ included in B (θ) is directly expressed using θ, SA (θ) and SB
(Θ) can be expressed as in the following equation (1). SA (θ) = θ + Aε (θ) SB (θ) = θ + Bε (θ) Equation (1)

Here, when the difference between the two is obtained, δ (θ) = SA (θ) −SB (θ) = Aε (θ) −Bε (θ) (2) It can be seen that they are canceled out and only the error component is obtained. This δ (θ) will be referred to as “relative angular position error”.

The above equation (1) is for the case where the phases of the outputs (SA (θ) and SB (θ)) of the first and second angle detectors 1 and 2 with respect to the rotation angle θ are the same. When this phase is shifted by φ degrees, SA (θ) = θ + Aε (θ) SB (θ + φ) = θ + φ + Bε (θ + φ) Equation (3) can be expressed.

The above equation (3) means that when the first angle detector 1 outputs the angle detection data SA (θ) corresponding to the angle θ, the second angle detector 2 Angle detection data S corresponding to angle θ + φ out of phase by φ degrees
B (θ + φ).

Here, assuming that φ is a known value, the phase of the outputs (SA (θ) and SB (θ)) of the first and second angle detectors 1 and 2 with respect to the rotation angle θ is intentionally set to φ. It can be set to be shifted. Therefore, SB (θ +
It is easy to subtract φ from φ) and deform as follows. SB (θ + φ) −φ = θ + Bε (θ + φ) Equation (4)

Thus, it is possible to easily obtain a "relative angular position error" which is obtained by extracting an error component of the angle detection data of the two detectors 1 and 2 when the phase is shifted by φ. If this is represented by δ (θ, φ), δ (θ, φ) = SA (θ) − {SB (θ + φ) −φ} = θ + Aε (θ) − {θ + Bε (θ + φ)} = Aε (θ) − Bε (θ + φ) ... Expression (5) Here, since the solution when φ = 0 corresponds to δ (θ) in the above equation (2), equation (5) will be used as a general equation including the above equation (2).

As described above, for a plurality of different values of φ, the angle detection data SA (θ) and SB (θ) (that is, SB (θ + φ)) according to the above equation (3) are obtained, and the equations (4) and (4) are obtained. According to equation (5), several “relative angular position errors” δ (θ, φ) corresponding to each φ can be obtained. In order to obtain the “relative angular position error” δ (θ, φ) corresponding to several values of φ in this manner, the angular position error Aε (θ) is calculated by performing an arithmetic processing on these in accordance with a predetermined calibration algorithm. Alternatively, Bε (θ) can be extracted, which is extremely effective.

In this case, when the range of 360 degrees per rotation is equally divided by an arbitrary number M and the value of the phase shift φ of the two detectors is determined, the angular position error Aε (θ) or Bε (θ) is extracted. Was found to be effective. For example, if M = 4, φ
= K (360 ° / 4) (however, k is an ordinal number of 0, 1, 2, 3). In this case, by calculating a simple average of the “relative angular position error” δ (θ, φ) corresponding to several values of φ, the angular position error Aε (θ) or Bε
It was found that (θ) could be extracted. Such a new algorithm for the self-calibration process will be named “equal averaging method”.

By shifting the mounting position of the second angle detector 2 with respect to the machine shaft 3 by a desired φ in mechanical angle with respect to the mounting position of the first angle detector 1 with respect to the machine shaft 3, two detections can be performed. Since the angle detection data SA (θ) and SB (θ) of the devices 1 and 2 can have a phase shift φ (that is, a relationship of SB (θ + φ) can be obtained), the first angle detector 1 The mounting position of the second angle detector 2 may be determined so that the relative angle between the mounting position of the second angle detector 2 and the mounting position of the second angle detector 2 forms the required phase angle φ.

This determination is basically performed as follows. First, the first angle detector 1 is attached to the machine shaft 3. When installing, use the angle detector 1
This may be appropriately performed without making the origin (sensor origin) coincide with the origin of the mechanical shaft 3. However, when the mechanical axis 3 is the target mechanical axis to be detected in an industrial machine or the like to which the angle detector 1 is applied, the origin (sensor origin) of the angle detector 1 and the origin of the mechanical axis 3 are used. It is a matter of course that if the mounting is carried out so that the values substantially coincide with each other, the later origin adjustment processing becomes easy.

Next, as schematically shown in FIG. 1A, the second angle detector 2 is mounted on the same mechanical shaft 3.
The angle detector 2 is attached so that the origin (sensor origin) of the angle detector 2 is shifted from the origin (sensor origin) of the first angle detector 1 by a desired angle φ. This is performed by determining the mounting position of No. 2. This indexing (that is, mounting the angle detector 2 at a predetermined mounting position) is simplified by mounting the second angle detector 2 on the machine shaft 3 while manually adjusting the angle to be shifted by the desired angle φ. Can be performed. By determining the mounting position in this manner, the angle detection data SA (θ) output by the first angle detector 1 and the second angle detector 2 are output for the same rotation angle of the mechanical shaft 3.
A phase shift corresponding to the mechanical angle φ is produced between the angle detection data SB (θ) and the angle detection data SB (θ). That is, the angle detection data S output from the first angle detector 1
If A (θ) is expressed as “SA (θ)” as it is,
Angle detection data SB output from second angle detector 2
(Θ) is represented by “SB (θ + φ)”. in this case,
The variable θ in “SA (θ)” corresponds to the mechanical angle θ when the origin (sensor origin) of the first angle detector 1 is 0, and the variable θ + φ in “SB (θ + φ)” is the mechanical angle. It corresponds to the mechanical angle θ + φ obtained by adding the initial phase value φ to θ. That is, for example, when the first angle detector 1 outputs the angle detection data “SA (0)” corresponding to the sensor origin θ = 0 corresponding to a certain rotational position of the mechanical shaft 3,
The angle detector 2 outputs angle detection data “SB (φ)” corresponding to a position shifted by a mechanical angle φ.

As an example of the case of performing the self-calibration process according to the above “equal division averaging method”, if the number of equal divisions is M = 4,
φ = k (360 ° / 4) (where k = 0, 1, 2, 3
From the ordinal number), the displacement of the mounting positions of the two detectors 1 and 2 is φ =
For each of the four index positions of 0 °, φ = 90 °, φ = 180 °, and φ = 270 °, the relative positions of the angle detectors 1 and 2 are determined.

The procedure of the self-calibration process according to the "equal averaging method" will be described below by taking the case of M = 4 as an example. Note that the self-calibration processing according to the “equal division averaging method”
The processing is performed by the computer processing in. FIG.
FIG. 7 is a flowchart schematically showing an example of an “error calculation program” executed by a computer at that time. In the "error calculation program", the value of the number of equal divisions M can be appropriately selected and set. When M = 4, M is set to M = 4 in the first step S1. Next step S
In 2, the ordinal number k is set to an initial value 0.

Next, in step S3, φ = k (360
Instruct the relative position of the angle detectors 1 and 2 to be determined at the relative phase angle φ of (° / M), and wait until the indexing operation is completed. This instruction is given via a display (not shown) of the computer. In response to this instruction, the indexing operation is performed manually by a human. In determining the relative mounting positions of the angle detectors 1 and 2, the first angle detector 1 is used as described above.
The mounting position of the second angle detector 2 may be fixed at a predetermined position, and only the mounting position of the second angle detector 2 may be changed according to a desired indexing position. (Of course, not limited to this,
The index corresponding to the desired phase angle φ can be determined by changing the mounting positions of the angle detectors 1 and 2, respectively, but this is merely a detour. )

In this example, the first indexing position corresponding to the ordinal number k = 0 is φ = 0 °, and the origin (sensor origin) of the first angle detector 1 and the origin of the second angle detector 2 The second angle detector 2 is located at a position where the origin (sensor origin) matches.
Attach (determine). When the indexing operation is completed, the computer process proceeds to the next step by operating a switch (not shown) or the like.

In the next step S4, while rotating the mechanical shaft 3, the angle detection data SA by the first and second angle detectors 1 and 2 over the entire rotation position (0 to 360 degrees) of the mechanical shaft 3 (Θ) and SB (θ + φ) are collected. The collected data SA (θ) and SB (θ + φ)
The two values collected at the same time are associated with each other on a one-to-one basis (as a data pair) and stored in a buffer in the RAM 22. For example, the value of data SB (θ + φ) obtained simultaneously at that time may be stored in a buffer using the value of data SA (θ) at a certain moment as an address designation index. For example, the detection resolution of the angle detector 1 is one rotation (0 degree to 360 degrees).
Assuming that “2 to the 16th power” = 65536 divisions per degree), the value of data SA (θ) is used as an addressing index in the buffer area of a total of 65536 addresses from 0 to 65535, and the data at that time is used. The value of SB (θ + φ) is stored.

In the next step S5, the angle detection data SA (θ) and SB (θ +
φ), based on the operation according to equation (5),
The relative angular position error δk (θ, φ) is obtained. The obtained relative angular position error δk (θ, φ) is temporarily stored in the RAM 22. Here, δk (θ, φ) indicates the relative angular position error δ (θ, φ) obtained corresponding to the k-th index position. In the case of this example, when k = 0, φ = 0 °
Therefore, it can be expressed as δk (θ, φ) = δ ₀ (θ, 0), and the specific relative angular position error δ (θ, φ) is δ ₀ (θ, 0) = Aε (θ ) −Bε (θ) Expression (6) In the example shown in FIG. 4, this is as shown in FIG. 4C, as shown in (c), each angular position θ from 0 ° to 360 ° from the error value Aε (θ) corresponding to each angular position θ from 0 ° to 360 °. Is obtained by subtracting the error value Bε (θ) corresponding to.
That is, the error value Aε (θ) is obtained at each angular position θ where the phases match.
From the error value Bε (θ).

In the next step S6, it is checked whether ordinal number k has reached the final value "M-1". If NO, the process goes to step S7 to increase the value of ordinal number k by one. then,
Returning to step S3, it is instructed to determine the relative mounting position of the angle detectors 1 and 2 with the relative phase angle φ of “φ = k (360 ° / M)” according to the value of the ordinal number k increased by one. Wait until the indexing operation is completed.

In the case of this example, the second indexing position corresponding to the next ordinal number k = 1 is φ = 90 °, and the second angle with respect to the origin of the first angle detector 1 (sensor origin). The second angle detector 2 is attached (determined) at a position where the origin (sensor origin) of the detector 2 is shifted by 90 degrees in mechanical angle. That is, when the first angle detector 1 outputs angle detection data corresponding to a mechanical angle of 0 ° in the detector 1, the second angle detector
Is set such that the angle detector 2 outputs angle detection data corresponding to a mechanical angle of 90 ° in the detector 2. Then, the same arithmetic processing as in steps S4 and S5 is performed, and the mounting angle difference φ
A relative angular position error δk (θ, φ) corresponding to 90 ° is obtained. Here, since k = 1 and φ = 90 °, δk
(Θ, φ) = δ ₁ (θ, 90 °), and a specific relative angular position error δ (θ, φ) is given by δ ₁ (θ, 90 °) = Aε (θ) −Bε (Θ + 90 °) Expression (7). In the example shown in FIG. 4, as shown in FIG. 4, this is based on the error value Aε (θ) corresponding to each angular position θ from 0 to 360 degrees, and each angle for one rotation starting from 90 degrees. This corresponds to a value obtained by subtracting the error value Bε (θ + 90 °) corresponding to the position θ + 90 °. That is, at each angular position θ and θ + 90 ° having a phase shift of 90 degrees, the error value Aε (θ) is subtracted from the error value Bε
(Θ + 90 °) is subtracted.

Next, in step S7, the value of the ordinal number k is further increased by 1, and the process returns to step S3, where the relative phase angle φ of “φ = k (360 ° / M)” corresponding to the increased ordinal number k is incremented by one.
Instructs to determine the relative mounting positions of the angle detectors 1 and 2, and waits until the indexing operation is completed. This time, the second indexing position corresponding to the ordinal k = 2 is φ =
180 °, and the origin (sensor origin) of the second angle detector 2 with respect to the origin (sensor origin) of the first angle detector 1
Is located at a position shifted by 180 degrees in mechanical angle.
Attach (determine). Then, the step S
The same arithmetic processing as in steps S4 and S5 is performed.
The relative angular position error δk (θ, φ) corresponding to the mounting angle difference φ = 180 ° is determined. Here, k = 2, φ = 180
Therefore, δk (θ, φ) = δ ₂ (θ, 180 °), and the relative angular position error δ (θ, φ) is δ ₂ (θ, 180 °) = Aε (θ) −Bε (Θ + 180 °) Expression (8). In the example shown in FIG. 4, as shown in FIG. 4, this is based on the error value Aε (θ) corresponding to each angular position θ from 0 ° to 360 °, and each angle for one rotation starting from 180 °. This corresponds to a value obtained by subtracting the error value Bε (θ + 180 °) corresponding to the position θ + 180 °. That is, the error value Bε (θ + 180 °) is subtracted from the error value Aε (θ) at each angular position θ and θ + 180 ° that are 180 ° out of phase.

Next, in step S7, the value of the ordinal k is further increased by 1, and the process returns to step S3, where the relative phase angle φ of “φ = k (360 ° / M)” corresponding to the increased ordinal k is incremented by 1.
Instructs to determine the relative mounting positions of the angle detectors 1 and 2, and waits until the indexing operation is completed. This time, the second indexing position corresponding to the ordinal number k = 3 is φ =
270 °, and the origin of the second angle detector 2 (sensor origin) with respect to the origin of the first angle detector 1 (sensor origin)
Is located at a position shifted by 270 degrees in mechanical angle.
Attach (determine). Then, the step S
The same arithmetic processing as in steps S4 and S5 is performed.
The relative angular position error δk (θ, φ) corresponding to the mounting angle difference φ = 270 ° is determined. Here, k = 3, φ = 270
Since a °, δk (θ, φ) = δ 3 (θ, 270 °) is, the relative angular position error [delta] (theta, phi) _{is, δ 3 (θ, 270 °} ) = Aε (θ) -Bε (Θ + 270 °) Expression (9). In the example shown in FIG. 4, as shown in FIG. 4, each error angle Aε (θ) corresponding to each angular position θ from 0 degrees to 360 degrees is obtained from each angle for one rotation starting from 270 degrees. This corresponds to a value obtained by subtracting the error value Bε (θ + 270 °) corresponding to the position θ + 270 °. That is, the error value Bε (θ + 270 °) is subtracted from the error value Aε (θ) at each angular position θ and θ + 270 ° that are out of phase by 270 degrees.

In the case of M = 4, if ordinal number k = 3, step S6 becomes YES and the process goes to step S8. In step S8, each index position (k = 0, 1, 2,... M−
The relative angular position error δk (θ, θ,
The simple arithmetic average for each θ in φ) is determined as follows.

(Equation 1) Here, it is the second term of the numerator on the right side of the above equation (10).

(Equation 2) Can be set to “0”, the right side of the above equation (10) becomes M · Aε (θ) / M = Aε (θ).

Therefore, by calculating the simple arithmetic average of each relative angular position error δk (θ, φ) for each θ as in the above equation (10), the angular position error Aε of the first angle detector 1 is obtained.
An error function corresponding to (θ) can be obtained. The obtained angular position error Aε (θ) is written to an appropriate memory, for example, an EEPROM in the ROM 23, and is stored in a nonvolatile manner. This error Aε (θ) is calculated by the first angle detector 1
Can be used as calibration data for calibrating an error in the angle detection data SA (θ). If the first angle detector 1 is used as the angle detector of the mechanical shaft 3 to be detected, only the error Aε (θ) for the angle detector 1 needs to be obtained as the calibration data. However, if the angular position error Bε (θ) of the other angle detector 2 is also desired to be obtained, the Aε (θ) obtained as described above is calculated using the above equation (10) or (6), (6). By substituting in 7), (8), etc., the error Bε (θ) of the angle detector 2 can be easily obtained.

Next, the second numerator of the numerator on the right side of the above equation (10)
The term, that is, the reason why equation (11) can be set to “0” will be described. In conclusion, this can be regarded as substantially "0" depending on how the number of equal divisions M is selected. M =
In the example of Equation ( ₄ ), the above equation (11) is defined as ξ ₄ (θ), and when this is expressed by a Fourier series, the angular position error Bε (θ) becomes

(Equation 3) From the above equation (11), that is, ξ ₄ (θ)
Can be expressed as the following equation (12).

[0047]

(Equation 4) Becomes Here, n is the harmonic order, and N is the angle detector 2
, And n = 1, 2, 3,..., N / 2.

In the above equation (12), the value of each term of the series for each value (1 to N / 2) of the degree n is
When the value of cos (n · 90 °) or cos (n · 45 °) becomes “0”, the solution of the term becomes 0, so that n = 1, 3,
.., Cos (n · 90 °) is 0 when odd-ordered, and when n = 2, 6, 10, 14,. (N · 45 °) is 0. That is, when M = 4, when the order n is not a multiple of 4, the solution of the term corresponding to the order n in Expression (12) becomes “0”, which can be ignored. Then, in the above equation (11), only terms of a multiple of 4 are left. That is, only the error component of the multiple wave of 4 affects the solution of the above equation (11). Here, as the mechanical structure of the angle detector 2, it is relatively easy to create a mechanical structure in which the error component of the multiple wave of 4 does not appear or is negligibly small. By using the angle detector 2 that does not produce an error component or is negligibly small, the solution of the above equation (11) can be substantially set to “0”. In theory, even if some higher-order error components remain, the level is often small enough to be practically negligible, so that the solution of the above equation (11) is substantially changed to “0”.
In many cases, you can leave it alone.

For example, if three angle detector poles are arranged at intervals of 120 degrees per circumference and calibration is performed to perform averaging processing, the error component of a multiple wave of 4 can be extremely reduced. It has been known. Therefore, when M = 4, the angle detector 2 having such a three-pole configuration (or a six-pole configuration or the like) may be employed. On the other hand, FIG.
In a quadrupole angle detector at 90-degree intervals as shown in (b), the solution of the above equation (11) is substantially reduced by setting the number of equal divisions M to 9 or 11 or 13 or the like. It has been confirmed that the "." Therefore, it is sufficient to do so. In addition, by designing the pole arrangement of the angle detector as appropriate and applying appropriate averaging processing, etc., a detector with very few error components such as a multiple wave of 3 and a multiple wave of 4 is configured. Therefore, the solution of the above equation (11) can be substantially set to “0” both theoretically and practically.

As described above, according to the “equal division averaging method”, each relative angular position error δk
By obtaining the simple arithmetic average of (θ, φ) for each θ, an error function corresponding to the angular position error Aε (θ) of the first angle detector 1 can be easily obtained. When the data of the angular position error Aε (θ) is stored in the memory as the angle accuracy calibration data of the first angle detector 1, the angle detection data of the first angle detector 1 corresponding to the angular position θ Using SA (θ) as an addressing index, the data of the angular position error Aε (θ) for angle accuracy calibration is stored corresponding to (paired with) the angle detection data SA (θ).

In operation, the first angle detector 1 is used as an angle detector for the mechanical axis to be detected (for example, the mechanical axis 3), and the current rotation of the mechanical axis to be detected is obtained from the conversion unit 1D in FIG. It outputs angle detection data SA (θ) corresponding to the position. FIG. 6 schematically shows an example of the “calibration program” when the computer executes the angle detection operation. In the computer, the current angle detection data SA of the mechanical axis to be detected given from the conversion unit 1D
(Θ) via the data bus 20 (step S
11) Using the captured current angle detection data SA (θ) as an addressing index, a specific current angle detection data SA (θ) is stored in a memory storing data Aε (θ) for angle accuracy calibration. Is read out, and a predetermined calibration operation, typically, “SA (θ) −Aε (θ)” is performed based on the data Aε (θ) (step S12). In this way, calibrated angle detection data is obtained, and this is latched by the latch circuit 26 (FIG. 3).
(Step S13). In FIG.
The routine of steps S11 to S13 is always performed, and the calibration process is constantly performed.

FIG. 7 shows the first and second angle detectors 1, 2
It is a figure which shows another Example of this. In this example,
The stator 2S of the second angle detector 2 is fixed in a predetermined positional relationship with respect to the stator 1S of the first angle detector 1, and the second angle detector 2S corresponding to the required phase angle φ The attachment position is determined by moving only the attachment position of the rotor 2R of the second angle detector 2.
Also in this case, the angle accuracy calibration processing can be performed in the same manner as in the above embodiment.

In each of the above embodiments, the data Aε (θ) for angle accuracy calibration has the same resolution as the detection resolution of the angle detector 1, that is, one-to-one for each value of the angle detection data SA (θ). , Does not necessarily have to be calculated. That is, one piece of data Aε (θ) for angle accuracy calibration is representatively obtained for each group (angle range) of several values of the angle detection data SA (θ), and this is used as an angle accuracy calibration common to the group. May be used as the data for use.

The structures and the detection resolutions of the two angle detectors 1 and 2 to be used may not be the same. For example,
As the angle detector 1, a rotor structure as shown in FIG. 1B (which provides one cycle of magnetoresistance change per rotation of 0 to 360 degrees) is used, while a multi-tooth structure is used as the angle detector 2. (Corresponding concave and convex teeth are also provided on a single coil core on the stator side), and a plurality of cycles of magnetoresistance change may be provided per 0 to 360 degrees per rotation. In that case, the absolute detectable range of one angle detector 1 is 0 degree to 360 degrees per rotation, but the absolute detectable range of the other angle detector 2 is 1 / n rotation (n is the number of teeth). Become. In that case, the output data of the second angle detector 2 can be expressed as “SB (nθ)”. That is, the detection data is obtained at an angle value nθ that is n times the actual mechanical angle θ, and the resolution of the detection data can be increased. At the time of calibration data collection, error data Aε (θ) and Bε (nθ) of both detectors 1 and 2 are obtained based on the outputs of the respective detectors 1 and 2 according to the above embodiment. At the time of normal measurement, the output data SA (θ) and SB (nθ) of the two detectors 1 and 2 are used as error data A
Calibration may be performed using ε (θ) and Bε (nθ), and a combination of both calibrated data may be used. That is, 1 / n
An angle in a narrow range of rotation is absolutely detected with high resolution based on the output of the detector 2, and an absolute value exceeding the angle is detected based on the output of the detector 1.

Further, in each of the above embodiments, the structure of the angle detector used and its detection principle are not limited to those shown in the above embodiments, and any type may be used.
For example, an optical encoder can be used even if the detection resolution is lower than that of the detector shown in the above embodiment. Also, a plurality of primary coils are excited by a two-phase (multiple-phase) AC signal such as sinωt and cosωt, and an output AC signal sin (ωt + θ) indicating an electrical phase angle θ corresponding to the rotation angle θ of the rotor is generated. Then, a known inductive rotational position detector of the type that digitally counts the phase angle θ can be used. In addition, primary coil and 2
It is also possible to use an induction-type rotational position detector that detects the rotational angle θ of the rotor by measuring the change in self-inductance of the exciting coil without having the secondary coil.

As an alternative embodiment of the present invention, the first detector 1 and the second detector 2 are not physically mounted on the exact same mechanical shaft, but are mounted on a mechanical shaft which can be regarded as substantially common. You may do so. For example, the detector 1 may be mounted on the mechanical shaft 3 and the detector 2 may be mounted on another mechanical shaft that moves in conjunction with the mechanical shaft 3. Further, the self-calibration method of the present invention is not limited to the rotation angle detector, but can be similarly applied to a linear position detector. In this case, the detectable range length L of the linear position detector is changed by one rotation (0 degree to 36 degrees).
0 degree), the same calibration method as in the above embodiment can be applied.

[0057]

As described above, according to the present invention, the angle calibration or the calibration of the detection data can be performed by the self-calibration method by the calibration target detector itself, so that it is very simple and a dedicated angle indexing mechanism is provided. The calibration can be performed easily by attaching the calibration target detector to any machine axis without using the sensor, and the detector is mounted on the machine axis itself to be measured at the site where the calibration target detector is mounted And an excellent effect such as that the accuracy calibration process can be performed. Furthermore, by employing the “equal division averaging method”, the calculation algorithm for calibration is extremely simple, and accuracy calibration processing of angle or position detection data can be easily performed.

[Brief description of the drawings]

FIG. 1 is a schematic axial sectional view and a schematic front view showing an embodiment of a self-calibration type angle detecting device according to the present invention.

FIG. 2 is a diagram showing a connection example of primary and secondary coils in one angle detector in FIG. 1;

FIG. 3 is a block diagram schematically illustrating an overall configuration of an embodiment of a self-calibration type angle detection device according to the present invention.

FIG. 4 is a graph illustrating an angular position error in two angle detectors in FIG. 1;

FIG. 5 is a flowchart showing an example of an error calculation program executed by the computer in FIG. 3;

FIG. 6 is a flowchart showing an example of a calibration program executed by the computer in FIG. 3;

FIG. 7 is a schematic axial sectional view showing another arrangement example of two angle detectors in the self-calibration type angle detection device according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 1st angle detector 2 2nd angle detector 1S, 2S Stator 1R, 2R Rotor PW Primary coil SW1-SW4 Secondary coil 1D, 2D conversion part 21 CPU 22 RAM 23 ROM

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Hiroshi Sakamoto 896-8 Yamada, Kawagoe-shi, Saitama (72) Inventor Tadashi Masuda 1-14-5 Hagoromocho, Tachikawa-shi, Tokyo F-term (reference) 2F063 AA35 BC04 BD16 CB19 CC05 DA24 DD03 EA03 GA22 GA29 GA30 GA36 GA53 KA01 KA03 LA03 LA16 LA19 LA20 LA22 LA23 LA29 2F077 AA20 FF04 FF34 TT21 TT26 TT38 TT42 TT59 TT66 TT75 UU09

Claims (10)

[Claims] A first angle detector and a second angle detector which are attached to a common machine shaft and output angle detection data according to a rotation angle of the machine shaft at a predetermined resolution; and The mounting position of the second angle detector with respect to the machine axis is relatively shifted by a predetermined angle with respect to the mounting position of the first angle detector with respect to. Determining the angle detection data of the first and second angle detectors for each rotational position of the machine shaft for each of the determined relative mounting angle positions, and for each angle detection data of the first angle detector. An error calculator for calculating at least an error value for each angle detection data of the first angle detector based on a difference between the first and second angle detectors; and at least the first angle according to the current rotational position of the detection target machine axis. In accordance with the current angle detection data output from the output unit, the error value obtained by the error calculation unit for the current angle detection data is used, and the value of the current angle detection data is calibrated in accordance with the error value. A self-calibration type angle detection device comprising a calibration unit that performs calibration. 2. The first and second angle detectors are each energized by a predetermined AC signal to generate an inductive output signal, and an induction coefficient in the coil is rotated relative to the coil. A magnetic response unit that changes according to the position,
A relative rotation position between the coil unit and the magnetic response unit changes according to a change in the rotation position of the machine shaft, and an induction output signal corresponding to the rotation position of the machine shaft is generated in the coil unit. The self-calibration type angle detection device according to claim 1, wherein 3. The first and second angle detectors respectively output an output AC signal indicating an electrical phase angle corresponding to a relative rotational position of the coil unit and the magnetic response unit to the induction output signal. 3. The self-calibration type angle detection device according to claim 2, wherein the absolute angle detection data is generated by digitally measuring an electrical phase angle corresponding to the relative rotational position in the output AC signal. . 4. The apparatus according to claim 2, wherein the second angle detector including both the coil unit and the magnetic response unit is located at the mounting angle position with respect to the first angle detector.
The self-calibration type angle detection device according to 1. 5. The method according to claim 5, further comprising: fixing an arrangement of the coil portions in the first and second angle detectors, and indexing the magnetic response portion of the second angle detector to the mounting angle position. 3. The self-calibration type angle detecting device according to claim 2, wherein the angle detector is indexed. 6. A storage unit for storing an error value calculated by the error calculation unit, wherein the calibration unit reads the error value stored in the storage unit in accordance with the current angle detection data, and reads the error value. The self-calibration type angle detecting device according to claim 1, wherein a value is used. 7. A first position detector and a second position detector for outputting position detection data having a predetermined resolution are mounted on a common machine shaft, and the first position detector is mounted on the machine shaft. With respect to the position, the mounting position of the second position detector with respect to the machine axis is relatively shifted by a predetermined position to determine a plurality of relative mounting positions, and for each of the determined relative mounting positions, Obtaining respective position detection data of the first and second position detectors over a predetermined movement range of the mechanical axis; and corresponding to each relative mounting position, each position detection data of the first position detector. Calculating the difference between the position detection data of the first and second position detectors, and obtaining relative position error data based on the difference; Calculating an error value for each position detection data of at least the first position detector based on the relative position error data for each of the relative mounting positions; At least according to current position detection data output from the first position detector,
Using the error value corresponding to the current position detection data,
Calibrating the value of the current position detection data according to the error value. 8. A machine-readable recording medium, comprising:
The method includes a group of instructions of a program for a detection accuracy calibration method executed by a computer. The detection accuracy calibration method includes a first position detector and a second position detector that output position detection data having a predetermined resolution. A detector is mounted on a common machine shaft and the first
The mounting position of the second position detector with respect to the machine axis is determined so as to form a predetermined relative position with respect to the mounting position of the position detector, and in this state, the accuracy of the position detector is determined according to the program. This program performs a process for calibration, and the program executes a process in which a relative mounting interval between the mounting position of the second position detector and the mounting position of the second position detector is different from each other. Calculating a relative mounting position of the second position detector with respect to the first position detector at any one of the plurality of index positions;
A first step of obtaining position detection data of the first and second position detectors with respect to all rotation positions of the machine shaft in this state; and a remaining one of the plurality of predetermined index positions. Then, relative mounting positions of the second position detector with respect to the first position detector are sequentially determined, and the first and second positions are determined for each of the determined positions over a predetermined movement range of the machine shaft in that state. A second step of obtaining position detection data of each of the position detectors; and, for each position detection data of the first position detector, corresponding to each of the index positions, of the first and second position detectors. A third step of obtaining a difference between the position detection data and obtaining relative position error data based on the difference, and for each position detection data of the first position detector, the phase for each of the index positions. Position based on the error data, a recording medium characterized by comprising a fourth step of calculating each error value for each position detection data of at least a first position detector. 9. The error value corresponding to the current position detection data according to at least current position detection data output from the first position detector in accordance with a current position of a detection target machine axis. 9. The recording medium according to claim 8, further comprising: a fifth step of calibrating the value of the current position detection data according to the error value by using the following. 10. The recording medium according to claim 9, wherein in the fourth step, the calculated error value is stored in a memory, and in the fifth step, the error value is read from the memory and used.