JP2607048B2 - Correction data creation method, rotation angle detection method, and rotation angle detection device in resolver - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for creating correction data, a method for detecting a rotation angle, and a device for detecting a rotation angle in a resolver having a single rotation detection resolver and a multipolar resolver.

[0002]

2. Description of the Related Art Conventionally, as shown in FIG. 2, a one-rotation detecting resolver 12 and a multi-polar resolver 13 are incorporated in a direct drive motor 11 together.
There is a resolver that detects the rotation position of the rotation shaft 14 with high resolution based on output data corresponding to the rotation angle of the rotation shaft 14 of the motor 11 detected by the motor 3. When the one-rotation detection resolver 12 and the multi-pole resolver 13 are used in combination as described above, the approximate rotation angle of the rotation shaft 14 is specified based on the output data corresponding to the rotation angle of the one-rotation resolver 12, and A high-resolution rotation angle is detected based on output data corresponding to the rotation angle of the pole resolver 13.

The rotor 12a provided on the rotation shaft 14 side of the one-rotation detection resolver is formed of a disk which is eccentrically supported on the axis of the rotation shaft 14 and has, for example, an eccentric amount of about 0. It is set to 6 mm. A stator side salient pole is provided at a position facing the outer peripheral surface of the eccentric rotor 12a on the fixed side of the motor 11. Also, a multi-pole resolver 13
Is formed in a gear shape having 132 salient poles, for example. Multi-pole rotor 1 on the fixed side of motor 11
Similarly, a stator side salient pole is provided at a position facing the outer peripheral surface of 3a.

When the reference signal is input to the coil 12b on the stator side of the one-rotation detection resolver 12, the eccentric rotor 1
As shown in FIG. 4, when the coil 2b is rotated, the rotation angle of the eccentric rotor 12a, that is, 360 ° whose position is determined by the gap between the stator side salient pole and the eccentric rotor 12a is defined as one cycle. sine wave D ₁ is output.
This sine wave D ₁ is converted by a resolver-to-digital converter into digital data indicated by a straight line E ₁ in FIG.

[0005] Further, the multi-pole rotor 13 has a reference signal input to the coil 13b on the stator side of the multi-pole resolver 13.
a is rotated from the coil 13b, as shown in FIG.
A sine wave D ₂ having one cycle of the rotation angle per pole whose position is determined by the positional relationship between each pole of a and the stator-side salient pole is output. This sine wave D ₂ is converted by a resolver-to-digital converter into digital data indicated by a straight line E ₂ in FIG.

However, in the resolvers 12 and 13, the output data of the resolvers 12 and 13 with respect to each actual rotation angle is affected by the processing accuracy of the resolvers 12 and 13 and the winding method of each excitation winding. error is superimposed (for example, indicated by a broken line F ₁ in FIG. 5). Therefore, each resolver 1
It is difficult to detect the rotation angle by using the output data of Nos. 2 and 13 as they are. Therefore, each resolver 1
The motor 11 in which the motors 2 and 13 are incorporated has its rotary shaft 1
4 with respect to the actual rotation angle (reference data)
Two or thirteen output data are measured, and the difference between the output data and the reference data is stored in the control computer as correction data for each output data. And the motor 11
Is used, an actual rotation angle for the output data is obtained by adding correction data stored in advance to the output data of each of the resolvers 12 and 13.

A method of creating correction data in a resolver as described above is disclosed by the present applicant in Japanese Patent Application Laid-Open No. 4-20813. The creation of the correction data is performed by the measuring device shown in FIG. A measurement motor 15 and a reference encoder 16 are coaxially connected to a rotating shaft 14 of the direct drive motor 11, and the direct drive motor 11 and the reference encoder 16 rotate in synchronization with the rotation of the measurement motor 15. Measurement motor 15
Is controlled by the control computer 17 based on the control program and the output data of the reference encoder 16.

When a measurement signal is output from the reference transmitter 18 to the stator side of the one-rotation detection resolver 12,
As indicated by the curve D ₁ in FIG. 4, one rotation detection resolver 12
An output signal corresponding to the rotation angle θ is output from the stator side. When the measurement signal is output from the reference transmitter 18 to the stator side of the multi-pole resolver 13, a curve D shown in FIG.
As shown by _2, the output signal corresponding to the rotation angle of the multipolar resolver 13 is output from the stator side. The output signal of the one-rotation detection resolver 12 is converted by a resolver-to-digital converter (hereinafter referred to as an RD converter) 20 having a resolution, that is, a data number of R (= 2 ¹² = 4096) 20 via a changeover switch 19 into digital data A _na (n). Is an integer from 0 to 40
95), and input to the control computer 17. The output signal of the multi-pole resolver 13 is converted into digital data I _mb (m is an integer by a resolver-to-digital converter (hereinafter referred to as an RD converter) 21 having a resolution of r (= 2 ¹⁰ = 1024). 0 to 1023)
After that, the data is input to the control computer 17.
The output data of the reference encoder 16 is shown by a straight line E _{3 in} FIG.
The characteristic is indicated by.

However, since the error is superimposed on the output data A _na actually output from the RD converter 20 for the above-described reason, the rotation angle θ-output data A _na characteristic is shown in FIG.
A characteristic as shown by a broken line F ₁ to. That is, in FIG. 5, the difference between the reference data indicated by a certain rotation angle output data A _na and the straight line E ₁ which error is superimposed for θ becomes error.

The measurement of correction data for correcting this error is performed as follows. First, the control computer 1
7 drives the measuring motor 15 and drives the RD converter 20
Every time the output digital value from the input data changes by a predetermined number of data intervals (hereinafter referred to as a set number; for example, 2 ⁰ ), the output data E _na of the reference encoder 16 at that time (n is an integer)
And the output data A _na of the one-rotation detection resolver 12 are captured and stored as output data (E _na , A _na ). The control computer 17 repeats the fetching of the output data A _na a predetermined number of times, and the average value <E of the number of output data E _na obtained by repeating the output data A _{na at} that time.
_na > is calculated and stored.

Similarly, the control computer 17 controls the RD
Every time the output digital value from the converter 21 changes by a set number (for example, 2 ⁶ ), the output data E _mb of the reference encoder 16 and the output data I _mb of the multipole resolver 13 at that time are taken in, and the output data (E _mb , I _mb ).
This is repeated a predetermined number of times, and the number of output data E _mb obtained by repeating the output data I _{mb at} that time is obtained.
_Are calculated and stored, respectively.

Next, the control computer 17 controls the motor 11
1 rotation detection resolver 1 with respect to the rotation angle of rotation shaft 14 (hereinafter referred to as initial position) in a state where is not driven.
2 output data (hereinafter referred to as initial position data)
The polar expression Ax ₀ of A ₀ is calculated by the following equation.

Ax ₀ = M · A ₀ / R (1) (where M is the number of poles of the multi-pole resolver 13) That is, the initial position data A ₀ is determined by the integer value [Ax ₀ ] of the polar expression Ax _0. The pole of the existing multipole resolver 13, that is, the initial pole [Ax ₀ ] is specified.

The control computer 17 determines this pole [Ax ₀ ].
, A pole [Ax _mb ] is _assigned to each output data I _mb of the multi-pole resolver 13. The pole [Ax _mb ] is allocated by determining both the _immediately preceding and _succeeding output data _Imb and the number of data r of the RD converter 21 (in this case, the maximum value r = the number of data r of the RD converter 21 = 1024).

[0015] Each output data I _mb to electrode [Ax _mb] is assigned, the output data I _mb is absolute reduction based on the assigned pole [Ax _mb], absolute data "I _mb" is calculated You. This absoluteization is performed based on the following equation.

_<< I _mb >> = r · ([A _Xmb ] −1) + I _mb (2) The control computer 17 stores the absolute data << I _mb >> corresponding to the calculated output data I _mb .

Next, the control computer 17 outputs the output data <E closest to the output data <E _na > of the reference encoder 16 paired with the output data A _na of the one-rotation detection resolver 12.
_mb > is read out from the output data I _mb of the multipolar resolver 13 forming a pair with the _mb >. Subsequently, the absolute data << I _mb >> for the read output data I _mb is read. And
From the read output data A _{na of} the one-rotation detection resolver 12 and the absolute data << I _mb >>, temporary one-rotation correction data ΔAx _na is calculated by the following equation.

ΔAx _na = R · << I _mb >> / (r · M) −A _na (3) The one-turn correction data ΔAx _na thus obtained is compared with the data number R of the RD converter 20 to obtain a regular data. 1
After being corrected rotation correction data ΔAx _na, PROM22
Is stored.

That is, the normal one-turn correction data Δ
Ax _na represents the difference between the straight line E _{1 and} the broken line F ₁ in the output data A _na , that is, the error. Next, the control computer 17 obtains correction output data (= A _na + ΔAx _na ) from the output data A _na and the one-rotation correction data ΔAx _na , and obtains Ib _na, which is a polar expression of the output data A _na , by the following equation.

That is, the polar expression Ib _na is expressed as follows: Ib _na / M = (A _na + ΔAx _na ) / R (4) From the relation, Ib _na = M · (A _na + ΔAx _na ) / R (5) , A value obtained by converting the pole expression Ib _na to an integer is used as a temporary pole [Ib
_na ].

Each temporary pole [Ib_na] Is the output data A_naPole of
Expression Ib_naAnd multipolar resolver data I _mbBased on the regular
Pole [Ib_na]. The regular
Pole [Ib_na] And multipolar resolver data I_mbFrom the complete abu
Solution data Ia_mbIs calculated by the following equation.

Ia _mb = r · [Ib _na ] + I _mb (6) Next, the control computer 17 _{calculates the} reference encoder obtained for the complete absolute data Ia _mb and the output data I _{mb of the} multipolar resolver. 16 output data E _mb
In order to obtain the multipole correction data ΔIx _mb which is the difference from
First, the reference data _Ipmb is calculated by the following equation.

That is, the reference data Ip _mb is Ip _mb / (r · M) = E _mb / P (7) P is the maximum data number of the reference encoder 16, and I _pm = E _mb · r · M / P (8) Then, the control computer 17 sends the reference data Ip _mb
Correction data Δ of multipolar resolver data I _mb based on
Ix _mb is calculated by the following equation.

ΔIx _mb = Ip _mb −Ia _mb (9) The multipolar correction data ΔIx _mb calculated in this way is a PROM
22. As described above, the one-rotation correction data ΔAx for the output data A _na of the one-rotation detection resolver 12
the direct drive motor 11 which multipolar correction data DerutaIx _mb is calculated with respect to the output data I _mb of _na and multipolar resolver 13 is used in block configuration shown in FIG. The motor 11 is connected to a reference oscillator 23, a changeover switch 24, a resolver-digital converter (hereinafter referred to as an RD converter) 25 having the same resolution (that is, the number of data) as the RD converter 21, and a control computer 26. The PROM 22 is connected to the control computer 26.

[0025] First, the control computer 26 takes in the output data (hereinafter referred to as initial position data) A _k of 1 rotation detection resolver 12 in the state where the motor 11 is not driven through the RD converter 25. Then, the control computer 26 reads the one-rotation correction data ΔAx _k corresponding to the initial position data A _k from the PROM 22, and calculates the polar expression Ib _k by using equation (5).

Ib_k= M · (A_k+ ΔAx_k) / R (5a) The control computer 26 calculates the calculated polar expression Ib._kSet
Digitized and the initial pole [Ib _k] Is calculated.

Next, the control computer 26 based on the calculated initial electrode [Ib _k], a complete absolute of output data (hereinafter referred to an initial position data) I _k of the multipolar resolver 13. That is, to calculate the complete absolute data Ia _k by a formula (6).

[0028] _{Ia k = r · [Ib k} ] + I k ... (6a) The control computer multipolar correction data DerutaIx _k corresponding to the complete absolute data Ia _k PRO
Read from M22, the initial position data I _k rotary shaft 14
The rotational position data H _k as the data corresponding to the rotation angle θ of the rotational position of the calculated by the following equation.

H _k = Ia _k + ΔIx _k (10) The control computer 26 outputs the rotation position data H _k as rotation initial position data corresponding to the rotation angle of the direct drive motor 11.

Thereafter, when the initial rotational position data H _k is obtained, the control computer 26 obtains the rotational position data H _k using only the output data I _mb of the multipolar resolver 13. That is, the control computer 26 calculates the difference between the multi-polar resolver data I _{k + 1} output from the multi-polar resolver 13 after a predetermined time and the initial position data I _k by the number r of data of the multi-polar resolver 13.
Compare with Then, the output data I of the multipolar resolver 13
_It is determined whether the pole [Ib _k ] where _k exists has moved to the next position and the pole [Ib _k ] is corrected. The control computer 26 calculates the complete absolute data Ia _k by a formula (6) based on the modified electrode [Ib _k].

[0031] _{Ia k = r · [Ib k} ] + I k ... (6b) further reads out the correction data Ia _k corresponding to full absolute data Ia _k, the rotational position data H _k corresponding to the rotation angle (10) Is calculated.

H _k = Ia _k + ΔIx _k (10a) Thus, the control computer 26 sequentially turns the multi-pole resolver data I as the direct drive motor 11 rotates.
_{k + 1, I k + 2} , ... uptake, calculation of complete absolute data Ia _k for each multipole correction data ΔIx
performs calculation of _k read and rotational position data H _k, and outputs the sequential rotational position data H _k.

[0033]

If the resolver 12 and the multi-pole resolver 13 are provided close to each other or are designed to be small in order to reduce the size of the resolver, the magnetic circuit of each other interferes with each other. rotation angle θ- output data a _na characteristics of 12 greatly varies as shown by curve F ₂ in FIG. As a result, there are places where the amount of change of the rotation angle θ with respect to the amount of change of the output data _Ana becomes extremely large.

Therefore, in such a location, one rotation compensation is performed.
Positive data ΔAx_naIs created, as shown in FIG.
Output data of the rotation detection resolver 12 (1 rotation setting data)
T) A _z-1, A_Z, A_{Z + 1}Correction data corresponding to each
Data ΔAx_Z-1, ΔAx_Z, ΔAx_{Z + 1}Changes greatly
You.

On the other hand, the direct drive motor 11 is used.
When used, excessive external force is applied in the radial direction of the rotating shaft 14
And the center of the rotor 12a of the one-rotation detection resolver 12
Offset to the center of the data. Eccentricity of eccentric rotor 12a
Since the amount is set to approximately 0.6 mm, 132 poles
The eccentric rotor 12a per pole of the multipole resolver 13 and the
The amount of change in the gap with the data-side salient pole is approximately 7 μm. Outside
When the force becomes excessively large, the center of the eccentric rotor 12a
And the center of the stator may deviate by more than 7 μm.
As a result, the eccentric rotor 12a of the one-rotation detection resolver 12
Eccentric rotor 12a and stator side salient poles
Is changed, the rotation of the eccentric rotor 12a is changed.
Output data A of angle θ and one-turn detection resolver 12_naRelationship with
The engagement changes. On the other hand, the multipolar resolver 13 is
Rotation angle θ and output data I _mbThe relationship is basic
Hardly change. Therefore, the rotation position of the rotation shaft 14
Output data A of one rotation detection resolver 12 with respect to θ_naWhen
Output data I of the multipolar resolver 13_mbThe relationship with is shifted.

[0036] That is, for example, as indicated by two-dot chain lines in FIG. 8, one rotation angle θ- output data A _na characteristics of the rotation detecting the resolver 12 with respect to the rotation angle θ- output data A _na characteristics during correction data measurement Change. As a result, a certain rotation angle θ _k
For example, _AZ-1 or _{AZ + 1} is output in place of the output data _AZ to be output to the.

Therefore, when the initial pole Ib _k is _obtained by using the equation (5a), that is, Ib _k = M · (A _k + ΔAx _k ) / R, the output initial position data _AZ-1 or _{AZ + 1} , the correction data ΔAx _Z−1 , ΔA corresponding to the initial position data A _Z−1 , A _{Z + 1.}
correction output data calculated by adding x _{Z + 1} (= A _k + Δ
Ax _k ) correspond to the values to be obtained at the rotation angles θ _Z−1 and θ _{Z + 1} , respectively.

Therefore, the initial pole [Ib _Z ] calculated based on the corrected output data Is the original pole [Ib _Z] and different poles [Ib _Z]. Therefore, the formula (6a), i.e. _{Ia X = r · [Ib X} ] + If using I _X determine the complete absolute data Ia _Z, completely absolute data Ia _k and this is calculated on the basis of the electrode [Ib _Z] The rotational position data H _Z calculated from the complete absolute data Ia _Z by using the equation (10) has a different result from the rotational position data H _Z that should be originally obtained. Further, thereafter, each rotation position data H _k is sequentially _obtained.
There the rotational position data H _Z containing all errors.

Therefore, as described above, the rotation angle θ−the output data
TA A_naOne-rotation detection resolver 12 having characteristics
Direct drive motor 1 incorporating resolver 13
1, the rotation angle θ-output data A_naSpecial
And the rotation angle θ _ZShould be output at
Output data A_ZOutput data A instead of_Z-1Or A
_{Z + 1}Is output, the correction data ΔA_Z-1Or ΔA_{Z + 1}
At each rotation angle θ _Z-1Or θ_{Z + 1}Compensation output equivalent to
Force data (= A_k+ ΔAx_k) Is obtained.

However, corresponding to the corrected output data
Rotation angle θ_Z-1Or θ_{Z + 1}And the rotation angle θ_zAnd the difference
(Θ_z-1−θ_z) Or (θ_{z + 1}−θ_z) And big,
The initial pole [Ib calculated based on the corrected output data_z]
Original initial pole [Ib_z]. As a result,
This shifted initial pole [Ib_zRotational position obtained based on
Position data H_zIs data containing an error. In addition,
[Ib] sequentially specified by the rotation of the motor 11_z] Is all
Displaced pole [Ib_z], This pole [Ib _z]
Rotational position data H calculated from_zAre all times containing errors
Shift position data H_zBecomes

A first object of the present invention is to solve the above-described problem. Even if an error occurs in the initial position data detected at the initial position, the first object of the present invention is to correct the error by correcting the initial position data. An object of the present invention is to provide a method for creating correction data that can suppress enlargement.

A second object is that even if an error occurs in the initial position data detected at the initial position, it is possible to suppress the expansion of the error due to the correction of the initial position data and detect the rotation angle with high accuracy. An object of the present invention is to provide a rotation angle detection method and a rotation angle detection device in a resolver that can be used.

[0043]

In order to solve the above-mentioned problems, the invention according to claim 1 obtains each reference data of a reference encoder for each one-turn setting data data obtained from a one-turn detection resolver, Obtain each reference data of the reference encoder for each multi-pole setting data obtained from the multi-pole resolver, create absolute data obtained by converting the multi-pole setting data into absolute data for each multi-pole setting data, Is obtained from each reference data of the reference encoder, and one rotation setting data and the absolute data of the multi-pole setting data corresponding to the one rotation setting data are used to calculate the one rotation setting data for the absolute data. One-rotation correction data is created, and reference encoding is performed on each of the multi-pole setting data. The multi-polar reference data for the multi-polar setting data is obtained from the reference data of the multi-polar setting data, and the multi-polar correction data is created from the multi-polar reference data for each multi-polar setting data and the absolute data obtained by converting the multi-polar setting data into absolute data. In the correction data creating method, the one-turn correction data is averaged with at least one or more adjacent one-turn correction data to create averaged correction data, and the averaged correction data is measured for each one-turn measurement of a one-turn resolver. One rotation correction data for the data.

According to a second aspect of the present invention, an initial pole of a multipolar resolver corresponding to the initial position data is obtained from the initial position data obtained from the one-rotation detecting resolver, and the initial pole is obtained from the initial pole and the multipolar resolver. In the rotation angle detection method for obtaining rotation initial position data from the initial position data, determining a pole corresponding to the output data from sequentially obtained output data of the multipolar resolver, and obtaining rotation position data from the poles and the output data, The averaged correction data for each one-rotation setting data of the one-rotation detection resolver and the multi-pole correction data for each multi-pole setting data of the multi-pole resolver, which are created by the correction data creation method for the resolver according to item 1, are stored. First, averaging correction data corresponding to the initial position data of the one-rotation detection resolver is obtained, and Data is corrected to calculate one-rotation initial position data, and the one-rotation correction initial position data and the initial position data of the multi-polar resolver are used to determine the initial pole of the multi-polar resolver. Calculate initial position absolute data obtained by converting the initial position data from the initial position data to absolute data, obtain multipolar correction data corresponding to the initial position absolute data, and correct the initial position absolute data with this multipolar correction data. The multi-pole corrected initial position data is calculated and output as rotation initial position data, and the sequentially obtained multi-pole output data is compared with previously obtained multi-pole output data, and the multi-pole output data corresponds to the corresponding multi-pole output data. Determine the pole of the pole resolver, based on the multi-pole output data and the pole determined,
Absolute data obtained by converting the multipolar output data into absolute data is calculated, multipolar correction data for the absolute data is obtained, and the absolute data is corrected with the multipolar correction data to calculate corrected rotational position data, and the rotational position is calculated. Output as data.

According to a third aspect of the present invention, there is provided a rotation angle detecting device provided with a one-rotation detecting resolver and a multi-pole resolver, wherein the one-rotation angle correction device according to the first aspect is provided. Storage means for storing averaging correction data for each rotation setting data of the rotation detection resolver, multipolar correction data for each multipolar setting data of the multipolar resolver, and averaging corresponding to initial position data of the rotation detection resolver. A one-rotation correction initial position data calculating means for calculating correction data and correcting the initial position data with the averaged correction data to calculate a one-rotation correction initial position data; Initial pole determining means for determining an initial pole of the multipolar resolver from the initial position data; and initial position data from the initial pole and the initial position data of the multipolar resolver. Initial position absolute data calculating means for calculating the initial position absolute data obtained by converting the absolute position data, multi-polarity correction data corresponding to the initial position absolute data is obtained, and the initial position absolute data is corrected by the multi-polarity correction data to obtain a multi-polarity correction data. Multipolar correction initial position data calculating means for calculating the polar correction initial position data and outputting it as rotation initial position data, and comparing the sequentially obtained multipolar output data with the previously obtained multipolar output data, Pole determining means for determining the poles of the multipolar resolver corresponding to the multipolar output data, and absolute data calculation for calculating absolute data obtained by converting the multipolar output data into absolute data based on the multipolar output data and the determined poles Means for obtaining multipolar correction data for the absolute data. The absolute data and a correction position data calculating means for outputting a rotational position data to calculate the corrected position data and corrected by the multipole correction data.

[0046]

Therefore, according to the first aspect of the present invention, each one
The two data are combined based on the rotation setting data and the respective reference data obtained for each multi-pole setting data. Further, absolute data in which each multi-pole setting data is converted into absolute data is created. Then, one-rotation correction data for the absolute data of the one-rotation setting data is created from each one-rotation setting data and the absolute data of the multi-pole setting data corresponding to the one-rotation setting data. The one-rotation correction data is averaged with at least one or more adjacent one-rotation correction data before and after the one-rotation correction data, and averaged correction data corresponding to each one-rotation setting data is created. The difference between adjacent averaged correction data created in this way is smaller than the difference between each one-turn correction data before being averaged.

According to the second aspect of the present invention, the averaging correction data for one rotation setting data of the one-rotation resolver and the multi-pole correction data for multi-pole setting data of the multi-pole resolver are stored. Averaged correction data corresponding to the initial position data of the one-rotation resolver is obtained, and one-rotation corrected initial position data corrected by the averaged correction data is calculated. The one-rotation corrected initial position data is used to determine the initial pole of the multipolar resolver, and absolute data obtained by converting the initial position data into absolute data is calculated from the initial pole and the initial position data of the multipolar resolver. The absolute data is corrected by the corresponding multipolar correction data, and multipolar correction initial position data as rotation initial position data is obtained. At the initial position, if output data deviated from the output data at the time of creating the correction data is taken in as initial position data, an adjacent average different from the averaged correction data corresponding to the actual initial position data corresponding to this output data. The correction is performed using the conversion correction data, and the one-rotation correction initial position data is calculated. The one-rotation correction initial position data calculated in this way is more corrected for the actual initial position than the one-rotation correction initial position data corrected by the one-rotation correction data before averaging the shifted output data. The difference between the output data at the time of data creation and the one-rotation correction initial position data obtained from the one-rotation correction data corresponding to the output data is reduced. Therefore, the initial pole obtained from the one-rotation corrected initial position data corrected by the averaging correction data and the initial position data of the multi-pole resolver is one in which the deviation of the initial position data from the corresponding initial pole is actually suppressed. Becomes

Further, the multi-pole output data obtained sequentially is compared with the multi-pole output data obtained before that, and the poles of the multi-pole resolver corresponding to the multi-pole output data are determined.
The absolute data of the output data is calculated from the determined pole and the output data. Then, the absolute data is corrected by the corresponding multipolar correction data, and corrected rotation position data as rotation position data is calculated.
Therefore, since the deviation of the initial pole of the multi-pole resolver from the initial position data including the error with respect to the actual initial pole is suppressed, the output data of the multi-pole resolver is identified based on the initial pole and identified. The deviation of the actual pole from the actual pole is suppressed. According to the third aspect of the present invention, each one-revolution set data of the one-revolution resolver created by the correction data creating device for creating the correction data by the method of creating the correction data according to the first aspect. And the multi-pole correction data for each multi-pole setting data of the multi-pole resolver are stored in the storage means. The one-rotation correction initial position data calculation means calculates average correction data corresponding to the initial position data of the one-rotation detection resolver, and calculates the one-rotation correction initial position data by correcting the initial position data with the average correction data. I do. The initial pole determining means determines the initial pole of the multipolar resolver corresponding to the initial position data from the initial position data obtained from the one-rotation detection resolver and the averaging correction data for the initial position data. The initial position absolute data calculation means obtains absolute data obtained by converting the initial position data from the initial position data of the initial pole and the multipole resolver. The multi-pole correction initial position data calculating means calculates multi-pole correction data corresponding to the initial position absolute data, and corrects the initial position absolute data with the multi-pole correction data to calculate multi-pole correction initial position data. The pole determining means compares the sequentially obtained multipolar output data with the previously obtained multipolar output data, and determines the pole of the multipolar resolver to which the multipolar output data corresponds. Absolute data calculation means calculates absolute data obtained by converting the multi-pole output data into absolute data based on the multi-pole output data and the determined pole. The correction position data calculating means calculates multi-polarity correction data for the absolute data, and corrects the absolute data with the multi-polarity correction data to calculate correction position data.

[0049]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below with reference to FIG. The generation of the correction data for the output data of each resolver in the present embodiment is performed by the correction data generation device shown in FIG. That is, 1
A measurement motor 15 and a reference encoder 16 are coaxially connected to a rotation shaft 14 of a direct drive motor 11 in which a rotation detection resolver 12 (hereinafter, referred to as a single rotation resolver) and a multipolar resolver 13 are built. Each resolver 12,
A reference transmitter 18 is connected to the input side of each of the 13 coils 12b and 13b. The output side of each coil 12b, 13b has a changeover switch 19 and an RD converter 20,
The control computer 17 is connected to the control computer 21 via the control computer 21.
In this embodiment, the resolution of the RD converter 21 is 2 ¹² (= R) as in the RD converter 20.

The control computer 17 has a central processing unit (hereinafter referred to as CPU) 27 and a temporary storage memory (hereinafter referred to as RA).
M) 28 and a PROM 22 which is a writable read-only memory. The PROM 22 stores a correction data creation program. CPU
Reference numeral 27 denotes a correction data creation program, which is one rotation setting data Ama [i] taken at a predetermined interval from output data Ama and Bma output from the resolvers 12 and 13 via the RD converters 20 and 21, respectively. Multi-pole setting data Bma
Output data EA [i] and EB [j] as reference data of the reference encoder 16 corresponding to [j] are taken in, and correction data ΔAma [i] and ΔBma [for each output data Ama [i] and Bma [j]. j] is calculated.

The CPU 27 converts the one-rotation correction data ΔAma [i] corresponding to the calculated one-rotation resolver 12 output data Ama [i] into the one-rotation correction data ΔAma before and after it.
[I−1], ΔAma [i + 1] and averaged correction data <ΔAma corresponding to the output data Ama [i].
[I]> is created. Furthermore, CPU
27 controls the rotation of the measuring motor 15 and the changeover switch 1
9 is switched. The PROM 22 stores the calculated averaged correction data <ΔAma [i]> and multipolar correction data ΔBma [j].

Next, a method of creating correction data by the above-described correction data creating apparatus will be described. First, CPU2
7 executes reference data creation processing. In the reference data creation process, first, the changeover switch 19 is switched to the one-turn resolver 12 to enter a standby state for taking in the output data Ama, and the measuring motor 15 is driven to
Is rotated from its initial position. Here, the initial position refers to a position immediately before driving of the motor 11. Control computer 1
Reference numeral 7 denotes an interval between output data Ama from the RD converter 20 and a predetermined number of data (hereinafter referred to as a set number) e with the rotation.
(= 2 ² = 4) Every time it changes, the output data Ama [i] 0, e, 2e,..., I · e,..., (X−2) e, (x− 1) From the reference encoder 16 corresponding to e (x = 2 ¹⁰ = 1024 = R / e; x is the number of data per rotation; 0 ≦ i ≦ x−1; ∴0 ≦ i ≦ 1023; i is an integer) Output data EA
[I] EA

[0], EA [1],... EA [i],..., EA [x-2], EA [x-1] Here, x (= 2 ¹⁰ = 1024) is the maximum data value R (= 2 ¹² = 2) which is the resolution of the RD converter 20.
4096) divided by the set number e (= 2 ² = 4), and the CPU 27 outputs 1 while the motor 21 rotates 360 °.
This is the number of data taken from the rotating resolver 12.

The output data Ama [i] at that time is
Of the output data EA [i] of the reference encoder 16 with respect to

(0)), (e, EA [1]), ... (iee, EA [i]), ... (x-2) e, EA [X-2]), ((x-1) e, EA [(x-1)]) are stored in the RAM 28.

Next, the CPU 27 sets the changeover switch 19 to 1
It switches to the 32-pole resolver 13, enters a standby state for taking in the output data Bma, drives the measuring motor 15, and rotates the motor 11 from the initial position. The CPU 27 outputs the output data Bma from the RD converter 21 with the rotation.
Every time the set number f (= 2 ⁷ = 128) changes, that is, 1
Output data Bma [j] (0, f, 2f, ..., (n-1) f) ₁ , (0, f, 2f, ..., (n-1)) as multi-pole setting data from the 32-pole resolver 20 f) ₂ ,... (0, f, 2f,..., (n-1) f) ₁₃₂ (however, n = ²⁵ = 32 = R / f; per pole) 0 ≦ j ≦ y−1; {0 ≦ j ≦ 4223; j is an integer; y = n × 132 = 4224) output data EB from the reference encoder 16 corresponding to
[J] EB

[0], EB [1], ..., EB [j], ..., EB [y-2], EB [y-1] are obtained. Here, n (= 2 ⁵ = 32) is a value obtained by dividing the maximum data value R (= 2 ¹² = 4096) of the RD converter 21 by the set number f (= 2 ⁷ = 128). / 132) °, ie, for each pole, CP
U27 is the number of data taken from the 132-pole resolver 20. Therefore, when the 132-pole resolver 20 rotates 360 °, the total number of data becomes y (= n × 132 = 4224).

Then, the output data Bma [j] at that time is output.
, The value of the output data EB [j] of the reference encoder 16, that is, (0, EB

(0)), (f, EB [1]), ... ((n-1) f, EB [n-1]), (0, EB [n]), (f, EB [n + 1]),... ((N-2) f, EB [y-2]) ((n-1) f, EB [y-1]) are stored in the RAM 28.

Next, the CPU 27 executes a one-rotation correction data creation process. The one-rotation correction data creation process is performed by calculating the one-rotation correction data ΔAma for the output data Ama [i] from the measured output data Ama [i] of the one-rotation resolver 12 and the corresponding output data EA [i] of the reference encoder 16.
[I] is created.

First, the CPU 27 sets | EA [i] −EB [h] | (0 ≦ h ≦ y−1∴0 ≦ h ≦ 4223, where h is an integer, for the output data EA [i] of the reference encoder 16. ) Is minimized, and the output data EB [h] corresponding to the number i of the output data EA [i] is obtained.
Is obtained. Then, the i-th one-rotation correction data ΔAma [i] of the one-rotation resolver 12 is calculated by the following equation, and P
It is stored in the ROM 22.

ΔAma [i] = h · fi · e × 132 = 128 · h−4 · i × 132 Here, the first term h · f on the right side of the above equation is a 132 pole resolver 1
3 is an absolute value of the output data Bma [h]. In the second term on the right side, i · e represents output data Ama [i] of the one-rotation resolver 12, and i · e × 132 represents this output data Ama [i] as output data Bma [i] of the 132-pole resolver 13. It is the value converted to. Therefore, f is 12
8, e is 4, so ΔAma [i] = 128 · h−4 · i
× 132.

Next, the CPU 27 executes a multipolar reference data creation process. The multi-pole reference data creation processing is performed by converting the output data EB [j] of the reference encoder 16 corresponding to each output data Bma [j] of the multi-pole resolver 13 into the output data B
The multipolar reference data Ip [i] for ma [i] is obtained by the following equation.

Ip [j] = EB [j] · y · f / Pmax (where Pmax is the maximum data value of the reference encoder 16) Here, the right side of the above expression is the output data EB [j] of the reference encoder 16 This is a value converted into output data Bma of the 132-pole resolver 13.

Next, the CPU 27 executes a multi-pole correction data creating process. The multipolar correction data creation processing includes multipolar reference data Ip [j] for each output data Bma [j] and absolute data j for the output data Bma [j].
The multipolar correction data ΔBma [j] is calculated from f and is stored in the PROM 22 by the following equation.

ΔBma [j] = Ip [j] −j · f where the second term j · f on the right side of the above equation is output data Bma
This is a value obtained by converting [j] into an absolute value.

Next, the CPU 27 executes an averaging correction data creation process. The averaging correction data creation processing is performed for each one rotation correction data Δ created in the one rotation correction data creation processing.
Ama [i] is averaged with the one-turn correction data ΔAma [i-1] and ΔAma [i + 1] adjacent to each other before and after the Ama [i] according to the following equation to create averaged correction data <ΔAma [i]>. Then, it is stored in the PROM 22.

<ΔAma [i]> = ΣΔAma [i + u] / 3; (u is −1, 0, 1) = (ΔAma [i−1] + ΔAma [i] + ΔAma [i + 1]) / 3 Averaged correction data <ΔAma [i]> for output data Ama [i] of the rotary resolver 12 and multipolar correction data ΔBma [j] for output data Bma [j] of the multipolar resolver 13 are created and stored in the PROM 22. Is done. These two correction data <ΔAma [i]>, ΔBma
The PROM 22 in which [j] is stored is the one-turn resolver 12
The multi-polar resolver 13 is used by being incorporated in a control computer 26 of a rotation angle detecting device in which the resolver 13 is used.

Next, the motor 1 in which correction data <ΔAma [i]> and ΔBma [j] for the output data Ama [i] and Bma [j] of the built-in resolvers 12 and 13 are prepared.
A rotation angle detection device for measuring the rotation angle of the output shaft 1 will be described. Note that the rotation angle detection device of this embodiment has basically the same configuration as the conventional example. That is, each coil 12b, 13 of each resolver 12, 13 built in the motor 11
The reference transmitter 23 is connected to the input side of b. The output side of each coil 12b, 13b is connected to a changeover switch 24 and R
It is connected to a control computer 26 via a D converter 25. The resolution of the RD converter 25 is 2 ¹² (= R), similarly to the RD converters 20 and 21 of the correction data creation device.

The control computer 26 has a central processing unit (hereinafter referred to as CPU) 29 and a temporary storage memory (hereinafter referred to as R).
AM) 30 and a PROM 22 as storage means. In this PROM 22, the averaged correction data <ΔAma
[I]> and multipolar correction data ΔBma [j] are stored. That is, the averaging correction data <ΔAma [i]>
One-rotation resolver 1 output via D converter 25
2 is correction data for each output data Ama [i], and multi-pole correction data ΔBma [j] is correction data for each output data Bma [j] of the 132-pole resolver 13 which is also output via the RD converter 25. .

The PROM 22 stores a rotational position calculation program. The CPU 29 uses this rotation position calculation program to output the output data Ama (hereinafter referred to as the initial position data A) for the input initial position θK of the one-rotation resolver 12.
K) corresponding to the averaged correction data <ΔAma
The one-rotation correction initial position data AKS corrected in [i]> is calculated. Further, the CPU 29 obtains one-rotation corrected initial position data AKS and output data Bma (hereinafter referred to as initial position data BK) corresponding to the initial position of the 132 pole resolver 13 by one.
A pole corresponding to the initial position data BK of the 32-pole resolver 13 (hereinafter referred to as an initial pole TBK) is determined. Also, CPU 29
Calculates the initial position absolute data BaK obtained by converting the initial position data BK into an absolute value from the initial pole TBK and the initial position data BK. Then, the initial position absolute data BaK is converted to the corresponding multipole correction data ΔBma.
The corrected multi-pole initial position data BKS as the rotational initial position data corresponding to the initial position θK is calculated by correcting in [j]. Further, the CPU 29 outputs the output data Bma (hereinafter referred to as Bma) of the 132-pole resolver 13 obtained every time a predetermined time elapses.
(T) is compared with the previously obtained output data Bma to newly determine a corresponding pole (hereinafter referred to as TBK (t)). Also, this pole TBK (t) and the output data Bma
From (t), absolute data BaK (t) obtained by converting output data Bma (t) into absolute data is calculated. Then, the absolute data BaK (t) is corrected with the corresponding multipolar correction data ΔBma (t), and corrected position data BKS as rotational position data for the rotational position is corrected.
(T) is calculated.

The above-mentioned data is temporarily stored in the RAM 30. Next, a method of detecting a rotation angle by the rotation angle detection device will be described.
First, the CPU 29 performs one-rotation correction initial position data calculation processing as one-rotation correction initial position data calculation means in a state where the motor 11 is at the initial position θK. In the one-rotation correction initial position data calculation process, first, the initial position data AK of the one-rotation resolver 12 and the initial position data BK of the 132-pole resolver 13 are fetched. The CPU 29 divides the acquired initial position data AK of the one-rotation resolver 12 by the set number f, and obtains a value obtained by rounding down the decimal portion. Then, the averaged correction data <ΔAma [i]> corresponding to the obtained value is set to P
Read from ROM22. That is, the following processing is performed.

<ΔAma [i]> = <ΔAma [[AK / f
+0.5]]> Here, [] is a Gaussian symbol and [AK / f
+0.5] means that the value obtained by dividing the initial position data AK by the set number f is rounded off. That is, <ΔAma [[A
K / f + 0.5]]] is the averaged correction data <Δ corresponding to the output data Ama [i] closest to the initial position data AK.
Ama [i]>. Next, the CPU 29 calculates corrected initial position data AKS which is a value obtained by correcting the initial position data AK of the one-rotation resolver 12 with the averaged correction data <ΔAma [[AK / f + 0.5]]> by the following equation. I do.

AKS = AK × 132 + <ΔAma [[AK /
f + 0.5]]> Here, the first term on the right side of the above equation is the initial position data AK of 13
It represents a value converted into output data Bma [j] of the bipolar resolver 13.

Next, the CPU 29 executes an initial pole finding process as initial pole finding means. In the initial pole determination process, first, the calculated corrected initial position data AKS of the one-rotation resolver 12 is divided by n · f (= R), which is the maximum data value of the RD converter 25, and the quotient AKS (H) and the remainder are obtained. Find AKS (L). Value obtained by adding 1 to the quotient AKS (H) is the initial pole Tc _k tentative 132-pole resolver 13.

AKS / (n · f) = AKS (H)... AKS (L) Next, the CPU 29 sets the initial position data BK and the remainder AKS
(L) and the difference is calculated as the maximum data value n ·
The value G divided by f (= R) is calculated by the following equation.

G = (BK−AKS (L)) / (n · f) This value G is the ratio of the difference between the initial position data BK and the remainder AKS (L) to the number of data per pole of the 132-pole resolver 13. Represents The CPU 29 determines a temporary initial pole TBK (= A) calculated from the corrected initial position data AKS based on the value G.
KS (H) +1) is corrected to a true initial pole which actually outputs the initial position data BK. That is, the true initial pole TBK (= RN / (n · f) +
Find out 1). Here, the value RN is determined as follows.

When -3 / 8≤G≤3 / 8, RN = A
KS (H) × (n · f) (In this case, the initial pole is TBK = AKS (H) +1) When 5/8 ≦ G <1, RN = (AKS (H) −
1) × (n · f) (In this case, the initial pole is TBK = AKS (H)) When −1 <G ≦ −5 / 8, RN = (AKS (H) +1)
× (n · f) (at this time, the initial pole is TBK = AKS (H) +2) That is, the value RN is the pole number of the pole immediately before the initial pole TBK, RD
This is a value obtained by multiplying the maximum data number R (= n · f) of the converter 25. In the case of −3 / 8 ≦ G ≦ 3/8, the pole TBK that has output the initial position data BK (that is, the initial pole of the 132 pole resolver 13) is the temporary pole calculated by the corrected initial position data AKS. This is the case where the pole is the same as the initial pole TBK (= AKS (H) +1), and the true initial pole TBK is AKS (H) +1
Becomes When 5/8 ≦ G <1, the initial pole TBK is the pole immediately before the temporary initial pole (AKS (H)), and the true initial pole TBK is AKS (H). Becomes Also, -1 <G
If ≦ −5 / 8, the initial pole TBK is the pole (AKS (H) +2) immediately after the provisional initial pole, and the true initial pole TBK is AKS (H) +2. .

Here, the initial position data BK and the remainder AKS
It is assumed that the difference from (L) is at most smaller than R / 2. This is because if the difference is larger than R / 2, two or more combinations occur in the positional relationship between the initial position data BK and the remainder AKS (L). This is because it is not possible to specify any. That is, the value G is in the range of -1 <G <1, and in the range of -1/2 <G <1/2, the value G corresponding to the initial pole TBK that has output the initial position data BK and the remainder AKS (L). Are the same poles. Also, -1 <G
In the range of <-1/2, it can be seen that the initial pole TBK is the pole immediately after the temporary pole. Conversely, in the range of 1/2 <G <1, it can be seen that the initial pole TBK is the pole immediately before the temporary pole.

In this embodiment, the range of the value G is set to the above three cases, namely, -3 / 8≤G≤3 / 8, 5 / 8≤G <1 and -1 <G≤-5 / 8. Except for the ranges of -5/8 <G <-3/8 and 3/8 <G <5/8, the initial position data BK exists due to the detection error of the initial position data AK and BK. This is to prevent the relationship between the pole and the temporary initial pole TBK determined from the corrected initial position data AKS from being erroneously determined. That is, when the value G in this range is obtained, the CPU 29 determines that the data is abnormal and sets the initial pole TBK
, And output data Ama and Bam obtained next by the rotation of the motor 11 are used as initial position data AK and BK.
To determine the initial pole TBK.

Next, the CPU 29 is operated by the 132 pole resolver 13.
Of the initial position absolute data calculation means as the initial position absolute data calculation means. The initial position absolute data calculation process includes the value RN calculated in the initial pole indexing process and the 132 pole resolver 1.
3 from the initial position data BK
Is calculated by the following equation.

BaK = RN + BK Next, the CPU 29 executes a multipolar correction initial position data calculation process as multipolar correction initial position data calculation means. The multi-polar correction initial position data calculation processing is performed by multi-polar correction data ΔB
The corrected initial position data BKS which is a value obtained by correcting the absolute data BaK by ma [j] is calculated by the following equation.

BKS = BaK + ΔBma [[BaK / f + 0.5]] That is, the calculated absolute data BaK is divided by the set number f, and a value obtained by rounding down the decimal portion is obtained. The absolute data BaK is corrected with the multipolar correction data ΔBma [j] corresponding to the obtained value, and the corrected initial position data B is corrected.
Calculate KS.

Then, the CPU 29 outputs the calculated corrected initial position data BKS as rotation initial position data at the initial position θK. Next, the CPU 29 executes a pole indexing process as a pole indexing unit, an absolute data calculating process as an absolute data calculating unit, and a corrected position data calculating process as a corrected position data calculating unit.

First, the CPU 29 determines the initial position data AK,
When a predetermined time t elapses after fetching BK, the output data Ama (t) and Bma (t) of the one-rotation resolver 12 and the multi-polar resolver 13 are fetched again via the RD converter 25. Here, the output data Ama (t) is
Means output data Ama which is taken t hours after the initial position data AK is taken from the one-rotation resolver 12,
The output data Bma (t) means output data Bma which is also taken from the 132-pole resolver 13 after t time. still,
This time t is set to a value such that the maximum rotation speed of the motor 11 does not jump over the rotation angle of the 132 pole resolver 13 with respect to one pole.

In the pole indexing process, first, it is determined whether or not the following inequality holds for the output data Bma (t) of the 132 pole resolver 13 taken in after the time t. | BK-Bma (t) | ≥R / 2 | BK-Bma (t) | represents the absolute value of (BK-Bma (t)). The above inequality is used to determine whether the pole TBK (t) where the output data Bma (t) of the 132 pole resolver 13 exists is the same as the initial position data BK, or has moved to the next pole before and after. That is, by comparing the difference between the initial position data BK and the output data Bma (t) with half the number R of data per pole of the 132 pole resolver 13,
It is determined whether the initial position data BK and the output data Bma (t) are the same pole or adjacent poles.

If the inequality expression does not hold, that is, if the pole TBK (t) where the output data Bma (t) of the 132-pole resolver 13 exists is the same as the initial pole TBK,
As absolute data calculation processing, absolute data Bak (t) is calculated by the following equation.

BaK (t) = RN + Bma (t) where RN = (TBK−1) × (n · f) Then, the CPU 29 calculates the corrected position data as follows:
Output data Bma (t) with multipolar correction data ΔBma [j]
The corrected position data BKS (t), which is a value obtained by correcting, is calculated by the following equation and is output as rotational position data.

BKS (t) = BaK (t) + ΔBma [[BaK
(T) /f+0.5]] That is, the calculated absolute data BaK (t) is divided by the set number f, and a value obtained by truncating the fractional part is obtained.
The absolute data BaK (t) is corrected with the multipolar correction data ΔBma [j] corresponding to the obtained value, corrected position data BKS (t) is calculated, and output as rotational position data.

When the inequality is satisfied, that is, the pole TBK in which the output data Bma (t) of the 132 pole resolver 13 exists
If (t) moves to the pole next to the initial pole TBK, the CPU
Numeral 29 performs the following calculation as pole finding processing.

RN = RN ± (n · f) This equation indicates that the pole in which the output data Bma (t) of the 132-pole resolver has moved to the pole next to the initial pole TBK.

Next, the CPU 29 calculates the absolute data BaK (t) by the following equation as an absolute data calculation process. BaK (t) = RN + BK where RN = RN ± (n · f) = (TBK−1) × (n ·
f) ± (n · f) Then, the CPU 29 calculates the corrected position data BKS (t) by the following equation as a corrected position data calculation process, and outputs it as rotational position data.

BKS (t) = BaK (t) + ΔBma [[BaK
(T) /f+0.5]] That is, the calculated absolute data BaK (t) is divided by the set number, and a value obtained by truncating the decimal part is obtained. The absolute data BaK (t) is corrected with the multipolar correction data ΔBma [j] corresponding to the obtained value, corrected position data BKS (t) is calculated, and output as rotational position data.

Thereafter, the CPU 29 sequentially outputs the output data Bma (2t) and Bma of the 132-pole resolver 13 every time t elapses.
(3t),... And each output data B
., and the previously obtained output data Bma (t), Bma (2t),... are compared with R / 2, and the output data Bma is obtained from the comparison result. (2
t), Bma (3t),... corresponding to the poles TBK (2t), T
The transition of BK (3t), ... is determined. From this comparison result, a new corresponding pole is specified, and this pole and the output data B
The absolute data BaK is calculated from ma (2t), Bma (3t),. Further, the absolute data BaK is corrected with the multipolar correction data ΔBma [j] to calculate corrected position data BKS (2t), BKS (3t), and outputs the calculated corrected position data as rotational position data at the rotational position.

Next, the averaging correction data <ΔAma
The operation of the rotation angle detection device using [i]> and the multipolar correction data ΔBma [j] will be described. FIG. 1 shows a solid line F _2.
In the rotation angle θ-output data Ama characteristic shown by
[W-1], Ama [w], Ama [w + 1] are each 1
It is assumed that the rotation setting data is Ama [i]. That is, the respective one-turn setting data Ama [w-1], Ama [w], Ama [w +
1] are separated by a set number f. Then, each one-turn setting data Ama [w-1], Ama [w], Ama
The rotation angles corresponding to [w + 1] are θ (w−1) and θ, respectively.
(w) and θ (w + 1).

First, when the CPU 29 takes in the initial position data AK of the one-turn resolver 12 at the initial position θK, the CPU 29 sets the value i of the one-rotation correction data ΔAma [i] to the initial position data AK to i = [AK / f + 0.5 ]. Here, it is assumed that the value i is w.

As shown in FIG. 1, one-turn correction data Δ calculated from initial position data AK with respect to initial position θK
One rotation setting data Ama [w-1] and Ama [w +] before and after one rotation setting data Ama [w] corresponding to Ama [w].
1], the rotation angle as shown by the solid line F ₂ theta-output data A
When the rotation angle θ (w-1) or θ (w + 1) separated by the ma characteristic corresponds to one rotation correction data ΔAma [w] and ΔA
The difference between ma [w-1] or ΔAma [w] and ΔAma [w + 1] increases.

With the motor 11 in the initial position θK,
External force in the radial direction to the output shaft and is applied ,, rotation angle .theta.K - Output data Ama characteristic deviates from the solid line F _2. Then, at the initial position θK, output data AZ different from the actual initial position data AK is fetched. Then, the output data A
It is assumed that one-turn correction data ΔAma [w-1] or ΔAma [w + 1] is obtained as one-turn correction data ΔAma [i] corresponding to Z. The one-rotation correction data ΔAma [w-1] or ΔAma in which the output data Ama is not averaged
The corrected initial position data corrected and calculated in [w + 1] is AKS "[Z, w-1] or AKS" [Z, w + 1]. The corrected initial position data AKS "[Z, w-1] or AKS" [Z, w + 1] is obtained by comparing the initial position data AK corresponding to the actual initial position .theta.K with the non-averaged one-rotation correction data .DELTA.Ama [w]. From the one-rotation correction initial position data AKS "[K, w].
Here, the one-rotation correction initial position data AKS "[Z, w]
Means one rotation correction initial position data calculated from the initial position data AZ and the correction data ΔAma [w]. Therefore, the calculated corrected initial position data AKS "[Z, w-
1] or AKS "[Z, w + 1]
BK [w-1] or TBK [w + 1] is a multipolar resolver 13
It is more likely that the initial position data BK differs from the pole TBK corresponding to the initial position data BK.

In the above description, when the correction to the initial position data AK is performed using the averaged correction data <ΔAma [i]>, the output data AZ becomes the averaged correction data <ΔAma [w].
-1]> or <ΔAma [w + 1]> and the one-turn correction initial position data AKS [Z, w-1] or AKS
[Z, w + 1] is obtained. In this case, the averaged correction data <ΔAma [w]> and <ΔAma [w−1]> or <ΔAma [w−1]>
The difference between Ama [w]> and <ΔAma [w + 1]> is the difference between the correction data ΔAma [w] and ΔAma [w-1] or <ΔAma
[W]> and ΔAma [w + 1]. As a result, the calculated one rotation correction initial position data AKS [Z,
w-1] or AKS [Z, w + 1]
The difference between the one-rotation correction initial position data AKS [K, w] obtained from K and the averaged correction data ΔAma [w] is the non-averaged correction data ΔAma [w-1] or ΔA
corrected initial position data A calculated using ma [w + 1]
KS "[Z, w-1] or AKS" [Z, w + 1], one-rotation correction initial position data AKS "[K obtained from the non-averaged one-rotation correction data .DELTA.Ama [i]. , W].

Therefore, the one-rotation correction initial position data AKS
The initial pole TBK [w-1] or TBK of the 132 pole resolver 13 specified from [Z, w-1] or AKS [Z, w + 1].
[W + 1] is less likely to deviate from the initial pole TBK specified by the one-rotation correction initial position data AKS [K, w] obtained from the output data AK and the averaged correction data <ΔAma [w]>. Then, absolute data BaK is calculated from the initial pole TBK and the initial position data BK of the 132 pole resolver 13. Further, the multipolar correction initial position data BKS is calculated from the absolute data BaK and the multipolar correction data ΔBma [j], and is output as rotation initial position data. As a result, even when a radial external force is applied to the output shaft, the rotation initial position data BKS with high accuracy with respect to the initial position θK is calculated.

When a predetermined time t has elapsed, the CPU 29 takes in the output data Bma (t) of the multipolar resolver 13. Then, the CPU 29 compares the output data Bma (t) with the initial position data BK to determine the transition of the pole TBK. Then, the pole TBK (t) corresponding to the output data Bma (t) is specified from the result, and the pole TBK newly specified is specified.
(T) and the absolute data BaK (t) are calculated from the output data Bma (t), and the multipolar correction data ΔBma (t) is calculated.
Is added to obtain corrected position data. Here, even when a radial external force is applied to the output shaft, the deviation of the initial pole TBK from the initial pole TBK corresponding to the actual initial position θK obtained from the initial position data AK of the one-rotation resolver 12 is suppressed. ing. Therefore, the output data B of the multipolar resolver 13 sequentially obtained based on the initial pole TBK
The poles determined by comparing ma are also specified with high accuracy. Then, the corrected position data BKS (t) calculated from the newly specified pole TBK and the output data Bma.
Is highly accurate rotational position data.

As described in detail above, according to the method and the apparatus for generating correction data of the resolver of the present embodiment, the one-rotation correction generated for each output data Ama [i] of the one-rotation resolver 12 is performed. One-turn correction data ΔAma [i−1], ΔAma in which data ΔAma [i]
[I + 1] and averaged correction data <ΔAma
[I]> is created. As a result, the difference between the averaging correction data <ΔAma [i]> becomes one rotation correction data ΔAma
It is smaller than the difference between [i]. Therefore, the initial position data deviates from the originally obtained initial position data due to the detection error, and as a result, the averaged correction data <ΔAma [w]> adjacent to the averaged correction data <ΔAma [w]> to be used for the original initial position data ΔAma [w-1]> or <ΔAma [w
+1]>, the initial position data AK is corrected.
The calculated one-turn correction initial position data AKS [Z, w-
It is possible to create one-rotation correction data that can suppress an increase in an error due to correction to the actual one-rotation correction initial position data AKS [K, w] of [1] or [Z, w + 1].

Further, according to the rotation angle detecting method and the rotation angle detecting device of the present embodiment, the initial position data AK of the one-rotation resolver 12 is corrected by the averaging correction data <ΔAma [i]> to perform one-rotation correction. Initial position data AKS is created, and the initial pole TBK of the 132-pole resolver 13 is specified by the one-rotation corrected initial position data AKS. Therefore, even when the external force in the radial direction is applied to the output shaft, even if the output data AZ deviated from the output data AK when the averaging correction data <ΔAma [i]> is generated at the initial position θK, this output data AZ Is corrected by averaging correction data <ΔAma [w-1]> or <ΔAma [w-1]> to generate one-turn correction initial position data AKS [Z, w-1] or AKS [Z, w +
1], the initial pole TBK of the 132 pole resolver 13
Is unlikely to be deviated from the initial pole TBK corresponding to the actual initial position θK. Further, the rotational position data BKS calculated from the output data BK and Bma of the initial pole TBK and the 132 pole resolver 12 is less likely to be shifted from the actual rotational position. Therefore, even when a radial external force is applied to the output shaft of the motor 11, highly accurate rotational position data can be obtained.

The present invention is not limited to the above-described embodiment, but may be configured as follows without departing from the spirit of the invention. (1) In the above embodiment, the present invention was carried out in the resolver including the one-rotation resolver 12 and the multipolar resolver 13.
In a resolver including the rotating resolver 12 and a plurality of multipolar resolvers having different numbers of poles, the one-rotation correction data ΔAma [i] of the one-rotation resolver 12 may be used as the averaging correction data <ΔAma [i]>.

The number of poles of the multipolar resolver 13 may be other than 132, such as 120 or 150. (2) In the above embodiment, the one-turn correction data ΔAma
[I] is one rotation correction data ΔAma for each one before and after it.
[I-1] and [Delta] Ama [i + 1] are averaged, but may be averaged with two preceding and succeeding one-turn correction data [Delta] Ama [i], or three or more one-turn correction data [Delta] Ama [i]. You may average.

In the above embodiment, the one-turn correction data Δ
Ama [i] is the same number of one rotation correction data Δ
Although Ama [i-1] and ΔAma [i + 1] are averaged, they may be averaged with different numbers of one-turn correction data before and after. That is, for example, the one-turn correction data ΔA following the rear side
Data obtained by averaging ma [i + 1] and ΔAma [i + 2] may be used as averaged correction data <ΔAma [i]>. (3) In the above embodiment, one-turn correction data ΔAma [i
-1], ΔAma [i], ΔAma [i + 1]
Although the arithmetic mean (arithmetic mean) was used, the averaged correction data <ΔAma [i]> obtained by averaging the results with geometric mean (geometric mean) or harmonic mean may be used.

[0103]

As described in detail above, according to the first aspect of the present invention, even if an error occurs in the initial position data detected at the initial position, the expansion of the error due to the correction to the initial position data is suppressed. Correction data that can be created.

According to the second and third aspects of the present invention,
Even if an error occurs in the initial position data detected at the initial position, it is possible to suppress the expansion of the error due to the correction of the initial position data and detect the rotation angle with high accuracy.

[Brief description of the drawings]

FIG. 1 is a graph for explaining creation of correction data for output data of a one-rotation resolver as one embodiment of the present embodiment and its operation.

FIG. 2 is a longitudinal sectional view showing a direct drive motor in which a resolver is incorporated.

FIG. 3 is a block diagram illustrating a correction data creation device.

FIG. 4 is a graph showing output signals of a single-rotation resolver and a multi-pole resolver.

FIG. 5 is a graph showing RD converter conversion output data of a one-rotation correction resolver and a multipolar resolver, and output data of a reference encoder.

FIG. 6 is a block diagram showing a product circuit block of the direct drive motor.

FIG. 7 is a graph showing RD converter conversion output data of a magnetically affected one-turn resolver.

FIG. 8 is a graph obtained by enlarging a part of the graph of FIG. 7;

[Explanation of symbols]

12: one-rotation resolver, 13: multi-pole resolver, 16: reference encoder, 22: PROM as storage means, 27
... Reference data creating means, one-rotation correction data creating means, multi-pole reference data creating means, multi-pole correction data creating means and CPU as averaging correction data creating means, 29. CPU as indexing means, initial position absolute data calculating means, multi-pole corrected initial position data calculating means, pole determining means, absolute data calculating means and corrected position data calculating means, Ama [i]... 1 rotation setting data, Bma [ j] ...
Multi-pole setting data, EA [i], EB [j]: Reference data, h · f: Absolute data, ΔAma [i], Δ
Ama [i-1], [Delta] Ama [i + 1] ... one rotation correction data, Ip [j] ... multipole reference data, [Delta] Bma [j] ... multipole correction data, <[Delta] Ama [i]> ... averaging correction data, T
BK: initial pole, ΔBma (t): output data, TBK (t):
Pole, AKS: One-rotation correction initial position data, BaK: Initial position absolute data, BKS: Multi-pole correction initial position data, BaK (t): Absolute data, BKS (t) ...
Correction position data.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁶ Identification code Agency reference number FI Technical display location H02K 24/00 H02K 24/00

Claims (3)

(57) [Claims] 1. A reference encoder (16) for each one-turn setting data (Ama [i]) obtained from a one-turn detection resolver (12) and a multi-pole resolver. The reference data (EB [j]) of the reference encoder (16) for each multi-pole setting data (Bma [j]) obtained from (13) is obtained, and for each multi-pole setting data (Bma [j]), Absolute data (h · f) obtained by converting the multi-pole setting data (Bma [j]) into absolute data is created, and the multi-pole setting data (Bma [j]) corresponding to each one-turn setting data (Ama [i]) is obtained. The reference encoder (16)
From the respective reference data (EA [i], EB [j]), and the multi-pole setting data (Ama [i]) corresponding to each one-rotation setting data (Ama [i]). B
ma [j]) and the one-turn correction data (ΔAma) of the one-turn setting data (Ama [i]) with respect to the absolute data (h · f) from the absolute data (h · f)
[I]), and from the reference data (EB [j]) of the reference encoder (16) for each of the multi-pole setting data (Bma [j]), the multi-pole for the multi-pole setting data (Bma [j]) The reference data (Ip [j]) is obtained, and the multipolar reference data (Ip [j]) for each multipolar setting data (Bma [j]) and the multipolar setting data (Bma [j])
Absolute data (j ·
f), the multi-pole correction data (ΔBma [j]) is generated from the correction data generation method in the resolver. ΔAma [i-
1], ΔAma [i + 1]), and averaged correction data (<ΔAma [i]>) is created. A method of creating correction data in a resolver as one rotation correction data for one rotation measurement data (Ama [i]). 2. An initial position data (AK) from an initial position data (AK) obtained from a one-rotation detection resolver (12).
Obtains the initial pole (TBK) of the corresponding multipole resolver (13), obtains the rotation initial position data from the initial pole (TBK) and the initial position data (BK) obtained from the multipole resolver (13), and sequentially (1) of the obtained multipolar resolver
3) From the output data (Bma (t)),
A rotation angle detection method for determining a pole (TBK (t)) corresponding to ma (t) and obtaining rotational position data from the pole (TBK (t)) and output data (Bma (t)). Averaging correction data (<ΔAma [i]>) for each one-rotation setting data (Ama [i]) of the one-rotation detection resolver (12) created by the method for creating correction data in the resolver described in “1. The multipole correction data (ΔBma [j]) for each multipole setting data (Bma [j]) of the resolver (13) is stored, and the initial position data (AK) of the one-rotation detection resolver (12) is stored.
Is obtained, and the initial position data (AK) is corrected with the averaged correction data (<ΔAma [i]>) to obtain the one-rotation corrected initial position data (<ΔAma [i]>). AKS), and calculates an initial pole (TBK) of the multipolar resolver (13) from the one-rotation corrected initial position data (AKS) and the initial position data (AK) of the multipolar resolver (13). Initial position absolute data (BaK) obtained by converting the initial position data (AK) from the initial position data (AK) of the multi-pole resolver (13) and the initial position data (AK).
Is calculated, multipolar correction data (ΔBma [j]) corresponding to the initial position absolute data (BaK) is obtained, and the initial position absolute data (BaK) is converted to the multipolar correction data (ΔBma).
ma [j]) and multi-pole corrected initial position data (BK
S) is calculated and output as rotation initial position data. The multipole output data (Bma (t)) obtained sequentially is compared with the multipole output data (BK) obtained before, and the multipole output data is obtained. The multipolar resolver (1) to which the data (Bma (t)) corresponds
3) The pole (TBK (t)) is determined, and the multipole output data (Bma (t)) and the pole (T
BK (t)) and the multi-pole output data (Bma
(T)) is calculated as absolute data (BaK (t)), multipolar correction data (ΔBma [j]) for the absolute data (BaK (t)) is obtained, and the absolute data (BaK (t) is obtained. )) To the multipolar correction data (ΔBma
[J]), a rotation angle detection method in a resolver that calculates corrected position data (BKS (t)) and outputs the calculated corrected position data as rotational position data. 3. A rotation angle detection device comprising a one-rotation detection resolver (12) and a multi-pole resolver (13), wherein the one-rotation detection resolver created by the method for creating correction data in the resolver according to claim 1. The averaging correction data (<ΔAma [i]>) for each one-rotation setting data (Ama [i]) in (12) and the multiplication correction data (Bma [j]) for each multipolar setting data (Bma [j]) in the multipolar resolver (13). Storage means (22) for storing the pole correction data (ΔBma [j]); and initial position data (AK) of the one-rotation detection resolver (12).
Is obtained, and the initial position data (AK) is corrected with the averaged correction data (<ΔAma [i]>) to obtain the one-rotation corrected initial position data (<ΔAma [i]>). A one-rotation correction initial position data calculating means (29) for calculating AKS); a multi-polar resolver (13) based on the one-rotation correction initial position data (AKS) and the initial position data (BK) of the multi-polar resolver (13). An initial pole determining means (29) for calculating an initial pole (TBK) of the above, and an initial position obtained by absolutely converting the initial position data (BK) from the initial pole (TBK) and the initial position data (BK) of the multi-pole resolver (13). Position absolute data (BaK)
Position absolute data calculation means (2)
9) and multi-polarity correction data (ΔBma [j]) corresponding to the initial position absolute data (BaK) are obtained, and the initial position absolute data (BaK) is converted to the multi-polarity correction data (ΔB
ma [j]) and multi-pole corrected initial position data (BK
S) and calculates multipolar correction initial position data calculating means (29) for outputting as rotation initial position data; multipolar output data (Bma (t)) obtained sequentially and multipolar output data previously obtained (BK), the multipolar output data (Bma (t)) corresponds to the corresponding multipolar resolver (1).
3) Pole determining means (2) for determining the pole (TBK (t))
9) and the multipole output data (Bma (t)) and the pole (T
BK (t)) and the multi-pole output data (Bma
(T)) absolute data calculation means (29) for calculating absolute data (BaK (t)), and multipolar correction data (ΔBma [j]) for the absolute data (BaK (t)). , The absolute data (BaK (t)) and the multipolar correction data (ΔBma
[J]), a corrected position data calculating means (29) for calculating corrected position data (BKS (t)) and outputting it as rotational position data.