ROTARY ENCODER

Abstract

A rotary encoder includes: a rotary disk with an angle code; a light source; a detector reading the angle code; and a processing unit acquiring a reading value. The light source includes at least two light-emitting elements spaced from each other. Every time the rotary disk is rotated by a predetermined angle, where an arbitrary angle from a rotation angle θ within a reading range on the detector is provided as φ, the processing unit acquires reading values fI(θ+φ) and fI(θ) with a first light-emitting element and a reading value fII(θ+φ) with a second light-emitting element, to calculate a reading value error due to deflection at an angle θ+φ based on the difference between the reading values fII(θ+φ) and fI(θ+φ), to obtain a difference gI(θ, φ) between the reading values fI(θ+φ) and fI(θ) such that the error is reflected, and to self-calibrate based on a change in the difference gI(θ, φ).

Description

CROSS-REFERENCE TO RELATED APPLICATION(S)

The present application claims priority under 35 U.S.C. § 119 to Japanese Patent Application No. 2018-185902 filed Sep. 28, 2018. The contents of this application are incorporated herein by reference in their entirely.

TECHNICAL FIELD

The present disclosure relates to a rotary encoder, and more particularly, to a rotary encoder capable of self-calibration.

BACKGROUND ART

A rotary encoder used for a surveying apparatus measures a horizontal angle and a vertical angle by using a rotary disk that rotates in conjunction with a telescope and a detector that reads a slit indicative of a scale or an angle code provided along the circumferential direction of this rotary disk.

The main causes of errors in measured values of the rotary encoder include the eccentricity of attachment of the rotary disk to a rotating shaft, and the scale or angle codes not attached at completely equal intervals.

To remove such an error, in the rotary encoder disclosed in Patent Document 1, every time the rotary disk is rotated by a predetermined angle, where a rotation angle of the rotary disk is provided as θ and an arbitrary angle from the rotation angle θ within a reading range on the detector is provided as φ, reading values f(θ+φ) and f(θ) within the reading range on the detector can be acquired, and the self-calibration of the reading value f(θ) can be carried out based on a change in a difference g(θ, φ) between the reading values f(θ+φ) and f(θ) when the rotary disk is rotated one turn.

RELATED ART DOCUMENT

Patent Document

Patent Document 1: Japanese Patent No. 4824415

SUMMARY OF INVENTION

Problems to be Solved by the Invention

However, in the rotary encoder of Patent Document 1, no consideration is given to an error caused by deflecting of the rotary disk toward and away from a light source during rotation (hereinafter referred to as “deflection”), and particularly, in the case of a rotary encoder for a surveying apparatus required to perform high precision measurement, the rotary encoder has a problem that calibration accuracy may not be sufficient.

The present invention has been made in view of the situations, and an object thereof is to provide a rotary encoder capable of carrying out highly-accurate self-calibration with consideration given to an error caused by deflection of a rotary disk.

Means for Solving the Problems

To achieve the aforementioned object, an aspect of the present invention provides a rotary encoder including: a rotary disk with an angle code; a light source for irradiating the angle code; a detector for reading the angle code; and a processing for acquiring a reading value f(θ) based on the angle code read by the detector. The light source includes a plurality of light-emitting elements, wherein a first light-emitting element is disposed at a position corresponding to an angle θ of the rotation, wherein a second light-emitting element is disposed at a predetermined distance x_L away from the first light-emitting element. Every time the rotary disk is rotated by a predetermined angle, where a rotation angle of the rotary disk is provided as θ and an arbitrary angle from the rotation angle θ within a reading range on the detector is provided as φ, the processing unit is configured to acquire reading values f_I(θ+φ) and f_I(θ) within a reading range on the detector in the case of using the first light-emitting element and a reading value f_II(θ+φ) on the detector in the case of using the second light-emitting element, to calculate a reading value error Δφ_I due to deflection at an angle θ+φ based on a difference h(θ+φ) between the reading value f_II(θ+φ) and the reading value f_I(θ+φ), to obtain a difference g(θ, φ) between the reading values f_I(θ+φ) and f_I(θ) such that the reading value error Δφ_I is reflected, and to self-calibrate the reading value f_I(θ) based on a change in the difference g_I(θ, φ).

In another aspect, preferably, the light source includes a third light-emitting element disposed at a position symmetrical to the second light-emitting element with respect to the first light-emitting element. The processing unit is configured to acquire a reading value f_I(θ−φ) within a reading range on the detector in the case of using the first light-emitting element and a reading value f_III(θ−φ) on the detector in the case of using the third light-emitting element, to calculate a reading value error Δ(−φ_I) due to deflection at an angle θ−φ based on a difference h(θ−φ) between the reading value f_III(θ−φ) and the reading value f_I(θ−φ), to obtain a difference g_I(θ, −φ) between the reading values f_I(θ−φ) and f_I(θ) such that the reading value error Δ(−φ_I) is reflected, to average changes in the difference g_I(θ, φ) and the difference g_I(θ, −φ), and to self-calibrate the reading value f_I(θ).

Benefit of Invention

According to the aspect, an object is to provide the rotary encoder capable of carrying out highly-accurate self-calibration with consideration given to an error caused by deflection of a rotary disk.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a configuration block diagram of a rotary encoder according to a first embodiment of the present invention.

FIG. 2A is a plan view of the vicinity of a detector of the rotary encoder according to the first embodiment.

FIG. 2B is a cross-sectional view taken along a line IIB-IIB of FIG. 2A.

FIG. 3A is a plane view for explaining a method of calculating an error due to a deflection in the rotary encoder of the first embodiment when a first light-emitting element emits light.

FIG. 3B is a cross-sectional view taken along a line IIIB-IIIB of FIG. 3A.

FIG. 4 is a view for explaining the method of calculating an error due to a deflection in the rotary encoder of the first embodiment and is a cross-sectional view similar to FIG. 3B when a second light-emitting element emits light.

FIG. 5 is a flowchart illustrating self-calibration procedures of the rotary encoder of the first embodiment.

FIG. 6A is a plan view of the vicinity of a detector of a rotary encoder according to a second embodiment of the present invention

FIG. 6B is a cross-sectional view taken along a line VIB-VIB of FIG. 6A.

FIG. 7 is a flowchart illustrating self-calibration procedures of the rotary encoder of the second embodiment.

DESCRIPTION OF EMBODIMENTS

Preferred embodiments of the present invention will now be described with reference to the drawings; however, the present invention is not limited thereto. In the embodiments, the same constituent elements are denoted by the same reference numerals and will not redundantly be described.

First Embodiment

1. Configuration of Rotary Encoder

A rotary encoder (hereinafter simply referred to as an “encoder”) 100 according to this embodiment is a so-called absolute encoder. The encoder 100 includes a rotary disk 2 fixed to a rotating shaft (not depicted), a light source 3 fixed to a rotating shaft bearing (not depicted), a detector 4 receiving the light from the light source 3 via the rotary disk 2, a converting unit 5, a processing unit 6, and a display unit 7.

The rotary disk 2 is a doughnut-shaped disk and has a periphery portion provided with slits 21 composed of narrow slits that mean 0 and wide slits that mean 1 and forming angle codes indicative of angles.

The light source 3 includes a first light-emitting element 31 and a second light-emitting element 32, and the light-emitting elements are LEDs, for example. The light-emitting elements 31, 32 are disposed to face the slits 21 of the rotary disk 2 at a predetermined distance H_DL (FIG. 2B). The first light-emitting element 31 and the second light-emitting element 32 are disposed at a predetermined distance x_L in a direction orthogonal to the diameter of the rotary disk.

The detector 4 is a CCD linear sensor, for example, and is disposed on the side opposite to the light source 3 with respect to the rotary disk 2 so as to face the light source 3 at a predetermined distance H_DC (FIG. 2B).

The converting unit 5 is an A/D converter and converts a detection signal from the detector 4 into a digital signal to output the signal to the processing unit 6.

The processing unit 6 is a microcontroller equipped with a ROM, a RAM, etc. for a CPU, controls light emission of the light source 3, decodes an angle code from the detection signal input from the converting unit 5, applies a known interpolation process to calculate an angle, and outputs the angle to the display unit 7 such as a liquid crystal display.

As depicted in FIGS. 2A and 2B, the detector 4 is disposed at a position orthogonal to the diameter of the rotary disk 2 to face the slits 21. A foot of a perpendicular line drawn from a center O of the rotary disk 2 to the detector 4 is provided as the point A. In addition, an appropriate point of the rotary disk 2 within a detection range of the detector 4 is provided as the point B. Reference sign M denotes a direction in which the angle code indicates 0°, and a true angle of the point A created from an M direction is provided as θ, while a true angle of an angle AOB is provided as φ. Therefore, the position of the point A is the position of the angle θ, and the position of the point B is the position of the angle θ+φ.

In this state, the first light-emitting element 31 is disposed at a position corresponding to the point A, and the second light-emitting element 32 is disposed at the predetermined distance x_L away from the first light-emitting element 31 in an extending direction of the detector 4.

As depicted in FIG. 2B, if deflection is not taken into consideration, light emitted from the first light-emitting element 31 and made incident on the point A is incident on a reading position A_CI of the detector 4. The light made incident on the point B is incident on a reading position B_CI of the detector 4. The processing unit 6 can read respective angle codes of the point A and point B at the reading positions A_CI and B_CI and can apply the interpolation process to determine angle reading values f_I(θ) and f_I(θ+φ) at the point A and point B, respectively. In normal angle measurement, the angle reading value f_I(θ) at the point A is the measurement value.

Since the reading values f_I(θ) and f_I(θ+φ) are periodic functions with a period of 360° and can therefore be expressed by formulas 1 and 2 by using Fourier series.

f _I(θ)=θ+ΣA_i*sin(iθ+α_i)  (1)

f _I(θ+φ)=(θ+φ)+ΣA_i*sin{i(θ+φ)+α_i}  (2)

(where i means a natural number (1, 2, 3, . . . M/2) expressing the order of a higher harmonic, Ai means amplitude of the i-th higher harmonic, α_i means an initial phase of the i-th higher harmonic, and M means a scale number, which is an even number).

As depicted in FIG. 2B, if deflection is not taken into consideration, light emitted from the second light-emitting element 32 and made incident on the point B is incident on a reading position B_CII of the detector 4. The processing unit 6 reads the angle code of the point B with the detector 4 and applies the interpolation process to determine an angle reading value f_II(θ+φ) at the point B. The details of the reading value f_II(θ+φ) will be described later.

2. Calculation of Deflection Error

An influence of deflection on an angle reading value will be described with reference to FIGS. 3A and 3B. The deflection means displacement of the rotary disk 2 toward and away from the light source 3 due to rotation. In FIGS. 3A and 3B, description will be made of the case that the deflection occurs in the direction toward the light source.

When the deflection ΔH occurs, the incident position on the rotary disk 2 of the light emitted from the first light-emitting element 31 for detection of the angle θ+φ is a point B′ shifted from the point B by a distance Δx_DI. Therefore, the angle code of the point B′ is read instead of that of the point B at the reading position B_CI. Thus, the reading value f_I(θ+φ) is (θ+φ)+Δφ_I (FIG. 3A). Therefore, −Δφ_I is an error caused by the deflection of the rotary disk.

The deflection error Δφ_I at the position of the angle θ+φ can be calculated as follows.

From FIG. 3B, a relationship is as expressed by formula 3 among the deflection ΔH, a deviation Δx_DI of the incident position due to the deflection ΔH, a distance H_LC from the first light-emitting element 31 to the detector 4, and a distance x_CI between the reading positions A_CI, B_CI.

x _CI /H _LC =Δx _DI /ΔH  (3)

Therefore, the deviation Δx_DI of the incident position can be expressed by formula 4.

Δx_DI=ΔH/H_LC·x_CI  (4)

From FIG. 3A, the tangent of the angle φ can be expressed by formula 5 by using a distance x_D between A and B and a length l of a perpendicular line drawn from the center O of the rotary disk to the point A.

tan φ=x_D/l  (5)

The tangent of (φ+Δφ_I) is expressed as formula 6 by using the deviation Δx_DI of the incident position.

tan(φ+Δφ_I)=(x_D−Δx_DI)/l  (6)

Therefore, if Δφ_I is sufficiently small, Δφ_I can be approximated as in formula 7.

Δφ_I≈−(Δx_DI)/1=−(ΔH/H_LC)(x_CI/1)  (7)

The deflection is also a periodic function with a period of 360° and is therefore expressed by formula 8 by using Fourier series.

Δφ_I=Σ(ΔH_i/H_LC)(x_CI/l)*sin{i(θ+φ)+βi}  (8)

By adding this formula to formula 1, a measured angle value f_I(θ+φ) at the point B (angle θ+φ) with consideration given to the deflection can be expressed by formula 9.

f _I(θ+φ)=θ+φ+ΣA_i*sin{i(θ+φ)+α_i}+Σ(ΔH_i/H_LC)(x_CI/l)*sin{i(θ+φ)+β_i}  (9)

In the case of φ=0, x_CI=0 is satisfied, and an error due to the deflection does not occur as expressed by formula 10. Specifically, the error due to the deflection does not affect the measurement of the angle θ itself.

f _I(θ)=θ+ΣA_i*sin(iθ+α_i)  (10)

Description will be made of the influence of the deflection ΔH in the case of obtaining the angle reading value f_II(θ+φ) at the point B_CII when the light is emitted by the second light-emitting element 32. The incident position on the rotary disk 2 of the light emitted from the second light-emitting element 32 is a point B″ shifted from the point B by a distance Δx_DII.

A relationship at a position of θ+φ can be expressed by formula 11 from FIG. 4 among the distance x_L between the first light-emitting element 31 and the second light-emitting element 32, a distance (angle reading position) x_CII between the reading positions A_CI, B_CII, the distance H_LC from the second light-emitting element 32 to the detector 4, a deviation Δx_DII of the incident position due to the deflection ΔH, and the deflection ΔH.

(x_L−x_CII)/H_LC=−Δx_DII/Δ_H  (11)

Therefore, the deviation Δx_DII of the incident position can be expressed by formula 12.

Δx_DII=(ΔH/H_LC)(x_CII−x_L)  (12)

Thus, as in the case that the light is emitted by the first light-emitting element 31, the angle reading value f_II(θ+φ) of the point B_CII at the position of θ+φ can be expressed by formula 13.

f _II(θ+φ)=−θ+φ+ΣA_i*sin{i(θ+φ)+α_i}+E(ΔH_i/H_LC)(x_CII−x_L)/l*sin{i(θ+φ)+β_i}  (13)

Therefore, a difference h(θ+φ) between the reading value f_II(θ+φ) according to the second light-emitting element 32 and the reading value f_I(θ+φ) according to the first light-emitting element 31 can be expressed by formula 14.

h(θ+φ)=f_II(θ+φ)−f_I(θ+φ)=Σ(ΔH_i/H_LC)(x_CII−x_CI−x_L)/l*sin{i(θ+φ)+β_i}  (14)

From formula 8, Σ(ΔH_i/H_LC)*{sin{i (θ+φ)+β_i}=Δφ_I/(x_CI/l) is acquired, and therefore, the error Δφ_I of the angle reading value due to the deflection at the position of θ+φ can be expressed by formula 15.

Δφ_I=x_CI/(x_CII−x_CI−x_L)*h(θ+φ)  (15)

From the above, an angle reading value taking an influence of an error into consideration is obtained by f_I(θ+φ)−Δφ_I.

Since the distance x_L between the first light-emitting element 31 and the second light-emitting element 32 is known, and the distance (angle reading position) x_CI between the reading positions A_CI, B_CI, and the distance (angle reading position) x_CII between the reading positions A_CI, B_CII can also be acquired, the error Δφ_I of the angle reading value due to the deflection at the position of the angle θ+φ can be obtained.

3. Method of Self Calibration

A method of self-calibration of the encoder 100 according to this embodiment is the same as the method disclosed in Patent Document 1 except that the error Δφ_I of the angle reading value due to the deflection is taken into consideration when the reading value f_I(θ+φ) at the position of θ+φ is acquired, and the outline is as follows.

First, the difference g_I(θ, φ) between the reading values f_I(θ+φ) and f_I(θ) can be expressed by a Fourier series as in formula 16 from formulas 1 and 2 according to the method described in Patent Document 1.

g _I(θ,φ)=φ+ΣA_i*sin(iθ/2)*cos{i(θ+φ/2)+α_i}=B₀+ΣB_i*sin{iθ+γ_i}  (16)

where

φ=B_o  (17)

A _i =B _i*sin(_iφ/2)  (18)

α_i=γ_i−_iφ/2+π/2  (19).

Therefore, at a position with a rotation angle θj where the rotary disk 2 is rotated by a predetermined angle (e.g., 5°) obtained by dividing 360° into N equal parts, reading values f_I(θ_j), f_I(θ_j+φ) at points A and B are acquired. Here, j=1, 2, 3, . . . N−1. By calculating g_I(θ_j, φ)=f_I(θ_j+φ)−f_I(θ_j) for each predetermined angle θ_j, the respective coefficients B_o, B_i, γ_i of the Fourier series in formula 16 can be calculated by a known method, and φ, A_i, α_i can be calculated from formulas 17, 18, 19. Therefore, the true angle θ can be obtained from formula 20, which is deformation of formula 1.

θ=f(θ)−ΣA_i*sin(iθ+α_i)=f(θ)−E(θ)  (20)

Here, E(θ) means an error function E(θ)=ΣA_i*sin(iθ+α_i).

4. Self-Calibration Procedures

Self-calibration procedures of the encoder 100 according to this embodiment will be described with reference to FIG. 5. The following automatic calibration process is usually carried out by the processing unit 6.

When the self-calibration is started, in step S101, the light is emitted by the first light-emitting element 31 so as to acquire the reading values f_I(θ_j), f_I(θ_j+φ) of the reading positions A_CI, B_CI on the detector 4 and the distance (angle reading position) x_CI between A_CI and B_CI, and this is carried out a predetermined number of times. After the measurement, the light emission of the first light-emitting element 31 is stopped.

Next, in step S102, the light is emitted by the second light-emitting element 32 so as to acquire the reading value f_II(θ_j+φ) of the reading position B_CII on the detector 4 and the distance (angle reading position) x_CII between A_CI and B_CII, and this is carried out a predetermined number of times. After the measurement, the light emission of the second light-emitting element 32 is stopped.

Next, in step S103, averages of the reading values f_I(θ_j), f_I(θ_j+φ), f_II(θ_j+φ), the distance (angle reading position) x_CI between A_CI and B_CI, and the distance (angle reading position) x_CII between A_CI and B_CII are calculated.

Next, in step S104, a difference h(θ_j+φ)=f_II(θ_j+φ)−f_I(θ_j+φ) between the reading values f_II(θ_j+φ) and f_I(θ_j+φ) at θ+φ is calculated and stored.

Next, in step S105, the error Δφ_I of the angle reading value due to the deflection at the position of θ+φ is calculated by formula 15 and stored.

Next, in step S106, a difference g_I(θ_j, φ) between the reading value f_I(θ_j+φ)−Δφ_I and the reading value f_I(θ) taking the influence of the deflection into consideration is calculated and stored.

Next, in step S107, it is judged whether all possible data has been acquired, i.e., whether g(θ_j, φ) has been acquired for each predetermined angle θ_j. For example, it is judged for θ whether g_I(θ_j, φ) has been acquired at intervals of 5° traversing 360°.

If all possible data has not been acquired (No), the process proceeds to step S107, the rotary disk 2 is manually or automatically rotated by a predetermined angle, and the process returns to step S101. Thereafter, steps S101 to S107 are repeated to obtain and store g_I(θ_j, φ) for each predetermined angle θ_j.

When the predetermined number of times of data acquisition has been completed in step S109 (Yes), the process proceeds to step S108 to calculate, based on g_I(θ_j, φ) for each predetermined angle θ_j, coefficients B₀, B_n, γ_n of the Fourier series of g_I(θ, φ) described in formula 16.

Next, in step S110, coefficients A_n, α_n of the Fourier series of the error function E(θ) of the reading value f(θ) are calculated based on the coefficients B₀, B_n, γ_n.

Next, in step S111, the error function E(θ) is obtained based on the coefficients acquired in step S110 and is stored, and the process ends.

By carrying out the automatic calibration of the encoder 100 as described above, the true angle θ can be accurately obtained from the reading value f(θ) by using formula 1 in subsequent measurements.

According to the encoder 100 according to this embodiment, when a rotation angle of the rotary disk is provided as θ and an arbitrary angle from the rotation angle θ within a reading range on the detector is provided as φ, the second light-emitting element is disposed at a predetermined distance in addition to the first light-emitting element for measuring the rotation angle θ, and this enables measurement of the angle reading value f_I(θ+φ) using the first light-emitting element 31 and the angle reading value f_II(θ+φ) using the second light-emitting element 32 at the position of the angle θ+φ as well as the respective angle reading positions on the detectors 4. Using these measured values and the known distance between the light-emitting elements makes it possible to obtain the error Δφ_I of the angle reading value at the angle θ+φ due to the deflection at the position of the angle θ+φ.

Therefore, the error function can be calculated by using the reading value f_I(θ_j+φ)−Δφ_I taking the error Δφ_I of the angle reading value due to the deflection, and the self-calibration can be carried out with this error function, which enables highly accurate self-calibration.

Second Embodiment

FIGS. 6A and 6B depict an encoder 100a according to a second embodiment. The encoder 100a has substantially the same configuration as the encoder 100 according to the first embodiment except that a light source 3a further includes a third light-emitting element 33. The third light-emitting element 33 is disposed symmetrically to the second light-emitting element 32 with respect to the first light-emitting element 31, that is, a perpendicular line drawn from the center O of the rotary disk 2 to the point A.

The third light-emitting element 33 is used for reading an angle at a position separated from the point A by the angle φ in the direction opposite to the point B, that is, at a position of an angle θ−φ. An error Δ(−φ_I) of an angle reading value due to the deflection at the angle θ−φ is also calculated by the same method as Δφ_I.

Next, self-calibration procedures of the encoder 100a according to this embodiment will be described with reference to FIG. 7. The basic process is the same as the procedures according to the first embodiment, and therefore, the same process will be cited without description.

When the self-calibration is started, in step S201, in addition to the process of step S101, an angle reading value f_I(θ_j−φ) and a distance (angle reading position) x_CI′ between A_CI and C_CI are acquired, and this is carried out a predetermined number of times. After the measurement, the light emission of the first light-emitting element 31 is stopped.

Next, in step S202, the measurement according to the second light-emitting element is carried out as in step S102.

Next, in step S203, the light is emitted by the third light-emitting element 33 so as to acquire a reading value f_III(θ_j−φ) of a reading position B_CIII on the detector 4 and a distance (angle reading position) x_CIII between A_CI and B_CIII, and this is carried out a predetermined number of times. After the measurement, the light emission of the third light-emitting element 33 is stopped.

Next, in step S204, averages of the reading values f_I(θ_j), f_I(θ_j+φ), f_II(θ_j+φ), f_III(θ_j−φ) and the angle reading positions x_CI, x_CI′ acquired in steps S201 to S203 are calculated.

Next, in step S205, h(θ_j+φ) and h(θ_j−φ) are calculated and stored as in step S104.

Next, in step S206, the errors Δφ_I, Δ(−φ_I) of the angle reading values due to the deflection at the respective positions of θ+φ, θ−φ are calculated and stored as in step S105.

Next, in step S207, g_I(θ_j, φ) and g_I(θ_j, φ) are calculated and stored as in step S106.

Thereafter, the processes in steps S208 to S211 are carried out as in steps S107 to S110 to calculate an error function of f_I(θ) obtained from each of θ+φ and θ−φ.

The error function is averaged in step S212, the error function is stored in step S213, and the process ends.

According to this embodiment, since the reading errors of the reading values at the angles θ+φ and θ−φ can be averaged, higher calibration accuracy can be achieved.

Although the preferred embodiments of the present invention have been described, the embodiments are examples of the present invention and can be combined based on the knowledge of those skilled in the art, and such embodiments fall within the scope of the present invention.

DESCRIPTION OF SYMBOLS

• 2 rotary disk
• 3, 3a light source
• 31 first light-emitting element
• 32 second light-emitting element
• 33 third light-emitting element
• 4 detector
• 6 processing unit
• 100, 100a rotary encoder

Claims

1. A rotary encoder comprising: a rotary disk with an angle code;a light source for irradiating the angle code;a detector for reading the angle code; anda processing unit for acquiring a reading value f(θ) based on the angle code read by the detector, whereinthe light source includes a plurality of light-emitting elements, wherein a first light-emitting element is disposed at a position corresponding to an angle θ of the rotation, wherein a second light-emitting element is disposed at a predetermined distance xL away from the first light-emitting element, and whereinevery time the rotary disk is rotated by a predetermined angle, where a rotation angle of the rotary disk is provided as θ and an arbitrary angle from the rotation angle θ within a reading range on the detector is provided as φ, the processing unit is configured to acquire reading values fI(θ+φ) and fI(θ) within a reading range on the detector in the case of using the first light-emitting element and a reading value fII(θ+φ) on the detector in the case of using the second light-emitting element, to calculate a reading value error ΔφI due to deflection at an angle θ+φ based on a difference h(θ+φ) between the reading value fII(θ+φ) and the reading value fI(θ+φ), to obtain a difference gI(θ, φ) between the reading values fI(θ+φ) and fI(θ) such that the reading value error ΔφI is reflected, and to self-calibrate the reading value fI(θ) based on a change in the difference gI(θ, φ).

2. The rotary encoder according to claim 1, wherein the light source includes a third light-emitting element disposed at a position symmetrical to the second light-emitting element with respect to the first light-emitting element, and whereinthe processing unit is configured to acquire a reading value fI(θ−φ) within a reading range on the detector in the case of using the first light-emitting element and a reading value fIII(θ−φ) on the detector in the case of using the third light-emitting element, to calculate a reading value error Δ(−φI) due to deflection at an angle θ−φ based on a difference h(θ−φ) between the reading value fIII(θ−φ) and the reading value fI(θ−φ), to obtain a difference gI(θ, −φ) between the reading values fI(θ−φ) and fI(θ) such that the reading value error Δ(−φI) is reflected, to average changes in the difference gI(θ, φ) and the difference gI(θ, −φ), and to self-calibrate the reading value fI(θ).