JPH02298817A - Rotary encoder - Google Patents

Abstract

PURPOSE:To realize an encoder being free from deterioration of precision in detection even when there is an error in fitting of a scale plate, by a method wherein two diffracted lights are made to share an optical path substantially and an optical system wherein first and second positions are set in the relation of conjugation is provided in the optical path. CONSTITUTION:A laser light of a light source 1 is divided into two light fluxes by a polarizing BS 41 and these fluxes are made to fall on a first point M1 of a disk plate 5 through a mirror 60 and a 1/4 wave plate 71. Then, plus and minus primary diffracted lights out of reflected diffracted lights in a plurality are superposed by the polarizing BS 41 and made to fall on a second point M2 through a lens 12, mirrors 61 and 62, a lens 13, a mirror 63 and a 1/4 wave plate 72. On the occasion, the points M1 and M2 are set in the relation of conjugation by the lenses 12 and 13 and the mirrors 61 and 62. Since the diffracted lights fall on the point M2 even when there is a positional slippage in fitting of a scale plate, according to this constitution, precision in measurement is not deteriorated and desired signals can be taken out of photodetectors S1 and S2.

Description

[Detailed Description of the Invention] [Technical Field] The present invention relates to a rotary encoder, and particularly to a rotary scale having a radial diffraction grating in which a plurality of grid patterns of transparent parts and reflective parts are periodically formed on the circumference. The present invention relates to a rotary encoder that irradiates a light beam and uses diffracted light from the diffraction grating to electrically detect the rotation angle and speed of a rotating scale.

[Prior art]

Conventionally, a diffraction grating is provided on the circumference of a disk (scale) connected to a rotating object, and the diffraction grating is irradiated with a laser beam.
A rotary encoder is known that detects the rotation angle and rotation speed of a disk by interfering diffracted light generated by a diffraction grating and detecting changes in brightness of the interference light.

Figures 12(a) and (b) are from Japanese Patent Application Laid-Open No. 63-9151.
FIG. 5 is a configuration diagram of a conventional rotary encoder described in No. 5. In the figure, 1 is a laser, 2 is a scale having a diffraction grating, 3 is a reflecting prism, 4 is a polarizing prism, 51 and 52 are light receiving elements, and 6 is a rotation axis of the scale 2. In FIG. 12, the light beam emitted from the laser 1 enters the diffraction grating 2 at a position M almost perpendicularly. M,
The ±1st-order diffracted light generated in the reflection prism 3 is reflected in the right angle direction by the first right-angled reflection surface 3a of the reflection prism 3, and is totally reflected twice each by the side surfaces 3c and 3d of the reflection prism 3. The light is reflected again in the right angle direction by the right angle reflecting surface 3b and is incident on the position M2 of the diffraction grating 2. Figure 12 (
b) is an explanatory diagram of the optical path inside the reflecting prism 3, and FIG.
FIG. 4 is a plan view of FIG. As shown in FIG. 12(b), the ±1st-order diffracted light generated by MI has a diffraction angle α+
, α-, and exits from the side surface 3c of the reflecting prism 3.

Total reflection occurs in 3D. Then, the light intersects near the center of the reflecting prism 3, and is further totally reflected on the side surfaces 3c and 3d.
The light is incident on the position M2 of the diffraction grating 2 at the same angles as the above-mentioned diffraction angles α+ and α−. Then, the ±1st-order pre-refracted light in M2 is parallel to the incident light from the laser 1 in Ml, overlaps in the opposite direction, and exits the diffraction grating 2.

Then, this interfering ±1st order pre-refracted light is passed through a polarizing prism 4.
The light is received by the light receiving elements 51 and 52 via the light receiving elements 51 and 52. When the diffraction grating 2 rotates by one grating pitch, the phase of the ±1st-order diffracted light changes by ±2π. Similarly, the phase of the ±1st-order pre-refracted light changes by ±4π with respect to the rotation of one grating pitch.

Therefore, as shown in FIG. 12, when the ±1st-order pre-refracted lights are caused to interfere with each other, a sine wave signal for four periods is obtained from the light receiving elements 51 and 52 by rotation of the grating l pitch of the scale 2.

If the total number of gratings is N, a sine wave signal of 4N periods can be obtained in one rotation. In addition, in FIG. 12, M, and M2
are in a positional relationship that is almost point symmetrical to each other with respect to the rotation center of the rotation shaft 6, and this prevents measurement errors from occurring even if there is eccentricity when attaching the scale 2 to the rotation shaft 6. . Further, from the light receiving elements 51 and 52, a 90° phase difference signal is obtained by a combination of the linearly polarized light of the laser 1, the elliptically polarized light due to total reflection within the reflecting prism 3, and the polarizing prism 4, and a 90° phase difference signal is obtained from the diffraction grating 2. The direction of rotation can also be determined.

In the conventional example configured as described above, the following problems occur.

(1) The optical path of the ±1st-order diffracted light to be interfered with is the reflection prism 3
Inside, pass through the separate sales route. Therefore, measurement errors are likely to occur due to environmental changes such as changes in ambient temperature.In particular, the larger the diameter of the scale 2, the longer the optical path within the reflecting prism 3, that is, the non-common optical path, resulting in errors. is likely to occur. In addition, when the laser oscillation wavelength changes due to such environmental changes, the optical path of the ±1st order diffracted light changes,
The ±1st-order diffracted light does not enter the position M2.

(2) If the diffraction grating forming surface of the scale 20 is tilted relative to the rotation axis 6, the diffracted light generated in Ml will not be re-injected into M2, and therefore a measurement error will occur due to the influence of the eccentricity described above.

[Summary of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a highly accurate rotary encoder that eliminates the problems of the prior art described above and eliminates measurement errors caused by environmental changes or the inclination of the scale 2.

In order to achieve the above object, the rotary encoder of the present invention irradiates a light beam from a light source to a first position of a rotary scale in which a diffraction grating is formed along the rotation direction, and generates a light beam at the first position. The first and second diffracted lights are transferred to a second position that is approximately symmetrical to the first position with respect to the rotation center of the rotation scale.
the first and second re-diffracted lights generated by diffraction of the first and second diffracted lights at the second position are caused to interfere with each other and are guided to a photodetector; In a rotary encoder that detects the rotational state of the rotary scale based on an output signal from It is characterized in that it is provided with an optical system that makes the positions of No. 2 and 2 have a conjugate relationship.

Further features and embodiments of the invention are described in the Examples below.

〔Example〕

FIG. 1 is a perspective view showing a first embodiment of the present invention, in which 1 is a light source consisting of a laser diode, etc.;
0 is a collimator lens, 31, 32, 33 is a prism,
41.42 is a polarizing beam splitter surface, 5 is a rotating disk plate (diffraction grating) forming a scale, 60. 61.
62. 63 is a mirror, 71. 72. 73 is 1/4
A wavelength plate, 8 is a 1/2 wavelength plate, 9 is a non-polarized beam splitter, 110 and 11 are polarizing elements (for example, a polarizing plate or a polarizing prism), 12 and 13 are lenses, and S1 and S2 are light receiving elements.

A laser beam with a wavelength λ emitted from a light source 1 is made into a parallel beam by a collimator lens 20, and is made incident on a prism 31 along symmetrical optical paths Ll and L2 by mirrors and polarizing beam splitter surfaces 41 provided at predetermined locations on the prism 31. The light beams R1 and R2 are divided into two traveling beams, reflected by a mirror 60, and passed through a quarter-wave plate 71. 1 point (Ml) at the same time. Here, among the plurality of diffracted lights diffracted by the diffraction grating and emitted from the point M1, the +1st-order reflected diffraction light of the light flux R1 and the -1st-order reflected diffraction light of the light flux R2 travel backward along the original optical paths Ll and L2, respectively. The luminous flux R1゜R is set in advance so that it is emitted in the direction
The angle of incidence θ of 2. θ. =sin-'(λ/2P). Furthermore, when the light beams R1 and R2 are split by the polarizing beam splitter surface 41, they become linearly polarized light whose polarization planes are orthogonal to each other. The plane of polarization of is switched. That is, since R1 is linearly polarized light (P polarized light) transmitted through the polarizing beam splitter surface, the 10th order diffracted light (R
1+) becomes S-polarized light through the quarter-wave plate 71, is reflected by the polarization beam splitter surface 41, and is emitted from the prism 31. Also, the light flux R2 is the polarizing beam splitter surface 4.
Since it is linearly polarized light (S polarized light) reflected at 1, the luminous flux R2
The first-order diffracted light (R2-) becomes P-polarized light through the quarter-wave plate 71, passes through the polarization beam splitter surface 41, and exits from the prism 31, overlapping with the light beam (R1+). The light beam (R1+) and the light beam (R2-) pass through the center (optical axis) of the lens 12 while remaining overlapping, and pass through the prism 33.
reflected by mirrors 61 and 62 and transmitted,
Passing through the 1/2 wavelength plate 8, the center (optical axis) of the lens 13
The light passes through and enters the prism 32. and prism 3
Mirrors and polarizing beam splitters installed at predetermined locations in 2.
The surface 42 causes the light beam (R1+) to travel along the optical path L3, and the light beam (R2-) to travel along the optical path L4. The light flux (R1+) and the light flux (R2-) are each mirror 6.
3 and passes through the quarter-wave plate 72, the second point (M2) of the radial diffraction grating provided on the rotating disk plate 5
) to the angle θ. incident at Here, 1. The /2 wavelength plate 8 converts the polarization plane of the +1st-order diffracted light (R1+) from S-polarized light to P-polarized light, and converts the polarization plane of the -1st-order diffracted light (R2-) from P-polarized light to S-polarized light. Also, the point M1 of the radial diffraction grating
and point M2 are set in a symmetrical positional relationship with respect to the rotation axis 0 of the rotating disk plate 5. Reflected and diffracted by the diffraction grating to point M
Luminous flux (R1+) among multiple reflected diffraction lights emitted from 2
The +1st-order diffracted light (R1++) travels backward along the original optical path L3, passes through the quarter-wave plate 72 again, becomes S-polarized light, is reflected by the polarizing beam splitter surface 42 in the prism 32, and is transmitted from the prism 32. The −1st-order diffracted light (R2-1) of the emitted light beam (R2-) travels backward along the original optical path L4, passes through the quarter-wave plate 72 again, becomes P-polarized light, and is sent to the polarizing beam splitter in the prism 32. It passes through the - plane 42, overlaps with the +1st order diffracted light (R1+10), and exits from the prism 32. When the two overlapping beams pass through the quarter-wave plate 73, they become circularly polarized lights whose planes of polarization rotate in opposite directions. Therefore, the polarization state of the beam that is a combination of these circularly polarized beams that rotate in opposite directions is linear. It becomes polarized light. The polarization direction of this linearly polarized light beam changes +1 according to the rotation of the rotating disk plate 5.
Next diffracted light (R1++) and -1 Kuki diffracted light (R2--)
It is determined by the phase difference between the wavefronts, and the phase difference is O9π/4,
2π/4, 3π/4.4π/4°5π/4. ..., 8
The polarization direction of the linearly polarized light beam changes as π/4 (in between
5°, 67.5°, 90', 112,5°.

135°, 157.5°,..., 225° (4
5°). Therefore, after this luminous flux is split into two luminous fluxes of equal light quantity by a non-polarizing beam splitter 9, only a specific polarized component of one of the luminous fluxes is separated using a polarizing element 10, extracted, and incident on the light receiving element s1. , if the other light beam is separated into only a specific polarized component using the polarizing element 11 and taken out and made incident on the light receiving element s2, the light receiving elements Sl and S2 each have a period corresponding to the amount of rotation of the rotary disc plate 5. A signal is output. Here, polarizing element l
If the polarized light components extracted by O and 11 are shifted by 45 degrees from each other, the timing of the change in brightness of the interference light incident on the light receiving elements Sl and S2 will be shifted by 174 cycles (π/2 in the phase of the output signal). By performing well-known electrical amplification and binarization processing on these two-phase periodic signals whose phases are shifted by 90 degrees from each other, the rotation angle and rotation direction of the rotary disk plate 5 can be detected.

In the rotary encoder of this embodiment, the optical path L1 when the light flux R1 enters the diffraction grating is almost equal to the optical path toward which the +1st-order reflected diffraction light R1+ of the light flux R1 exits, and the light flux R2 enters the diffraction grating. Optical path L2 and luminous flux R when
The optical path of the first-order reflected diffraction light R2- of the second-order reflected diffraction light R2- is set to be approximately equal to the optical path L3. Since L4 and each optical path of the ±1st-order diffracted light are also satisfied, even if the grating pitch P of the radial diffraction grating of the rotating disk plate 5 becomes a small value comparable to the wavelength λ of the light beam from the light source 1, The incident angles of the light fluxes R1 and R2 with respect to the points M1 and M2 and the exit angles of the ±1st order diffracted light and the ±1st order diffracted light can be set to about 30°. Therefore, unlike conventional encoders, when the grating pitch P is set fine (this causes problems such as difficulty in extracting the diffracted light).
＜, constitutes a high-resolution encoder.

Here, in this embodiment, a lens 12, a lens 13, Mirror 61 and mirror 62 are
It is arranged so that the following conditions are met.

First, a finite area centered on point M1 on the rotating disk plate 5 is mapped by the lens 12 to a finite area (intermediate image plane) centered on point P3, and then a finite area centered on this point M3. is mapped by the lens 13 onto a finite area centered on the point M2 on the disk plate 5. Under this condition, the light flux emitted from the point M1 at any angle can be made incident on the point M2 as long as it is not interrupted on the way. That is, in this embodiment, ' is the lens 12, the lens 13, the mirror 61 . The optical system including the mirror 62 establishes an optically conjugate relationship between the point Ml and the point M2. This prevents deterioration of measurement accuracy.

Next, the optical path of the diffracted light (R1+) emitted from point M1 is shifted from the optical path L1 by an angle Δθm (angle Δθx15.Δθy1. when decomposed into x + y components), and the optical path L
An optical path L3 in which an incident angle of an optical path (L7) that enters point M2 when traveling along an optical path (L5) different from 1 is set in advance.
The mirror 61. Other reflective surfaces such as mirror 62 and mirror 60, 63 are arranged.

The same applies to the optical path of the diffracted light (R2-). With this condition, the exit direction of the ±1st order diffracted light beams emitted from the point M2 becomes constant regardless of the position and orientation of the rotating disk plate 5, and the optical paths of the ±1st order diffracted lights that overlap at the polarizing beam splitter 42 are mutually Since they are always parallel, ±1st order diffracted light (
The signal obtained by photoelectrically converting the interference light due to the superposition of R1++) and (R2--) becomes stable.

We will prove this below.

First, the deviation of the optical path when the rotation axis 0 (rotation center) of the rotating disk plate 5 and the center of the radial diffraction grating are deviated by Δr will be explained with reference to FIGS. 2 and 3. The center of the radial diffraction grating on the rotating disk plate 5 is
As the diffraction grating rotates, it moves around the rotation axis 0, and the pitch of the diffraction grating at the point M1 changes periodically. As shown in Fig. 2, when the center of the radial grating is at point a, the pitch of the diffraction grating at point M1 is Δp(=2πΔr/N,
However, N becomes thinner by the number of radial diffraction gratings), and the point M
Since the pitch of the diffraction grating in 2 becomes thicker by Δp,
+1st-order diffracted light (R1+) of luminous flux R1 emitted from point M1
The first-order diffracted light (R2-) of the luminous flux R2 has an output angle θ5.degree. that is larger than the incident angle θ1.degree.θ2 (θ1=θ2=θ.) of the optical paths L1 and L2 by ΔθX. The optical path L becomes θ6 and deviates from the optical path L1 or L2.
5 or L6 and returns to the prism 31. Each light beam exits from the prism 31 and enters the lens 12 at a position offset from the center (optical axis), but due to the action of the lens 12, it travels toward point M3. And the ±1st-order diffracted light (R1+) and (R2-) that passed through point M3 are 1
/2 wavelength plate 8 (not shown), the optical path is bent by the lens 13, passes through the prism 32, and becomes the optical path L7.
Proceeding through L8, Δ from O3°O4 (θ3=θ4=θ.)
Incident angle θ7, which is smaller by θX. It enters point M2 at θ8.

At point M2, the pitch of the diffraction grating is thicker by Δp, so for example, the +1st-order diffracted light (R1++) of the luminous flux (R1+)
The exit angle θ++ is θ++=sin-' [λ/(p+Δp) sinθ
7), but θ5=sin-' [λ/(p-ΔI)) si
Substituting nθ1)θ7=2θ, −θ5, sinθ0=λ/(p+Δp) −sin [2θH−5in−′ [λ/(p−Δp)
−sinθ1)]=λ/(p+Δp) −5in [2
sin-'(λ/2p)-sin-' (λ/(p-
Δp)−λ/2p) ]=λ/ (p+Δp) −[2(λ/2p) −[λ/(p−Δp)−λ/2p
l] = λ/2p, that is, θ+"=θ1=θ3, and the +1st order diffracted light (R1 +
) is kept constant.

The traveling direction (emission direction) of the -1 Kuki diffracted light (R2--) is also kept constant for the same reason.

Next, when the center of the radial diffraction grating is at the point as shown in Fig. 3, the parameters θ], φl of the incident light flux in the coordinate system (t, r, z) (see Fig. 4) are expressed as θ1=
θ.

φ1=0, but the array direction of the diffraction grating at point M1 (axis)
is shifted by jan-' (Δr/r) in angle,
Since it is the +1 axis, the relationship between the light beam R1 incident on the diffraction grating and the diffraction grating is as shown in FIG. In FIG. 4, θ, 1 is the angle of incidence of the light beam on the rotating disk plate 5, φ, s
[=tan-' (Δr/r)] is the angle formed with the array direction (11 axes) of the diffraction grating. In addition, the 21st axis and the r1 axis are determined as shown in FIG. 4, and the angles formed between the optical path L1 of the incident light beam and each coordinate axis are α1□, β11, . Set to γ11
- and cosα,,=3inθll”cO8φII −(
1) cosβ1. =5iHθu'sinφ11 −
(2) γ1. =θ1. =θ1 ・
... (3). Therefore, the coordinate system (t,,r,,z
O12=cos-'((1-cos" a 12-c
os"β+□) +4 ]φ12 = tan-' (c
osβ12 / COS Q 12 ) (α12<90
) = tan-' (cosβ12/CO8α12)
+ 180 (α1□〉90) =90 (α1□〈90.β1□=0) Knee 9
0 (α12>90.β12=O). 'Here, Δr is sufficiently smaller than r, and the incident angle of the luminous flux R1 at the point M1 is θ.

When [=sin-' (λ・N/ (2yr r)l], the diffracted light (R
1+) output direction parameter θ5. φ5 is θ, = θ1 □ = θ1 φIs: φ12 = −2・φ11 It can be immediately determined by calculation that the optical path L5 of the +1st-order diffracted light (R1+) of the light flux R1 is aligned with the disk surface as shown in Figure 2. The angle formed with the t-axis when projected onto is -2・
φ5. When the luminous flux is incident on point M2,
The optical path L of the incident light beam is determined by the combination of the lens 12, the lens 13, the mirror 61, the mirror 62, and other reflective surfaces.
7 is projected onto the lattice array plane (t-r plane) of the rotating disk plate 5 and the angle it makes with the t-axis is shifted by -2·φ, 1, so the +1st-order diffracted light (R1++) of that light beam is The angle formed with the t-axis when projected onto the lattice arrangement surface of the rotating disk plate 5 is returned to zero. That is, it completely matches the optical path L3.

After the 1st-order diffracted light (R2-) of the luminous flux R2 exits from the point Ml and enters the point M2, the optical path of the 1st-order diffracted light (R2--) that exits from the point M2 also changes to the optical path L for the same reason.
0, which exactly matches 4. Ru.

Next, the rotating disk plate 50 rotating shaft 0 and the rotating disk plate 5
The manner in which the optical path is corrected when the normal line erected to the lattice array plane is relatively tilted by Δξ will be explained with reference to FIGS. 5 and 6. Since the incident angle and the incident direction (angle) of the light beam R1 with respect to the rotating disk plate 5 vary as the rotating disk plate 5 rotates, the +1st-order diffracted light (R1
The output angle θ5 and the output azimuth φ5 of +) also vary. For example, when the normal to the rotating disk plate 5 is at position C as shown in FIG. 5, the incident light flux R in the coordinate system (t+r, z)
1 parameter θ1. φ1 is θ, 200φ = O, but the angle α between the optical path L1 and the coordinate axis t1 is 1,
The angle β between the optical path L1 and the coordinate axis r1 is 1, the optical path L1 and the coordinate axis 2. The angle γ2. Each position is α,,=90
−○, ...(11) cosβ++
=8inα11*sinΔξ−(12)cos 7
11 = sin a 11 ・cosΔξ −(13
), the parameters of the incident light flux in the coordinate system (tII rll zl) are θ,, =cos”” (sin a 11 LIco
sΔξ)φ11 = tan-' (cosβ,,/co
saB).

Here, assuming that the center of deflection is at the center of the rotating disk plate 5, the incident position of the light beam R1 onto the rotating disk plate 5 is Δx [= rth tanΔξ・tanθ, ko, Δy
[=r (1/ cosΔξ−1) point Ml shifted by
1, and the incident position of the light beam R2 shifts by -ΔX, Δy to point M12, so the parameters of the incident light beam in the coordinate system (t21 r21 z2) centered on the incident point Mll of the light beam R1 are θ, 2=θ,.

φ, 2=φ11-tan-' [ΔX/(Δy+r)1
The angle α1° with the coordinate axis t2, the angle β1□ with the coordinate axis r2, and the angle γ, 2 with the coordinate axis z2 are as follows: cos a 12 = sin θ12”cosφ12CO
5β12=sinθ, 2-sinφ12γ12 ”β1
2, the emission direction α of the +1st-order diffracted light in the coordinate system (j21 r21 22), 3. β, 3. By substituting α13. β13. All you have to do is calculate γ, 3. Here, Δξ is close to 0, and the incident angle of the luminous flux R1 at the point M1 is θ. [==sin-'(λ・N/(2π
When it is designed in r months, the coordinate system (t+ r+
parameter α14° β14. When converted to γ14, α 14 = α 1a cosβ14=C(FiΔξ” cosβ13 Sin
Δξ−COS 713cos γ+a=sinΔξ*
cosβ13 +cosΔξ@COS 713θ6.
If you change it to φ5, θ6: γ14 φ5 ==tan-' (cosβ14 / COS
α14 ) (α, 4<90) = tan-1 (cos β14 /cos a 14)
+180 (α, 4 > 90) = 90 (α14 ” 90 + β,,<
90) = -90 (α, 4 = 90. β, 4 > 90
). Calculating this, we get the luminous flux R1 as shown in Figure 5.
The optical path L5 of the +1st order diffracted light (R1+) of
The angle formed with the t-axis when projected onto the lattice array plane is φ5
When the light beam is deviated by 100 degrees and enters the point M21, the optical path L7 of the incident light beam is projected onto the lattice arrangement surface of the rotating disk 5 by a combination of the lens 12, the lens 13, the mirror 61, the mirror 62, and other reflecting surfaces. Since the angle formed with the t-axis is φ7 = φ5, the +
The angle formed with the t-axis when the first-order diffracted light (R1++) is projected onto the disk surface is returned to zero. That is, the traveling direction (emission direction) of the +1st-order diffracted light (R1++) emitted from the point M21 becomes parallel to the optical path L3. Similarly, the first-order diffracted light (R2-) of the luminous flux R2 is emitted from the point M12, and is emitted from the point M2.
2, the first-order diffracted light (R
The optical path of 2--) is also parallel to the optical path L4.

Next, when the normal to the rotating disk plate 5 is at position d,
Incident light flux R1 in the coordinate system (x+Y+z).

The parameter θ1. φ1 is θ1=θ.

φ, −〇, but the angle α12 between the optical path L1 and the coordinate axis t1,
The angle β17 between the optical path L1 and the coordinate axis r1, and the angle γ and 1 between the optical path L1 and the coordinate axis z1 are α, , = 90
−θ1+Δξ β,,=90 γ7. = θ1−Δξ Therefore, the parameters of the incident light flux in the coordinate system (t, , rI+zI) are θ, 1=θ1−Δξ φlI20. Here, assuming that the center of the deflection is at the center of the rotating disk plate 5, the incident position of the light beam R1 on the rotating disk plate 5 coincides with Ml, so the +1st order diffracted light (R1+) of the light beam R1 is emitted. Direction θ1□.

φ12 is sinθ11+sinθ, 2=+1− λ/pφ1□
=0, that is, θ, 2 = sin” (+1 @λ/p-5inθ11
) = 5in-' [2・sinθ1-sin (θ1
-Δξ))=θ1+Δξ Therefore, the emission direction θ6. in the coordinate system (t, r, Z). φ5 is θ5=θ1□+Δξ 20. +2・Δξ φ5=0. Similarly, the emission direction θ of the −1st-order diffracted light of the luminous flux R2
6. φ6 becomes θ6=θo−2·Δξ φ6=0. The incident azimuth θ7, φ7 of the optical path L7 of the luminous flux (R1+) incident on point M2 is determined by the combination of lens 12, lens 13, mirror 61, mirror 62, and other reflective surfaces.
is θ7=θ5 φ7=O, and the incident direction θ13. in the coordinate system (t3+r3t z3). φ, 3 is θ13=θ7+Δξ=θ. +3・Δξ φ4. = 0, so if we calculate the emission direction θ, 4°φ, 4 of the first-order diffracted light (R1++) emitted from M2 using the diffraction angle calculation formula, θ14=θ. +Δξ φ14:0, and the output directions θ++ and φ++ in the coordinate system (t+ ”+ 2) are θ 1+ = θ 3 = θ.

φ++00 Therefore, it completely matches the optical path L3. -1st order diffracted light (R
2-) enters point M2 and exits from point M2. The output direction of the -1st-order diffracted light (R2--) is also calculated as follows: θ-=θ.

φ−=0 This completely matches the optical path L4.

In this way, when mounting the rotating disk plate 5, there may be a positional error or a change in the oscillation wavelength of the light source 1.
Even if the course of the diffracted light emitted from the first position (Ml) is shifted, the optical system (12, 13° 61, 62) of the present invention allows the diffracted light to move to the second position (M2) where it re-enters the rotating disk plate 5.
) does not change, in principle there is no deterioration in measurement accuracy due to changes in the reference position for reading the diffraction grating, and furthermore, the traveling direction ( By configuring the apparatus to keep the emission direction constant, the interference fringe pattern formed by the two diffracted lights is not disturbed. Therefore, the diffraction grating can be read stably.

In addition, in the embodiment shown in FIG. 1, the optical paths of the ±1st-order diffracted lights (R1+, R2-) between the polarizing beam splitter surfaces 41 and 42 are almost the same, so the light beams R1 and R
2 over a long optical path from point M1 to point M2, the variation in the optical path length difference between the light beams due to changes in ambient temperature, etc. is small. Therefore, only the signals accurately corresponding to the rotation of the rotating disk plate 5 are transmitted to the light receiving elements Sl and S.
It will be output from 2. Further, in this embodiment, the optical system is set in advance so that the optical path lengths of the light beams R1 and R2, which form interference light with each other, are equal.

FIG. 7 shows a modification (second embodiment) of the embodiment shown in FIG. In this embodiment, the incident angle of the light beams R, 1, and R2 incident on the rotating disk plate 5 is θ. With some changes from , the outgoing optical path (LL, L2. L3. L4) and the returning optical path (
L5. L6. L7. L8) is slightly shifted, but the basic structure is the same. In FIG. 1, the +1st-order diffracted light (R1+) of the luminous flux R1 emitted from the point M1 and the specularly reflected light (zero-order diffracted light) of the luminous flux R2 travel along the same optical path LL, and their planes of polarization are different from each other at the polarizing beam splitter surface 41. Due to the difference, only the first-order diffracted light (R1+) of the light beam R1 should be reflected, but depending on the imperfections of the optical components, a part of the specularly reflected light of the light beam R2 is also reflected by the polarizing beam splitter surface. Since the light is transmitted, ghost light (noise) is generated. Regularly reflected light (zero-order diffracted light) of the luminous flux R1 at point Ml, +1st-order diffracted light (R1+) emerging from point M2
) specularly reflected light (zero-order diffracted light) and -1st-order diffracted light (R2-)
The same applies to the specularly reflected light (zero-order diffracted light). Therefore, if the optical path is set as shown in Fig. 7, the optical path of the ghost light will be shifted from the optical path of the diffracted light for forming the interference light, so that the ghost light will be prevented from entering the light receiving elements Sl and S2. I can do it.

In each of the above embodiments, the +1st order and the 11th order are used as the orders of the diffracted light. However, as shown in FIG. The direction in which the traveling direction of the plane is shifted is marked as +, and the opposite direction is marked as -. Furthermore, it goes without saying that the order of the diffracted light forming the interference light is not limited to the first order, but may also be the second order or higher order.

FIG. 9(a) is a perspective view of the third embodiment of the present invention, FIG.
b) is a developed optical path diagram of the embodiment of FIG. 9(a). In FIG. 9, 1 is a semiconductor laser, 20 is a collimator lens, 5 is a rotation scale on which a diffraction grating is formed along the rotation direction, and 41.42 is a polarizing beam splitter 1.
Sl and S2 are light receiving elements, 81° and 82 are reflecting mirrors, and 7 is 17
A four-wavelength plate, 9 a beam splitter, and 110 and 11 polarizing plates, the polarization directions of which are shifted by 45° from each other. Further, the direction of the linearly polarized light of the laser 1 is at 45 degrees with respect to the direction of polarization of the polarizing beam splitter 41. laser 1
The emitted light beam is converted into a parallel light beam by the collimator lens 20, and then split by the polarizing beam splitter 41 into a transmitted light beam (p-polarized light) and a reflected light beam (s-polarized light) into equal light quantities. The two divided beams enter the position M1 of the radial diffraction grating of the rotation scale 5 at an angle (diffraction angle) given by θ in equation (1) below.

θ=sin-'λ/p...(1)
Here, λ is the oscillation wavelength of the laser 1, and p is the grating pitch at the position M1 of the diffraction grating of the rotary scale 5. Incidentally, the plane of incidence of the two beams of light incident on the position M1 is
This is a plane parallel to the grating arrangement direction (tangential direction) of the diffraction grating 2 in .

The ±1st-order transmitted diffracted light of the two light beams generated at the position M1 is emitted from the rotation scale 5 in a direction perpendicular to the grating arrangement plane of the diffraction grating. Then, this ±1st-order diffracted light is reflected by the reflecting mirror 83.
The light is reflected at right angles (parallel to the grating arrangement plane) and is directed via lenses 101 and 102 to a reflecting mirror 84 located at a position approximately symmetrical with respect to the rotation axis O of the rotation scale 5. Then, this reflecting mirror 74 makes the ±1st-order diffracted light enter the position M2 of the diffraction grating of the rotary scale 2 perpendicularly from the same direction again. At position M2, the ±1st-order diffracted light is again emitted at an angle θ of equation (1).

These +1st-order re-diffracted lights are reflected by reflecting mirrors 85 and 86, enter the polarizing beam splitter 42, overlap each other, pass through the quarter-wave plate 7, and are split into two beams by the beam splitter 9. Each light beam enters the light receiving elements Sl and S2 via the polarizing plate 10.11. In this way, a sine wave signal generated as a result of the interference of the ±1st order diffracted light is obtained from the light receiving elements SL and S2.

Next, the optical path of the ±1st-order diffracted light will be explained using the optical path development diagram of FIG. 9(b). In FIG. 9(b), the transmitted light beam (solid line light beam) that passes through the polarizing beam splitter 41 is incident on the position M1 of the diffraction grating, and the first-order diffracted light is emitted vertically from the grating arrangement surface of the diffraction grating. The +1st-order diffracted light is reflected by the polarizing beam splitter 41 (luminous flux indicated by a broken line) and is incident on M1, and is emitted in the vertical direction from the grating arrangement surface of the diffraction grating.

Let the first-order diffracted light be the 11th-order diffracted light. The ±1st-order diffraction lights generated at position M1 overlap, and follow a common optical path from position M1 to position M2 through an optical system consisting of reflecting mirrors 83 and 84 and lenses 101 and 102. At position M2, ±1st order diffracted light is generated again for each of these ±1st order diffracted lights, but due to the action of the polarizing beam splitter 42, the 9th order diffracted light
Only the light beams indicated by solid lines and broken lines in FIG. 2B are incident on the light receiving elements Sl and S2.

That is, when the +1st-order diffracted light of the p-polarized light (luminous flux +, p shown by the solid line) from position M1 enters the position M2, ±1st-order diffracted light is generated again. The light is reflected by the mirror 85, passes through the polarizing beam splitter 42, and enters the light receiving elements Sl and S2. However, the -1st order diffracted light (not shown) is reflected by the reflecting mirror 86,
Since the light passes through the polarizing beam splitter 42, it does not enter the light receiving elements Sl and S2. On the other hand, the 1st-order diffracted light of the S polarization from Ml (luminous flux -+ s of the broken line) is at the position fiM2
Among the ±1st-order diffracted lights that are incident and generated, the -1st-order diffracted light (luminous flux -1 indicated by the broken line) is reflected by the reflecting mirror 86, reflected by the polarizing beam splitter 42, and enters the light receiving elements Sl and S2. However, the +1st-order diffracted light (not shown) is reflected by the reflecting mirror 85 and then reflected by the polarizing beam splitter 42, so that it does not enter the light receiving elements Sl and S2. In this way, +
The light that has undergone first-order diffraction twice (solid line luminous flux) and the light that has undergone -1st-order diffraction twice (dashed line luminous flux) are overlapped and output from the polarizing beam splitter 42, resulting in 1/ 4 wavelength plate 7
, beam splitter 9, polarizing plate 10. 11, it becomes two interference lights, and each interference light goes through the light receiving elements Sl and S.
2. As in the conventional example shown in FIG.
A sine wave signal corresponding to the period is obtained. Further, the light beam emitted from the polarizing beam splitter 42 passes through the quarter wavelength plate 7,
As in the previous embodiment, linearly polarized light whose polarization direction rotates as the rotary scale 5 rotates, but the polarizing plate 10. 11
Since the polarization directions of the two light beams are shifted by 45 degrees from each other, signals whose phases are shifted by 90 degrees from each other are obtained from the light receiving elements Sl and S2. ' In FIG. 9, the optical path of the ±1st-order diffracted light that reaches positions M1 to M2 of the diffraction grating of the rotary scale 2 is a common optical path, and measurement errors are less likely to occur due to environmental changes such as changes in the surrounding temperature. In addition, as in the conventional example shown in FIG.
Assembly and adjustment are easy because there is no complicated optical path such as crossing diffracted lights. In addition, the lens 10 arranged on the common optical path
1. Reference numeral 102 is set so that Ml and M2 are in a conjugate relationship with each other, and the same effect as that of the previous embodiment can be obtained.

FIG. 10 is a schematic diagram of a fourth embodiment of the present invention.

In the figure, 1 is a semiconductor laser, 20 is a collimator lens, and the light beam from the laser 1 is made to become a parallel light beam. 30a (30b) is a polarizing beam splitter that reflects the S-polarized light and transmits the P-polarized light of the incident light beam. 40a (40b) and 50a (50b) are mirrors, and 5 is a rotating scale on which a radial diffraction grating is formed.

Here, mirrors 40a (40b), 50a (50b
) are such that the two beams from the polarizing beam splitter 30a (30b) are incident on the diffraction grating at angles corresponding to the diffraction angles of the diffraction orders of different signs, for example, the 1st and 11th orders. There is. 70a (70b) is a polarization conversion means, which is made of a λ/4 plate. 80a (
Reference numeral 80b) denotes a reflecting means, which is made of, for example, a mirror or an optical member in which a reflective film is applied to the end face of an end face imaging type gradient index lens.

In this example, the reference numbers 30a (30b), 40a (
40b).

50a (50b), 70a (70b) and 80a
Detection unit 1 having each element corresponding to (80b)
Two oOa (100b) are provided.

Next, the detection operation of the rotary encoder in this embodiment will be explained.

A collimator lens 2 converts the laser beam from the laser 1 into a parallel beam, and a polarizing beam splitter 30a splits the beam into two beams: P-polarized light and S-polarized light.

Of these, the S-polarized light reflected by the polarizing beam splitter 30a is reflected by the mirror 40a and is incident on the position M1 on the diffraction grating surface on the rotation scale 5 at an angle corresponding to the first-order diffraction angle.

On the other hand, the P-polarized light transmitted through the polarizing beam splitter 30a is reflected by the mirror 50a, and is incident on the position M1 on the diffraction grating surface on the rotation scale 5 at an angle corresponding to the 11th order diffraction angle.

The S-polarized light that has been first-order diffracted by the diffraction grating is transmitted and emitted approximately perpendicularly from the grating array plane, passes through a quarter-wave plate 70a serving as a polarization conversion means, becomes circularly polarized light, is reflected by a reflection means 80a, and is directed to the polarization direction. becomes reversed circularly polarized light and returns to its original optical path. Then, it passes through the quarter-wave plate 70a again and re-enters the position M1 on the diffraction grating as P-polarized light. The first-order P-polarized light that has been re-diffracted by the diffraction grating is reflected by a mirror 40a and passes through a polarizing beam splitter 30a.

On the other hand, the P-polarized light that has been diffracted in the 1st order by the diffraction grating is transmitted and emitted from the grating array surface approximately perpendicularly, passes through the quarter-wave plate 70a, becomes circularly polarized light, is reflected by the reflection means 80a, and is directed to the polarization direction. becomes reversed circularly polarized light and returns to its original optical path.

Then, it passes through the quarter-wave plate 70a again and re-enters the position M1 of the diffraction grating as an S-polarized light beam. Then, the -1st-order S-polarized light re-diffracted by the diffraction grating is reflected by the mirror 50a, and then reflected by the polarization beam splitter 30a.

From this, the polarizing beam splitter 30a superimposes and extracts two light beams, a p-polarized light beam and an S-polarized light beam, and guides these two light beams to a mirror 109. mirror 10
The two light beams guided to the mirror 9 overlap each other with their polarization planes perpendicular to each other, and the two light beams reflected by the mirror 109 are reflected by the mirror 110 via the lenses 101 and 102, and then
The light is guided to a detection unit 100b having the same configuration as described above. Then, similarly to the detection unit 100a, the two beams are subjected to first-order diffraction twice at the position M2 on the diffraction grating, and then guided to the extraction quarter-wave plate 7.

At this time, the two light beams overlap each other with their planes of polarization perpendicular to each other. Since the mutual phase difference δ is the phase difference between the light beam that has undergone +1st-order diffraction four times and the light beam that has undergone -1st-order diffraction four times, it becomes I6 π X . That is, when x=IP (P is the lattice pitch), 2
The phase shift between the light beams is 16π.

The composite wave of the two beams that has passed through the quarter-wave plate 7 becomes a linearly polarized beam, and its polarization direction rotates by half a rotation while the phase between the two beams shifts by 2π, so it passes through the beam splitter 9 and becomes a linearly polarized beam. The luminous flux passing through the rays undergoes one period of brightness and darkness change, and is outputted as an electrical signal of one period by the light receiving element si.

On the other hand, the light beam reflected by the beam splitter 9 is caused by the periodic brightness and darkness of the light beam that passes through the polarizing plate 1 because the polarizing plate 11 is arranged with the direction of the polarization plane shifted by 45 degrees with respect to the polarizing plate 10. The phase of change in is shifted by 90°.

Therefore, the phase of the periodic signal output from the light receiving element S2 is always shifted by 90° compared to the phase of the periodic signal output from the light receiving element Sl.

If the total number of folded gratings on the rotating scale is N, then the change in the phase difference δ between the two light beams during one rotation of the scale 5 becomes δ=16πN by substituting x=NP. From this, two light receiving elements 15.16 to 16
A periodic signal (sine wave signal) with a period of πN/2π=8N is obtained.

In this embodiment, the lenses 101 and 102 are arranged so that the positions M1 and M2 of the diffraction gratings on the rotary scale 5 are in a conjugate relationship with each other, so that the common optical path of the ±1st-order diffracted light (the optical path from the polarizing beam splitter 30a to 30b) It is located inside. Therefore, like the encoders of the first to third embodiments described above, a system is constructed in which measurement accuracy is unlikely to deteriorate due to changes in ambient temperature or the inclination of the rotary scale 5. Also,
In this example, the light that has undergone +1st-order diffraction four times and the light that has undergone 11th-order diffraction four times form interference light, so a resolution four times as high as in each of the above-mentioned examples can be obtained. .

As the diffraction grating used in the encoder of the present invention, not only the well-known amplitude type diffraction grating but also a phase type diffraction grating such as a relief type diffraction grating or a hologram can be used. Further, as a light source, use of a semiconductor laser such as the above-mentioned laser diode is suitable for miniaturizing the device. As the semiconductor laser, various types can be selected, such as a single mode type and a multimode type.

FIG. 11 shows an example in which the rotary encoder shown in FIG. 1 is used in a device for writing track signals on a hard magnetic disk 113. The rotary encoder shown in FIG. The rotary disk plate 5 of the encoder is placed on a rotary disk (not shown), and the encoder reading unit 111 is placed on top of the rotary disk plate 5 of the encoder. Note that the reference number 115 indicates an air bearing.

〔Effect of the invention〕

As mentioned above, even if the traveling direction of each diffracted light beam emitted from the first point of the diffraction grating changes somewhat due to positional deviation or tilt when the scale plate is attached, the second point of the diffraction grating where each diffracted light beam re-enters the scale plate.
Since the position of the grating does not change, in principle, there is no deterioration in measurement accuracy due to a change in the reference position for reading the grating. or,
Since each diffracted light beam has a nearly common optical path, measurement accuracy does not deteriorate due to ambient temperature fluctuations, etc. Furthermore,
The interference fringe pattern formed by the two diffracted lights is not disturbed by configuring the device so that the traveling directions (emission directions) of the respective diffracted lights to be superimposed on each other emitted from the second position are kept constant. Therefore, it is possible to stably read interference signals and provide a high-resolution, high-precision rotary encoder. In addition, the installation of the rotating disk plate is relatively unaffected by errors, and the rotating disk plate, light emitting and receiving optical systems (reading optical units) can be separated, making it easy to create a "built-in type" rotary encoder. can.

[Brief explanation of the drawing]

FIG. 1 is a perspective view showing a first embodiment of the present invention. FIGS. 2 to 6 are explanatory diagrams showing optical path correction in the case where there is an error in the attachment of the rotating disk plate of the rotary encoder of the first embodiment. FIG. 7 is a perspective view showing a second embodiment of the present invention. FIG. 8 is an explanatory diagram showing how to assign signs to the orders of diffracted light. FIGS. 9(a), (b) and 10 are the third embodiments of the present invention, respectively.
and the remaining figures of the fourth embodiment. FIG. 11 is a diagram showing an example in which the encoder of FIG. 1 is used in a track signal writing device for a hard magnetic disk. FIGS. 12(a) and 12(b) are explanatory diagrams showing an example of a conventional encoder. l・・・・・・・・・・・・・・・・・・・・・・・・
...Light source (laser diode) 20...
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・Collimator lens 31.32.33
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・・・・Prism 41.42・
・・・・・・・・・・・・・・・Polarizing beam splitter plane 5・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・・・・・・・
...Rotating disk plate 60.61, 62.63
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・Mirror 71.72.73・・・・・・
・・・・・・・・・・・・・・・・・・・・・・・・
・・1/4 wavelength plate 8・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・1. /2 waves ■■ 9・・・・・・・・・・・・・・・・・・・・・・・・
...Non-polarizing beam splitter-io, 11...
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・・・・Polarizing element Sl, S2 ・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・・・・・・・
...... Light receiving elements 12, 13...
・・・・・・・・・・・・・・・・・・・・・・・・
・・・・・・・・・・・・・・・・・・・・・・・・
Len Y Fig. 2, incident on the enemy nz ↑ 5 luminous flux ~ 2 % 7 Fig. entry dimension vertically

Claims (1)

[Claims] A light beam from a light source is irradiated onto a first position of a rotating scale in which a diffraction grating is formed along the rotation direction, and the first and second diffracted lights generated at the first position are directed to the center of rotation of the rotating scale. the first and second re-diffracted lights generated by diffraction of the first and second diffracted lights at the second position. In a rotary encoder that causes interference to be guided to a photodetector and detects a rotational state of the rotary scale based on an output signal from the photodetector, the first and second diffracted lights have substantially a common optical path with each other. A rotary encoder characterized in that an optical system is provided in the common optical path to bring the first and second positions into a conjugate relationship.