JP5902891B2 - Encoder and calibration method - Google Patents

Description

  The present invention relates to an encoder and a calibration method using the encoder.

  The calibration principle of the device that calibrates the angle is disclosed in Non-Patent Documents 1 and 2, for example. According to this, a method called “equal division averaging method” is adopted, and the two rotary encoders are simultaneously calibrated by comparing two rotary encoders (reference encoder and encoder to be measured) whose scale position deviation is not known. Further, it is necessary to refer to the value of the reference encoder at the angle point of the encoder under measurement during the measurement process. However, since the angle point of the encoder to be measured does not normally coincide with the angle point of the reference encoder, it is necessary to use the interpolation signal of the reference encoder. Here, the angle point is a term representing an angular position when the encoder signal is binarized and converted into a pulse signal, for example, and when there are 1000 angle points, a scale for dividing one rotation of 360 ° into 1000 is shown. means. The interpolated signal is a signal having a phase value calculated by performing arctangent calculation based on sinusoidal analog encoder signals that are 90 ° out of phase with each other. For example, when one period of an analog encoder signal is divided into 80, the phase angle is calculated by dividing the range of 0 to 2π into 80 divided regions and outputting a pulse every time the region is switched, To improve.

  However, in general, the encoder signal has a problem that the error of the interpolation signal varies depending on the distortion (amplitude, offset, phase difference, harmonic component, etc.) of the sinusoidal signal. That is, when distortion is present in the sine wave signal of the encoder, an error is generated where one cycle of the sine wave signal should be divided at equal intervals. For example, such problems are taken up in Patent Documents 1 and 2 and the like, and a method of correcting the distortion in advance and recording it in a database and correcting it later is shown. However, in general encoders, the distortion of the sinusoidal signal varies depending on the mounting state, scale drawing error, etc., and it is difficult to specify the uncertainty, which is inconvenient as a calibration apparatus. Therefore, the above-described angle calibration apparatus uses the “time average method” as the interpolation method.

  In the time average method, as shown in FIG. 5, the rotating shaft 101 of the encoder 200 to be measured and the rotating disk 101 of the head 100 of the reference encoder are connected to a common rotating shaft 102 and rotated at a constant speed by a motor 103. When referring to the value of the reference encoder corresponding to the angle point of the encoder 200 to be measured, the pulse generation time T (i) of each angle point of the encoder 200 to be measured and the pulse generation time of the angle point of the reference encoder before and after that Tm (i) is measured. Then, assuming that the reference encoder rotates at a constant speed between each angle point, the angular position information of the reference encoder is calculated from the time information by a method of linear interpolation. This method is characterized in that the principle is clear and is not affected by the waveform distortion of the encoder 200 to be measured or the reference encoder. In particular, when a high-resolution encoder (that is, an encoder with many scales) is used as the reference encoder, due to the inertia of the rotating structure, the speed between adjacent angular points of the reference encoder is infinitely constant. It has been adopted.

  However, since this method cannot be measured unless the measurement object is rotated, a stationary object such as a polygon mirror cannot be calibrated. Therefore, conventionally, a method of calculating a value by combining (interpolating) the angle point of the reference encoder near the angle point of the polygon mirror with an interferometer in combination with an interferometer is used. That is, when a stationary object is calibrated, there are problems such as the fact that the work cannot be directly calibrated and the work becomes complicated, and the uncertainty of the laser interference length measuring apparatus must be added.

  Also, recent encoders are shifting from so-called incremental encoders that count periodic signals to absolute encoders. In the case of an absolute encoder, functions such as the output of angular position information based on commands rather than one-sided data output have been changed when outputting angle information, which is inconvenient for the calibration principle based on continuous rotation. There is.

JP 2003-161644 A JP 2006-90738 A

Journal of Japan Society for Precision Engineering, Vol. 67, No. 7, 1091-1095, 2001.07, Development of a high-precision rotary encoder calibration system (1st report) AIST Metrology Standards Report, Vol. 1, No. 19, pp. 19-23, 2002.01, Development of high-precision rotary encoder calibration equipment

  An object of the present invention is to provide a technique useful for an encoder used for calibration of an object.

One aspect of the present invention is an encoder for sequentially positioning an object and calibrating the object, wherein a light source and a plurality of marks are arranged and move relative to the light source together with the object. Splitting the light from the scale and the light source into two polarized light beams having different polarization directions, and diffracting the two polarized light beams obtained by the splitting by the mark, and diffracting light of positive order and negative order. By generating light and making the generated two diffracted beams interfere with each other in the circular polarization state opposite to each other, linearly polarized light whose polarization direction is rotated by the pitch of the mark and the relative movement speed of the scale is generated. An optical system that divides the linearly polarized light generated by the optical system into a first linearly polarized light and a second linearly polarized light, and a fixedly arranged first polarization on which the first linearly polarized light is incident. A first unit that includes a plate and detects the number of rotations of the polarization direction of the first linearly polarized light by detecting light that has passed through the first polarizing plate during the relative movement of the scale; and the second straight line A second polarizing plate that rotates with respect to a polarization direction of the second linearly polarized light that is incident on the polarized light, and detects light that has passed through the second polarizing plate in a state where the scale is stationary. A second unit for obtaining a phase within one rotation of the second linearly polarized light, and outputting the quantity relating to the relative movement based on the number obtained by the first unit and the phase obtained by the second unit. And an output unit.

  According to the present invention, for example, a technique useful for an encoder used for calibration of an object can be provided.

It is a block diagram of the rotary encoder of 1st Embodiment. It is a figure which shows a mode that a polygon mirror is calibrated with the rotary encoder of 1st Embodiment. It is a block diagram of the rotary encoder of 2nd Embodiment. It is a block diagram of the rotary encoder of 3rd Embodiment. It is a figure which shows a mode that a to-be-measured encoder is calibrated using a prior art.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[First Embodiment]
FIG. 1 is a configuration diagram of an encoder for measuring a displacement or an angle according to a first embodiment that can be used as a reference encoder for calibrating a stationary polygon mirror or encoder. FIG. 2 is an explanatory diagram when a polygon mirror is calibrated using the reference encoder of FIG. The light emitted from the semiconductor laser (light source) 1 is made into substantially parallel light by the collimator lens 2, enters the roof portion of the first polarizing prism 31, is reflected inside, and is polarized by the polarizing beam splitter surface 41. It is divided into two polarized lights (P wave and S wave) having different directions. The P wave transmitted through the polarization beam splitter surface 41 passes through the optical path L 1, is reflected by the reflecting prism 63, passes through the ¼ wavelength plate 71, and is obliquely incident on the position P 1 of the rotating disk (scale) 101. A plurality of marks are arranged on the rotating disk 101 at a constant pitch, and move relative to the head including the semiconductor laser 1. The rotating disk (scale) 101 generates positive-order diffracted light and negative-order diffracted light by diffracting obliquely incident P waves and S waves with marks. The P-wave + 1st order diffracted light obliquely incident on the position P1 of the rotating disk 101 returns to the original optical path, re-transmits through the quarter-wave plate 71, and passes through the polarizing beam splitter surface 41 in the polarizing prism 31. This time, the light is reflected and output as an S wave (+ 1st order light) to the optical path to the prism 3.

  On the other hand, the S wave reflected by the polarization beam splitter surface 41 passes through the optical path L2, is reflected by the reflecting prism 63, passes through the quarter wavelength plate 71, and is positioned on the radial diffraction grating 5 of the rotating disk 101. Incidently incident on P1. The S-wave −1st-order diffracted light obliquely incident on the position P 1 on the radial diffraction grating 5 returns to the original optical path, retransmits through the quarter-wave plate 71, and the polarizing beam splitter surface 41 in the polarizing prism 31. Then, the light passes through and is output as a P wave (−1st order light) to the optical path to the prism 3.

  These two light fluxes of P wave (−1st order light) and S wave (+ 1st order light) are reflected by the reflecting surfaces 64 and 65 in the prism 3 with the optical axes being overlapped, and are mutually reflected by the half-wave plate 8. Are changed to P wave (+ 1st order light) and S wave (-1st order light), respectively. The P wave (+ 1st order light) and the S wave (-1st order light) are then incident on the roof of the second polarizing prism 32, and after being reflected inside, the two polarized light beams are split by the polarizing beam splitter surface 42. Is done.

  First, the P wave (+ 1st order light) transmitted through the polarization beam splitter surface 42 passes through the optical path L3, is reflected by the reflecting prism 66, then passes through the quarter-wave plate 72, and is inclined to the position P2 of the rotating disk 101. Incident. The P-wave + 1st order re-diffracted light obliquely incident on the position P2 of the rotating disk 101 is transmitted again through the quarter-wave plate 72, and is then reflected by the polarization beam splitter surface 42 in the polarizing prism 32. In the optical path to the / 4 wavelength plate 73, S wave (+ 1 + 1st order diffraction) is output. On the other hand, the S wave (−1st order light) reflected by the polarization beam splitter surface 42 passes through the optical path L 4, is reflected by the reflecting prism 66, passes through the quarter wavelength plate 72, and is positioned on the rotating disk 101. Incidently incident on P2. The S-wave −1st-order rediffracted light that is obliquely incident and reflected at the position P2 of the rotating disk 101 returns to the original optical path, retransmits through the quarter-wave plate 72, and passes through the polarization beam splitter surface 42. This time, the light is transmitted and output as a P wave (−1-1 order diffraction) on the optical path to the quarter-wave plate 73. Here, the meaning of the sign of the diffraction order applies a rule in which the diffraction order of the light beam diffracted toward the rotational direction when the diffraction grating 5 rotates clockwise is applied, for example, S wave (+ 1 + 1 order diffraction) Means s-polarized light with a + 1st order diffraction performed twice.

  The S wave (+ 1 + 1 order diffraction) and the P wave (-1-1 order diffraction) before passing through the quarter-wave plate 73 are independent linearly polarized light whose polarization directions are shifted from each other by 90 °. When the rotating disk 101 rotates by one pitch, the S wave (+ 1 + 1st order diffraction) has a wavefront phase shifted by 2π × (+2), and the P wave (−1-1 order diffraction) has a wavefront phase of 2π × (−2). ) Just shift. Therefore, the change in the phase difference between the two light beams is 8π / pitch. The pitch means the pitch between adjacent marks on the disk 101.

By passing through the quarter-wave plate 73, both the S-wave (+ 1 + 1st order diffraction) and P-wave (-1-1st order diffraction) light beams are in the circular polarization state opposite to each other and interfere with each other. One linear polarization with a rotating polarization direction is generated. The polarization direction of the linearly polarized light beam depends on the phase difference between the original two light beams, and when the phase difference is shifted by 2π, the polarization direction of the linearly polarized light is half-rotated. That is, in the optical system of the present embodiment, the change in the phase difference between the original two light beams was 8π / pitch, so the rotation of the polarization direction of the linearly polarized light becomes 2 rotations / pitch. The optical system including the first polarizing prisms 31, 32, quarter-wave plates 71, 72, 73, etc. is an integral fraction of the period (1/4) determined by the mark pitch and the relative movement speed of the scale. ) To generate linearly polarized light whose polarization direction rotates with a period of a magnitude of. Therefore, the optical system improves the measurement resolution of the movement of the scale (rotating disk 101) by an integral multiple. When the rotation of the disk 101 is stopped, the polarization direction of the linearly polarized light beam is fixed in the direction determined by the phase difference between the two light beams. The linearly polarized light beam having such characteristics is split by the non-polarizing beam splitter (dividing unit) 9 into two light beams of the first linearly polarized light and the second linearly polarized light having the same polarization characteristics.

  The first linearly polarized light that has passed through the non-polarizing beam splitter 9 passes through a fixedly disposed polarizing plate (first polarizing plate) 10 and enters the light receiving element 51. The signal obtained by the light receiving element 51 is binarized by the comparator 91 and becomes an incremental signal S0. As in the prior art, the origin signal resets to zero and the counter 96 counts the incremental signal S0. The first polarizing plate 10, the light receiving element 51, the comparator 91, and the counter 96 obtain the number of rotations of the polarization direction of the first linearly polarized light that has passed through the polarizing plate 10 during the relative movement of the disk 101 as a scale. It constitutes one unit.

The second linearly polarized light reflected from the non-polarizing beam splitter 9 passes through the polarizing plate 11 connected to the rotating body and enters the light receiving element 52. The polarizing plate 11 constitutes a second polarizing plate that rotates at a predetermined angular velocity with respect to the polarization direction of the second linearly polarized light. It is assumed that the direction of the transmission axis of the polarizing plate 11 is rotated at a constant speed (at a high speed) by the motor M and its motor driver 93. Then, the output of the light receiving element 52 is maximized when the direction of the transmission axis of the polarizing plate 11 coincides with the polarization direction of the second linearly polarized light, and the output of the light receiving element 52 is minimized when they are orthogonal. When the output of the light receiving element 52 is binarized by the comparator 92, a rotation period signal S2 in the polarization direction of the second linearly polarized light is obtained. If the direction of the transmission axis of the polarizing plate 11 is controlled by the motor driver 93 with time, a timing signal S1 is obtained from the control system. The rising times Tr (1), Tr (2),... Of the linearly polarized light rotation period signal S2, and the rising times T (1), T (2), T (3),. Is as shown in FIG. When the relative movement of the disk 101 as a scale is stopped, the polarization direction of the second linearly polarized light stops in a specific direction. The stationary polarization direction of the second linearly polarized light can be obtained by detecting the rotation angle of the rotating polarizing plate 11. The arithmetic unit 95 detects the polarization direction of the linearly polarized light, thereby calculating the phase φ within one rotation period of the linearly polarized light when the disk 101 is stationary. The times Tr (1), Tr (2),... And the times T (1), T (2), T (3),... Are the rise times of the polarization rotation period signal S2 and the timing signal S1. Instead, it may be a fall time. The polarizing plate 11, the motor M, the light receiving element 52, the comparator 92, the motor driver 93, and the calculator 95 constitute a second unit that obtains a phase within one rotation of linearly polarized light when the scale is stationary.
The counter 96 is reset to zero with the origin signal and counts the incremental signal S0 as in the prior art, thereby dividing the integer position information N (360 ° around the entire circumference by four times the number of gratings of the diffraction grating 5). When ordinal). The position information calculation unit 97 integrates the position information N of the integer part and the phase information φ calculated by the calculator 95, and outputs the encoder position information POS. The position information calculation unit 97 constitutes an output unit that outputs the movement amount of the scale by integrating the results obtained by the first unit and the second unit.

  In the first embodiment, the time change of the intensity of the second linearly polarized light that has passed through the polarizing plate 11 when the disk 101 is stationary is detected by the light receiving element 52, and the polarization of the second linearly polarized light is detected by the calculation unit 95 from the detection result. The direction is calculated.

  FIG. 2 is a side view of a configuration when a polygon mirror (object) is calibrated using the above-described rotary encoder (reference encoder). The rotating disk 101 of the reference encoder is coupled to the lower surface of the air bearing 102, and further, the head 100 of the reference encoder is installed upside down in the lower part of FIG. The head 100 of this reference encoder is mounted on the sub-rotation stage in accordance with the installation rules used in the “equal division averaging method” so that the installation orientation of the head 100 can be changed. Details are not described here. A main rotary stage 105 is attached to the upper part of the air bearing 102, and a polygon mirror 201 is placed on the upper part. A zero-cross signal 203 is generated from the autocollimator 202 at a timing when the polygon mirror 201 faces the autocollimator 202 placed on the work surface 106. The main rotary stage 105 is controlled by the controller 104 and the motor 103 so that the main rotary stage 105 can be positioned by the zero cross signal 203. Through the above-described process of combining the polygon mirror 201 and the reference encoder, the polygon mirror 201 and the rotating disk 101 of the reference encoder can rotate coaxially.

  First, the rotation of the main rotation stage 105 is controlled, and positioning is performed in order so that the normal line of each mirror surface of the polygon mirror 201 faces a predetermined direction (a direction facing the autocollimator 202). At each positioning, the formation position of each mirror surface of the polygon mirror 201 and the rotation angle POS of the head 100 of the reference encoder are measured. Then, the computing unit 107 compares the measurement result of the formation position of each mirror surface in the measurement process with the measurement result of the rotation angle by the reference encoder, and performs a predetermined calculation (such as an equal division average method). A correction value of angle control data for each mirror surface is determined. The same applies when the encoder is calibrated. First, the position or angle of the object is measured using both the first encoder to be calibrated and the second encoder that is the reference encoder. After the measurement step, the correction value (calibration data) of the output of the first encoder is determined by comparing the measurement result by the first encoder and the measurement result by the second encoder. For example, when the absolute encoder is calibrated, the main rotary stage 105 is positioned for each angle point of the absolute encoder. Then, each time positioning is performed, the absolute encoder is calibrated by comparing the absolute position information of the absolute encoder with the position information POS of the head 100 of the reference encoder and performing a predetermined calculation (equal division averaging method or the like).

[Second Embodiment]
FIG. 3 is a configuration diagram of the encoder of the second embodiment in which the second unit for detecting the reference phase of the first embodiment of FIG. 1 is changed. In the second embodiment, a rotary encoder E is installed in the rotation control system of the polarizing plate 11, and the rotation angle of the polarizing plate 11 is detected by the rotary encoder E. Then, the rotary encoder E detects the polarization direction of the polarizing plate 11 that matches the polarization direction of the second linearly polarized light, whereby the polarization direction of the second linearly polarized light is calculated. In the case of the second embodiment, it is not necessary for the rotation of the polarizing plate 11 to be a constant speed rotation as in the first embodiment, and the measurement accuracy of the rotary encoder E becomes the phase calculation accuracy as it is. The accuracy required for the interpolation signal is about 1000 divisions per rotation, but it is very easy to realize this with a general rotary encoder E. In FIG. 3, the radial grating of the rotary encoder E is integrally recorded on the outer peripheral portion of the polarizing plate 11, and the rotation angle reading encoder head E is arranged close to the radial grating, so that the size can be reduced. . In addition, this rotary encoder E can use not only an optical type but a magnetic type, an electrostatic type, etc.

[Third Embodiment]
FIG. 4 is a configuration diagram of the encoder of the third embodiment. In the third embodiment, two rotary encoders E1 and E2 using polarized light are installed instead of the rotary encoder E using a radial grating, and the rotation angle of the polarizing plate 11 is detected based on these values. In FIG. 4, two reference polarizing plates having 0 ° and 45 ° azimuth are arranged in proximity to the outer periphery of the polarizing plate 11. Then, by receiving the transmitted light from the polarizing plate 11 and one of the two reference polarizing plates, a two-phase signal in which the phase of two cycles is shifted by 90 ° per rotation is generated. Based on the rotation angle position of the polarizing plate 11 calculated based on the two-phase signal, the rotation of the polarizing plate 11 is controlled via the servo driver 93 and the motor M.

[Other Embodiments]
The present invention is not limited to the configurations of the first to third embodiments, and various modifications and changes can be made within the scope of the gist.
In the first to third embodiments, the optical system of the grating interference type rotary encoder using polarized light is composed of two polarization beam splitter surfaces, three quarter-wave plates and one half-wave plate. . However, the number of polarizing beam splitter surfaces, quarter wavelength plates, and half wavelength plates can be changed. In the first to third embodiments, the film surface of the polarizing prism is used as the polarizing beam splitter surface, but a mirror may be used. Furthermore, although the diffraction grating is a reflection grating in the first to third embodiments, it may be a transmission grating.
In the first embodiment, the number of light receiving elements 51 for incremental signals is only one. However, as the light receiving element 51, two or four light receiving elements having a phase difference of 90 ° are used to output a two-phase or four-phase signal, one phase is set as a main counting phase, and the other phases May be used for determination to eliminate instability of count up / down.

The rotating polarizing plate 11 can be changed to a rotating wave plate. That is, the second unit includes a rotating wave plate and a fixedly arranged polarizing plate, and detects a change with time of the intensity of the second linearly polarized light that sequentially passes through the rotating wave plate and the fixedly arranged polarizing plate. Then, the polarization direction of the second linearly polarized light is calculated from the detection result.
The arithmetic unit 95 that calculates the interpolation phase using time in the first embodiment may realize an equivalent function with another algorithm or flow. For example, it has a built-in timer, detects the generation time of the rising pulse of the reference phase signal in order, compares it with the control time of the rotation control signal, converts the control signal of the rotation control system into a rectangular wave, Various methods such as a method of calculating a phase from the time (or integral value) of AND with a wave can be adopted.
-The positioning (stop) operation of the main rotary stage 105 of the first embodiment is for easy understanding of the explanation. Actually, measurement is possible even if the stop state is not complete by repeating measurement continuously. It is. However, in that case, it should be noted that the accuracy is lowered in relation to the rotation speed of the polarizing plate inside the reference encoder.
In the first to third embodiments, the reference encoder is a rotary encoder. However, the present invention can also be applied to a linear encoder which is a linear encoder. In the case of a linear encoder, the content is the same by replacing angle points with coordinate points.

In the present invention, the following effects can be obtained.
Inside the reference encoder, the light beam for generating the normal angle point signal and the light beam for generating the reference phase are divided from the same light beam. Then, the angle point signal is obtained as the same pulse generation time as before, and is obtained as a time function of the pulse signal by the constant speed rotating body in which the reference phase is separately incorporated. Therefore, since the uncertainty does not depend on the signal distortion of the reference encoder or the object to be measured (encoder, polygon mirror) but depends only on the rotation unevenness of the constant speed rotating body, the uncertainty of the interpolation signal between the angle points is estimated. Is possible.
・ Because the object to be measured can be interpolated, the polygon mirror and absolute encoder can be calibrated.

Claims (9)

An encoder for sequentially positioning an object and calibrating the object,
A light source;
A plurality of marks arranged, and a scale that moves relative to the light source together with the object ;
The light from the light source is divided into two polarized light beams having different polarization directions, and the two polarized light beams obtained by the splitting are diffracted by the mark to generate positive-order diffracted light and negative-order diffracted light. An optical system that generates linearly polarized light whose polarization direction is rotated by the pitch of the mark and the relative movement speed of the scale by causing the two diffracted lights thus generated to interfere with each other in a circularly polarized state in opposite directions. When,
A splitting unit that splits linearly polarized light generated by the optical system into first linearly polarized light and second linearly polarized light;
The first linearly polarized light is detected by detecting light passing through the first polarizing plate while the scale is relatively moved, including a fixedly arranged first polarizing plate on which the first linearly polarized light is incident. A first unit for obtaining a number of rotations in a direction;
A second polarizing plate that rotates with respect to a polarization direction of the second linearly polarized light that is incident on the second linearly polarized light, and detects light that has passed through the second polarizing plate when the scale is stationary. A second unit for obtaining a phase within one rotation of the second linearly polarized light in the state;
An output unit that outputs an amount related to the relative movement based on the number obtained by the first unit and the phase obtained by the second unit;
An encoder comprising: The second polarizing plate is a polarizing plate that rotates at a predetermined angular velocity,
The said 2nd unit detects the change with respect to time of the intensity | strength of the light which passed the said 2nd polarizing plate, The polarization direction of the said 2nd linearly polarized light is obtained from this detection result, The said 2nd unit is characterized by the above-mentioned. Encoder. The second polarizing plate is a rotating polarizing plate,
The second unit detects an intensity of light that has passed through the second polarizing plate and a rotation angle of the second polarizing plate, and obtains a polarization direction of the second linearly polarized light from the detection result. The encoder according to claim 1.   The encoder according to claim 3, wherein the second unit includes a rotary encoder that detects a rotation angle of the second polarizing plate.   The second unit includes a rotating wave plate and a fixedly arranged polarizing plate, detects the intensity of light sequentially passing through the rotating wave plate and the fixedly arranged polarizing plate, and from the detection result The encoder according to claim 1, wherein a polarization direction of the second linearly polarized light is obtained. The encoder according to any one of claims 1 to 5, wherein the calibration is performed by performing the positioning without bringing the object into a stopped state. The encoder according to any one of claims 1 to 6 , wherein the encoder is a rotary encoder. A method for calibrating an encoder comprising:
Using both the first encoder to be calibrated and the second encoder according to any one of claims 1 to 7 to measure the amount of relative movement of the object;
Calibrating the first encoder based on the measurement result by the first encoder and the measurement result by the second encoder;
Including methods. A method for calibrating a polygon mirror,
Combining the polygon mirror and the scale of the rotary encoder according to claim 7 so as to be coaxially rotatable;
In the rotary encoder, the rotation angle of the scale in a state where the polygon mirror and the scale are rotated coaxially and the polygon mirror is sequentially positioned so that the normal of each mirror surface of the polygon mirror faces a predetermined direction. Each measuring step,
Calibrating each mirror surface of the polygon mirror based on the measurement result in the rotary encoder;
Including methods.

Images