JPH0157291B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an encoder, and particularly to an encoder for digitally reading the absolute position or angle of a moving body with high precision and high resolution. In general, in encoders that digitally read lengths, angles, etc. with high precision, it is sufficient to subdivide the scale, but there is a manufacturing limit to the subdivision of the scale, so high precision beyond this limit is required.
To obtain high resolution, it is necessary to further interpolate the minimum scale. Conventionally, as an interpolation method for this purpose, for example, in an optical encoder, two slits are arranged so that the outputs of a code plate and two light receiving elements differ by 1/4 pitch in order to read information from an optical grating. There is a method of interpolating within the minimum scale to 1/4 by processing the 90° phase shift signal obtained from these two sensors, and a method of interpolating within the minimum scale to 1/4, and a method of interpolating the grating and sensor relative to A method is known in which interpolation is performed using an output value output from a sensor depending on the position. These methods further include, for example, an incremental method that uses relative movement between a scale and a sensor for measurement, and an absolute method that uses an encode plate that encodes absolute position information. In the incremental method, it is inconvenient because it is necessary to align the origin of the so-called reference position so that the counted value becomes zero when the reference position of the scale graduation is in front of the sensor, and it is also vulnerable to noise. There are drawbacks. On the other hand, in the absolute method, for example, the absolute position information of the encoder plate is always output as the photoelectric conversion output of the sensor, but as mentioned above, the measurement accuracy is limited due to manufacturing limitations when subdividing the scale. The encoder has two bright and dark
Since it is encoded using a binary code based on values, N tracks of codes are required on the code board to obtain a resolution of 1/2 ^N. Therefore, there are disadvantages such as the pattern becomes complicated and a corresponding number of pairs of light sources and photoelectric conversion elements are required. As an encoder that eliminates these drawbacks, for example, an encoder described in Japanese Patent Application Laid-Open No. 55-82919 is known. FIGS. 1a and 1b show the general structure of this encoder. That is, as shown in FIG. 1a, one track of the code plate 2 is divided into N blocks, and each block contains pattern information 11 for interpolation and address information indicating the block address as shown in FIG. 12 (for example, a binary code) arranged in a transparent and non-transparent optical grating, and minute photoelectric conversion elements as sensors are arranged in series at equal pitches in parallel with the code plate 2, and each element An encoder is known that uses an element array 3 (hereinafter referred to as a line sensor) from which independent photoelectric conversion outputs can be extracted as time-series serial signals. A reading method when this conventional encoder is applied to a linear scale is illustrated in Japanese Patent Application No. 53-155815 (Japanese Unexamined Patent Publication No. 55-82919) previously filed by the present applicant. This is the code plate 2 as shown in Figure 2.
A line sensor 3 is installed in parallel with the line sensor 3, and the code plate 2 is irradiated with the photoelectric 1 through the collimator lens 4. The light beam transmitted through the code plate 2 has a light intensity corresponding to a specific pattern on the code plate 2, and is projected onto the line sensor 3 by the imaging lens 5. Therefore, if the total length of the line sensor 3 is appropriately selected, the time-series serial output signal of the line sensor 3 includes block marker information 10A indicating the reference position of the block on the code plate 2, block start position information 10B,
Interpolation pattern information 11 and block address information 12 are output in time series. The interpolation method here is, for example, based on the number of gratings of the code plate 2, which is made up of periodically arranged transparent and non-transparent optical gratings, and the line sensor 3.
A vernier relationship is generated by creating a difference between the number of sensor elements in This is the so-called Vernier interpolation method. During measurement, an arbitrary bit of the line sensor 3 is predetermined as an index bit. The number of pulses from the start bit to the index bit of the line sensor 3 determined by detecting the block start position information 10B by the line sensor 3 is obtained. With this number of pulses, a value indicated by an index bit obtained by dividing one block of the code plate 2 at intervals equal to one bit of the line sensor 3 is obtained. This is taken as bit information. In other words, it is a rough reading. In addition, after separating the vernier information into odd and even numbers, a 2-bit interpolated value is obtained by finding the point where the output of the odd number is smaller than the output of the even number, and this becomes a fine reading. In this way, digit alignment can be achieved by setting the minimum resolution of coarse reading to be smaller than the interpolation range of fine reading. Furthermore, the digits of the 2-bit interpolated value and the bit information are aligned, and the address information of the block is combined and converted into position information to obtain the absolute position of the moving object. The above is a method of reading an absolute position in one track using a conventional encoder. Although the above reading method has been described for reading one block, it is usually necessary to read a plurality of blocks in order to improve reliability, correct reading errors, etc. However, in the conventional encoder described above, when reading a large number of blocks, it is necessary to increase the size of the illumination optical system for illuminating the large number of blocks on the code plate. Furthermore, when reading a pattern in which each block is independent, the further the block is from the optical axis, the greater the error due to the influence of magnification, aberration, etc., and it may therefore be outside the range of correction. Furthermore, for example, when a linear encoder is used for long lengths, the number of bits of address information in each block increases, which increases the length of each block, making it difficult to read multiple blocks. The present invention was made in view of the above circumstances, and
The purpose of this is to be able to read multiple blocks with multiple blocks near the optical axis without requiring a large illumination optical system or a large reading range, and to freely select the reading range of multiple blocks. An object of the present invention is to provide an encoder that can digitally read the absolute position or angle of a moving object with high precision and high resolution. According to the present invention, there is provided a code plate in which a plurality of pieces of interpolation information such as address information and vernier information forming part of a predetermined pattern arrangement such as an M-sequence arrangement are arranged alternately in series; It is equipped with a line sensor that reads information on the board and a processing circuit that processes the output signal of the line sensor, and each block is configured by the plurality of address information and pattern information, and each block is An encoder characterized in that an address is specified by a pattern array such as an M-sequence array obtained from the plurality of pieces of address information forming each block is obtained. An embodiment of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 3 is an explanatory diagram of a code plate of an encoder according to an embodiment of the present invention. In FIG. 3, the code plates 2 each contain address information 20a to 20e forming part of a predetermined pattern arrangement such as the M series.
and pattern information 21a-2 consisting of a plurality of pieces of information.
1e are alternately arranged in series, and each of the address information 20a to 20e is arranged in the same direction as the predetermined pattern arrangement such as the M sequence in the serial arrangement direction. There is. In FIG. 3, the code plate 2 is shown for convenience of explanation.
only a part of it is shown. Furthermore, each of the address information 20a to 20e in FIG.
Alternatively, it is composed of a grid pattern of "1" (hereinafter referred to as a character). In this embodiment, for example, four pieces of address information 20a to 20b and three pieces of pattern information 21a are used.
-21c constitute one block 30. Similarly, in FIG. 3, for example, one piece of address information is shifted to the right, and the address information 20b~
20e and pattern information 21b to 21d constitute one block 31. Thereafter, a large number of blocks 32, . . ., blocks N each consisting of four pieces of address information and three pieces of pattern information are constructed in the same manner. The address of each of these blocks is specified by a pattern arrangement obtained from a plurality of pieces of address information forming each block. For example, in FIG. 3, the address of block 30 is
It is designed to be specified by a pattern arrangement obtained from four pieces of address information 20a to 20d composing the address information, that is, a pattern arrangement obtained from four characters. In this case, the numbers of address information and pattern information constituting each block are shown as four and three pieces respectively in FIG. 3, but the number of these pieces of information constituting the block depends on the code board. can be set to any number when patterning. Therefore, the number of characters for reading the address of each block is not limited to the four characters described above, but can be set to any number corresponding to the number of address information constituting each block. Next, an example will be described in which the address of each block is specified by a pattern array obtained from four characters. For example, as shown in the character arrangement in Table 1(a), an M-series pattern arrangement is used as the predetermined pattern arrangement, and in this case, the address of each block is specified by each of the pattern readings shown in Table 1(b). It is assumed that the address value of each block is specified from the read value.

[Table] Table 1(b) Pattern reading In other words, each address information constituting each block, that is, the character, is a part of the M-series predetermined pattern arrangement, ie, “0” or “1” as shown in Table 1(a). For example, in FIG.
The read value of the character is 0000, which is (0) from Table 1(a), and the read value of the four characters 20b to 20d, which specify the address of the next block 31, is 0001, which is (1). Therefore, if you create a conversion table of block address values for read values as shown in Table 1(b), you can be sure of the address information of any block, that is, no matter which character you start reading from, the address of that block from the read value. can be known. Furthermore, since the address information of each block, that is, the characters, are arranged as described above, for example, when reading the address of an adjacent block, it is only necessary to shift the address by one character, so reading multiple blocks can be done within a small reading range. Since the number of characters can be set arbitrarily as described above, the reading range of a plurality of blocks can be freely selected. Furthermore, in FIG. 3, the pattern information 21a to 21e of each block 30 to 32 is, for example, a pattern of information for interpolation, and the difference between the number of grid patterns in this section and the number of elements of the line sensor corresponding to this is determined. This is a Vernier interpolation pattern. When each block is composed of four pieces of address information and three pieces of pattern information as described above, for example, in block 30, 21a to 21a are
If the vernier information is patterned to 1/3 of each of the three pattern information 21c and interpolated reading is performed by combining the three pattern information, the length of each block can be shortened. In this case, if the phases of the three pattern information are made continuous, the processing in the processing circuit described later will be simplified. Furthermore, if each address information of each block, that is, each character used to designate the block address, is composed of three or more bits on the line sensor, there is an effect that it can be completely distinguished from the vernier information. However, each character may be composed of one bit, and in this case, a pattern may be provided between the character and the vernier information so that the two can be distinguished. A schematic diagram of the structure of the code plate 2 described above is shown in FIG. In Figure 4 A ₄
~ _A10 is each address information of the code board 2, that is, each character, and each 3 bits 1 on the line sensor 3
It is made up of groups. B ₄ to _{B 10} are vernier interpolation patterns for length measurement formed as a set of three pieces of pattern information. These Vernier interpolation patterns are designed to provide a difference between the number of gratings and the corresponding number of elements of the line sensor 3 that reads them, so as to perform interpolation with a resolution of one pitch unit or less of the line sensor 3. In FIG. 4, if the code plate 2 is detected at position A by the line sensor 3 in which the index bit (IB) is set in advance, then at this position A, the line sensor 3 detects the 2nd bit between the reading start bit R and the index bit IB. 1 character and 2 characters after index bit IB A ₄ to _{A 7}
Since each address information is read as a time-series serial signal, the read address pattern value in this case is 0101, and the block address value can be decoded as 3 from the conversion table shown in Table 1(b). Similarly, if it is detected at position B, the address pattern read value is 1010 and the block address value is 4, and similarly, the C position is 0100 and the block address value is 5, the D position is 1001 and 6, and the E position is 0011. This can be interpreted as 7. Then, rough reading in each block is performed using the character closest to the front of the index bit IB, for example, in block _C1 , the sensor corresponding to the grid next to _A5 character, that is, the first grid _B51 of pattern information _B5 . It can be read by counting the number of elements from element M to index bit IB, and even finer reading can be done using pattern information such as B ₄ , B ₅ , B ₆
Since there is a difference between the number of grids and the number of sensor elements that read them, interpolation is performed by one pitch unit or less of the sensor. Below, when the sensor comes to positions B to D, reading can be performed using the same principle. The configuration and effects of the code plate of an encoder according to an embodiment of the present invention have been described above, and below, an embodiment of the present invention to which this code plate is applied will be described in detail with reference to FIG. 5. FIG. 5 shows an embodiment in which the present invention is applied to a linear encoder for detecting the position of a moving table of a machine tool or the like. The optical code plate 2 is fixed to a part of the moving table 50, the line sensor 3 is installed parallel to it, and when the code plate 2 is irradiated from the light source 1 through the condenser lens 4, the light transmitted through the code plate 2 is The light beam has a light intensity that corresponds to a specific pattern on the code plate 2, and this light beam is transmitted to the line sensor 3.
projected on top. Therefore, if the total length of the line sensor 3 is appropriately selected, the address information and interpolated reading information of the code plate 2 are output in time series as a time series serial output signal from one scan of the line sensor 3. Here, the approximately center bit of the line sensor 3 is set in advance as the index bit IB. Next, the output of the line sensor 3 is processed by the processing circuit described below. For example, the output 51 of the line sensor 3 is amplified to an appropriate level by an amplifier 52, and a character is detected by a character detector 54. Detector 54
detects a character and flips the flip-flop 5.
Set 7,58. Flip-flop 57 turns on analog switch 61 and sends out interpolation information. At the same time, the gate 62 is opened and the binary counter 63 is opened.
A clock pulse synchronized with the output of the line sensor 3 generated by the clock generation circuit 53 is sent to
At the same time as the interpolation information ends, the flip-flop 57 is turned off. The interpolation information in this case is vernier interpolation, and the vernier information of the line sensor 3 is sent to the A/D converter 70. Similarly, the second and third
Vernier information after a character is detected, such as a number, is divided and sent and written into the data register 71. Let this be data B. Furthermore, by passing the outputs of the flip-flop 58 set by the character detection and the flip-flop 59 set by the index bit detector 55 to the gate 64, the gate 65 is opened from the first character detection to the index detection, and at the same time the line sensor 3 continues to send a clock synchronized with the output of Therefore, the binary counter 66 calculates the number of pulses from the first detected character to the index bit, that is, the index bit on the scale obtained by dividing the reading block of the code plate 2 at intervals equal to one bit of the line sensor 3. The value indicated by is output. Furthermore, the conversion circuit 66' converts the value from the index bit to the character before it as explained in FIG. This conversion can be easily performed since the number of pulses in one block is known in advance. Let this be data A. Further, the multivibrator 56 is turned on by the number of bits corresponding to the sensor of the character, and is set by the binary counter 63 indicating the delimitation of interpolation information. This turns on analog switch 60 and sends one of the characters to comparator 67. A comparator 67 quantizes the signal at an appropriate threshold level. The second as well as the interpolation information,
The third character is read and converted into parallel binary number information by a serial-to-parallel converter 68. Further, this is converted into actual block address information by a conversion circuit 69. Let this be data C. The data A, B, and C generated by one scan of the line sensor 3 are input to the microprocessor 72. The microprocessor 72 approximates the outputs of the odd-numbered elements and the outputs of the even-numbered elements as straight lines (or curves) from the A/D-converted data B, which is vernier information, and at the intersection of both straight lines, the odd-numbered element becomes an even number. Find the point smaller than the th. The fine reading value obtained from this intersection is digit-aligned with the rough reading value obtained from data A, and this is further combined with data C to convert it into position information, which is displayed on the display 7.
Display on 3. The above is an explanation of the reading of one block. However, as described above, by assigning one character to read multiple blocks in the same manner as described above, and displaying the average value on the display 73, high accuracy and high accuracy can be achieved. A digital reading of the resolution can be obtained. Furthermore, in the embodiment of the present invention shown in FIG. 5, if the processing circuit for processing the output of the line sensor 3 is configured such that the processing after character detection and index bit detection is performed by the microprocessor 72, the code board The information in step 2 can be read more easily. Another embodiment of the invention in this case is shown in FIG. The embodiment shown in FIG. 6 is the same as that shown in FIG. 5 except for the configuration of the processing circuit, so the same parts are given the same reference numerals and the explanation thereof will be omitted. In FIG. 6, the output 51 of the line sensor 3 is amplified to an appropriate level by an amplifier 80, held for each bit by a sample hold 81, and stored in a first digital memory 83 via an A/D converter 82. The output peak value is stored as a digital signal. The contents of the first digital memory 83 are stored in the microprocessor 8.
5, the data is transferred to the second digital memory 84, and at the same time, the first digital memory 83 is cleared. The contents transferred to the second digital memory 84 are subjected to signal processing by the microprocessor 85 and displayed on the display 86. The signal processing by the microprocessor 85 at this time is performed, for example, as shown in the flowchart of FIG. In the above description of the embodiment, the coarse reading is performed up to one pitch of the sensor element, and the fine reading is interpolated within two pitches of the sensor element. This can be achieved if the minimum resolution is smaller than the interpolation range of fine reading, so for example, fine reading is limited to 1 pitch, and rough reading is performed by extracting the output of one sensor element of a line sensor as analog information and using it as appropriate. Of course, digit alignment may be performed by quantizing with a threshold circuit such that the minimum resolution for coarse reading is smaller than the interpolation range for fine reading. Furthermore, in the embodiment of the present invention, an example was explained in which the arrangement of block address information is separated into one character, but the present invention is not limited to this example, and for example, the arrangement of block address information is separated into two characters, It is clear that the same detection as described above can be performed even if the number of characters is separated into any number of characters, such as three characters. Furthermore, the arrangement of this address information and its conversion are not limited to those exemplified in Tables 1(a) and (b), and may be other M-sequence arrangements and their conversion.
For example, in the case of separation for 4 characters, it goes without saying that the array and its conversion as shown in Table 2 may be used.

【table】

Claims (1)

[Scope of Claims] 1 A plurality of address information forming part of a predetermined pattern array and pattern information consisting of a plurality of pieces of information are arranged alternately in series, and each of the address information A line sensor comprising a code plate arranged in the predetermined pattern in the arrangement direction, and a plurality of sensor elements arranged in series in parallel with the code plate or its image in order to read information on the code plate. and a processing circuit configured to read information on the code plate and process electrical signals outputted in time series from the line sensor to determine the absolute position or angle of the moving object. , the plurality of pieces of address information and pattern information form one block each, and the address of each block is specified by a pattern arrangement obtained from the plurality of pieces of address information forming each block. An encoder characterized by: 2. In the device described in claim 1, by creating a difference between the number of information of the pattern information and the corresponding number of sensor elements of the line sensor, the number of sensor elements of the sensor element is one pitch or less. 2. The encoder according to claim 1, wherein the encoder performs interpolation with a resolution of . 3. In the device described in claim 1, the processing circuit uses the address information of the block specified by the pattern arrangement obtained from the plurality of pieces of address information and the line sensor's information obtained from the reference position information. Claim 1, characterized in that the absolute position or angle of the moving body is determined from read information up to 1 pitch unit and interpolated information within 1 pitch unit of the line sensor. Encoder listed. 4. The device according to claim 1, characterized in that the address information forming part of the predetermined pattern array is composed of 3 or more bits on the line sensor. An encoder according to claim 1. 5. The device according to claim 1, characterized in that a plurality of pieces of pattern information constituting each block are interpolated together at a resolution of 1 bit or less of the sensor element. An encoder according to claim 1. 6. In the item described in claim 1, the address information and the pattern information are each constituted by optical information formed from optical gratings having different light transmittances or reflectances. An encoder according to claim 1, characterized in that: 7. The device according to claim 1, wherein the predetermined pattern arrangement arranged on the code plate is constituted by an M-sequence arrangement. encoder. 8. In the item described in claim 1, the address information and the pattern information are each constituted by magnetic or electromagnetic induction information,
2. The encoder according to claim 1, wherein the line sensor is constituted by a magnetically sensitive element array.