JPH0684974B2 - Travel time detection circuit - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to speed control of a controlled object, and in particular, a moving amount or rotation amount (hereinafter abbreviated as moving amount) of the controlled object is detected by an encoder, and the controlled object is controlled. The present invention relates to an improvement of a circuit that counts the time required for a certain amount of movement to obtain movement time data.

(Prior Art) According to the serial dot printer, therefore, the carriage is accelerated to a constant speed, the print head on the carriage is driven to print dots on the paper while maintaining the constant speed, and then the printer is decelerated and stopped at a predetermined position. In this device, dots cannot be printed at a predetermined position unless the position and speed of the carriage are accurately detected, and the character shape is disturbed.

For this purpose, the control circuit shown in FIG. 1 has been conventionally used.

FIG. 1 is a block diagram showing a conventional carriage control circuit, which is equipped with a microcomputer, ROM (load only memory), RAM (random access memory), I / O port, timer circuit, etc. A control unit, 2 is a motor drive circuit, 3 is a servo motor, 4 is a carriage, 5 is a slit disk and two optical sensors, and an A phase having a phase difference of 90 degrees each time the servo motor 3 rotates by a certain angle. An encoder that outputs a B-phase signal, 6 is a waveform shaping circuit, 7 is a regulator using a semi-fixed resistor, and 8 is each rising edge and falling edge of the A phase and B phase output from the waveform shaping circuit 6. To form a pulse, which is output to the UP signal line when the A phase is advanced, and is output to the DOWN signal line when the B phase is advanced (hereinafter referred to as edge direction detection circuit). ), 9 is an edge Is reset by the pulse output from the direction detection circuit 8, then the pulse interval counter circuit for counting between clock pulses until it is reset, 8a
Is an OR circuit.

The conventional carriage control circuit shown in FIG. 1 operates as follows. First, a control signal is given from the control unit 1 to the motor drive circuit 2, a control current is supplied to the servo motor 3, and the servo motor 3 rotates. When the servo motor 3 rotates, the carriage 4 is driven and the encoder 5
Outputs the A-phase and B-phase signals. This encoder 5
The output of is waveform-shaped by the waveform-shaping circuit 6 to be pulsed. The threshold level of the waveform shaping circuit 6 can be changed by the adjuster 7. The edge / direction detection circuit 8 forms a thin pulse at each edge of the A-phase and B-phase output pulses output from the waveform shaping circuit 6, and outputs the thin pulse to either the UP signal line or the DOWN signal line. This output pulse is directly fed back to the control unit 1 for position control, and at the same time, the pulse interval counting circuit 9 is used for speed control.
It is fed back to the control unit 1 via. The control unit 1
The digital movement time data obtained by the pulse interval counting circuit 9 is compared with a target value which is predetermined by the steady rotation speed of the servo motor 3, and when the rotation speed of the servo motor 3 is lower than the target value, the servo motor A control signal is output to the motor drive circuit 2 to increase the drive current of the servo motor 3 and to decrease the drive current of the servo motor 3 when the rotation speed of the servo motor 3 is higher than the target value.
By the above, the rotation speed of the servo motor 3 is controlled to a target speed. Further, the control unit 1 measures the up pulse and the down pulse obtained by the edge / direction detection circuit 8 (for example, an up / down counter is used inside the control unit 1, and when one of the up pulses is input, the up / down counter is Is incremented by 1 and the value of the up / down counter is decremented by 1 when the down pulse is input by 1), and the current rotation angle of the servo motor 3 is known. By the above, the servo motor 3 is stopped / reversed at the desired rotation angle.
A control signal is output to the motor drive circuit 2 so as to control the drive current and phase of the servo motor 3 so that operations such as speed switching can be performed.

In the servo system operating in this way, when the pulse interval of the pulses output from the edge / direction detection circuit 8, that is, the moving time does not accurately reflect the rotation speed of the servo motor 3, the control unit 1 performs the calculation. An error occurs in the moving time data to be generated, and the error causes uneven rotation of the servo motor 3, or even if the error is small enough not to cause uneven rotation, the fluctuation power is additionally consumed, and thus the servo motor 3 is consumed. As a result, the temperature rises.

In order to prevent the above phenomenon from occurring, the encoder 5 is manufactured so as to reduce the error in the interval between the sensors, the encoder 5 is mounted at a constant mounting angle, and the mutual position of the A phase and the B phase is adjusted by the adjuster 7. Phase difference is completely 90
There is a measure such as adjusting the threshold level so that the angle (electrical angle) is adjusted. However, the above has the drawback that the encoder 5 is expensive, and the above, the number of man-hours required for assembly and adjustment increases, and the total cost increases. There was a drawback that raised.

Even if all of the above measures are taken, it is not possible to prevent changes in the characteristics of the optical sensor due to temperature changes and changes in the output of the optical sensor due to adhesion of dust, oil, etc. It has the drawback of limiting the temperature range and requiring maintenance.

Further, this drawback similarly occurs in the magnetic encoder and the reflection type optical encoder.

(Object of the Invention) An object of the present invention is to detect an accurate moving time data by using an encoder with low accuracy, requiring no adjustment or maintenance, and automatically following a temperature change. It is in.

(Structure of the Invention) In order to achieve this object, the present invention outputs from an encoder that detects a movement amount or a rotation amount of a controlled object, a waveform shaping circuit that makes the output of the encoder into a pulse, and a wave shaping circuit. , An edge detection circuit that detects an edge of a pulse with periodical time variation, a counting circuit that counts the time between each edge detected by the edge detection circuit, and k counting circuits. Count value x _n , x _n-1 , x _n-2 ,
... x _{n-k + 1} , and movement time data y before 1 sampling
_The following calculation is executed from _n-1 and the corrected travel time data y
a moving time detection circuit comprising a digital filter for outputting _n , However, α is a constant in the range of -1 <α <0, and k is k ≧ 2.
, Or an encoder that detects the amount of movement or rotation of the controlled object, a waveform shaping circuit that turns the encoder output into a pulse, and a periodical temporal fluctuation that is output from the waveform shaping circuit. An edge detection circuit that detects the edges of the pulse that has become a pulse, a counting circuit that counts the time between each edge detected by the edge detection circuit, and k of the counting circuit.
A digital filter that outputs the corrected movement time data y _n by executing the following calculation from the +1 count values x _n , x _n-1 , x _n-2 , ... X _{n-k + 1} , x _nk A moving time detection circuit comprising However, β is a constant in the range of −0.73 <β <0, and k is k ≧
2 is provided.

(Embodiment) FIG. 2 is a block diagram showing an embodiment of the present invention.
Is a correction circuit composed of a recursive digital filter having a transfer function represented by the following equation (1).

However, in the operation of the embodiment having the above configuration, the threshold level of the waveform shaping circuit 6 cannot be adjusted, the output of the pulse interval counting circuit 9 is corrected by the correction circuit 10, and the control unit stores it as movement time data. Since it is exactly the same as the conventional one shown in FIG. 1 except that it is input to the correction circuit 10, the correction circuit 10 does not adjust the accuracy of the encoder 5 or the threshold level of the waveform shaping circuit 6, and changes the threshold level. The process until accurate travel time data is obtained will be described in detail below, and the other description will be omitted.

First, in the encoder 5, if the distance between the sensors with respect to the slit is deviated, a phase error occurs in the output waveform between each phase. Further, in the waveform shaping circuit 6, when the threshold level changes from the solid line to the broken line as shown in FIG. 3, the width of the output pulse changes as shown by the arrow from the solid line to the broken line.

The pulse train output from the edge / direction detection circuit 8 is as shown in FIG. 4 due to the above-described phase error and pulse width variation. That is, as shown in FIG. 4 (a), when the A phase and the B phase of the input pulse have a phase difference of 90 degrees and the pulse widths are equal, the output pulses are evenly spaced. However, if there is a phase error between the A and B phases of the input pulse, the output pulse will fluctuate as shown in FIG. 4 (b), and the A and B phases of the input pulse will change.
If the pulse width of the phase changes, the output pulse changes as shown in FIG. 4 (c). Actually, since the above-mentioned two kinds of factors are combined, the input pulse and the output pulse are as shown in FIG. 4 (d). FIG. 5 shows the pulse interval between the nth pulse and the (n + 1) th pulse of the output pulse shown in FIG. 4 (d). As is clear from FIG. 5, in the two-phase encoder, the same error occurs every four times. Therefore, the output value of the pulse interval counting circuit 9 also contains the same error every four times. By using the analogy that an error that repeats every four times is considered as a kind of harmonic wave and the fundamental wave is extracted, this kind of error can be removed by a digital filter. For this,
In the present invention, the pulse intervals shown in FIG. 5 above are treated as sample values. It should be noted that this phase error and pulse width variation are referred to as ambient temporal variation.

A cyclic digital filter having such characteristics has the following equation (2) between the input values x _n , x _n-1 , x _n-2 , x _n-3 and the output values y _n , y _n-1. To establish.

Where | a | <1 (x _n + x _n-1 + x _n-2 + x _n-3 ) / 4 in the above formula (2) is 4
This is an average taken to eliminate the repeat error for each time. Also, y _n-1 is the previous output value, and the coefficient (1-
α) and α are for weighting both in order to improve the response. When this equation (2) is z-transformed, the following equation (3) is obtained.

The above equation (3) is rearranged into the following equation (4).

Therefore, the transfer function of the correction circuit 10 becomes the above equation (1).

That is, the output value of the correction circuit 10 is none other than the moving average value obtained by weighting the past three moving time data and the current moving time data which are determined by the constant α.
Servo motor 3 because it can eliminate repeated errors
The travel time data that almost correctly reflects the current speed of

Next, a constant α is obtained in order to further embody the correction circuit 10. In order to evaluate the response of the cyclic filter having the characteristics shown in equation (1) to changes in the rotation speed of the servo motor, a ramp signal was input as input and the sampling number was set to 15
FIG. 6 shows the sum of squares of the difference between the input and the output up to U (the sampling value is 1, 2, 3 ... 15) calculated at 0.1 intervals from α of −0.9 to 0.9. Since the lower the sum of squares, the better the response, the range of α such that −1 <α ≦ 0 is good.

Furthermore, as a constant α that simplifies the actual circuit configuration The equation (1) when the above is set is as shown in the following equation (5).

FIG. 7 is a frequency characteristic diagram of the cyclic filter shown in equation (5). In FIG. 7, since the gain drops at the frequency ratios of 0.25 and 0.5, it is possible to eliminate the repetition of the error every four times.

Further, if the filter having the characteristic shown in the equation (5) is represented by the difference equation, the equation (6) is obtained.

In equation (6), x _n is the current input value, x _n-1 (i = 1,
2, 3) are input values before i sampling, y _n are current output values, and y _n-1 are output values one sampling before.

FIG. 8 is a diagram showing a configuration example of a digital filter having the characteristic of equation (5), where x _n is an input value, 11 is a unit delay element that delays by one sample time, 12 is a multiplier, 13 is an adder, y _n is an output value, and the subscript of the multiplier 12 indicates the multiplication coefficient.
FIG. 9 is a block diagram of a circuit embodying the configuration example of FIG. 8. 20, 21, 22, and 27 are registers for holding one sample time data, 24 is a pulse interval counting circuit 9, and a register. A multiplexer that selectively outputs the output values of 20, 22, and 27, and an arithmetic unit 25 that has an accumulator 25a and adds or subtracts the output value of the multiplexer 24 to output to the register 27 and the control unit 1 of FIG. , 26 output timing signals to the pulse interval counting circuit 9, registers 20, 21, 22, 27 and arithmetic unit 25, respectively, and output to the multiplexer 24 a mode signal indicating a register to be selected, and a timing signal generation circuit ( Hereinafter referred to as a mode generation circuit).

The operation of the circuit shown in FIG. 9 will be described with reference to the flowchart of FIG. First, in step 1, an up pulse or a down pulse is input to the mode generation circuit 26. As a result, the mode generation circuit 26 starts operating,
In step 2, the operation of clearing the accumulator 25a of the arithmetic unit 25, selecting the value of the register 22 in step 3, inputting it to the accumulator 25a, and adding it is repeated three times. This gives a value of 3x _n-3 . Similarly, in step 4, the value in register 21 is stored three times in the accumulator 25a.
To obtain (3x _n-2 + 3x _n-3 ) and shift the value in register 21 to register 22. In step 5, register 20
The value of is added to the accumulator 25a three times, and (3x _n-1 + 3x
_n-2 + 3x _n-3 ) and the value in register 20 is registered in register 21
And further, in step 6, the value of the pulse interval counting circuit 9 is added to the accumulator 25a three times (3x _n + 3x
_n-1 + 3x _n-2 + 3x _n-3 ) and shift the value of the pulse interval counting circuit 9 to the register 20, reset the pulse interval counting circuit 9 in step 7, count clock pulses again, and calculate the pulse interval. Is counted and the value of register 27 is selected, and it is shifted to the upper 2 bits and the accumulator
Input to 25a and subtract. This gives (3x _n + 3x _n-1 + 3x _n-2
+ 3x _n-3 −4y _n-1 ) is obtained. In step 8, the value of the accumulator 25a is shifted to the lower 3 bits and the control unit 1
And register 27. As a result, the value output to the control unit 1 becomes y _{n in the} expression (6), and this value is held in the register 27 as y _n-1 . After this, the process returns to step 1 and the above-mentioned operation is repeated.

In the above description, by convention, the value is shifted from one register to another register, but this shift does not change the value of the original register.
To be precise, the contents of one register are read and written to another register.

As described above, in the first embodiment, of the errors included in the moving time data output from the pulse interval counting circuit 9, the error repeated every four times can be removed. Therefore, in the servo system using the two-phase encoder, it is not necessary to increase the manufacturing accuracy and the mounting accuracy of the optical sensor with respect to the slit of the encoder 5, and the adjuster for adjusting the threshold level of the waveform shaping circuit 6 is unnecessary. Of course, no adjustment work is required. Even if there are variations in the characteristics of the optical sensor of the encoder 5 and the circuit constants of the waveform shaping circuit 6, they are not affected. The encoder 5 and the waveform shaping circuit 6 due to temperature changes, aging changes, etc. It is possible to realize a servo system which is not affected by the characteristic change of the encoder 5 and the characteristic deterioration of the encoder 5 due to the adhesion of dust and oil and the like, and which has a uniform quality even with no adjustment and maintains stable characteristics for a long time.

Although the cyclic digital filter is used as the correction circuit in the above embodiment, the non-cyclic digital filter can be used as the correction circuit.

That is, when the correction circuit is a non-recursive digital filter, the input values x _n , x _n-1 , x _n-2 , x _n-3 , x _n-4 and the output value y _n are as follows (7). It suffices if the difference equation of the equation is satisfied.

The above equation (7) is weighted to improve the responsiveness by multiplying the average value of the four time series data including the current value and the average value of the four time series data including the one before the current value by a coefficient. It was done. This equation (7) is z-transformed,
When arranged, it becomes the following formula (8) Therefore, the transfer function H (z) is given by the following equation (9).

The following equation (10) can be obtained by rearranging the above equation (9).

FIG. 11 shows an example of the configuration of the non-recursive filter having the characteristics shown in the equation (10). With respect to the non-recursive digital filter having the characteristic shown in the equation (10), the responsivity to the change in the rotation speed of the servo motor was evaluated by the same method as in the case of the recursive digital filter. In Fig. 12, the ramp signal is input as input and the sampling number is 15
1,2,3 ... 15) and the sum of squares of the difference between the input and output is obtained by calculating β from −0.9 to 0.9 at 0.1 intervals. From FIG. 12, the range of β such that −0.73 ≦ β ≦ 0 is good. Comparing this with the characteristic of α in FIG. 19, it can be seen that the characteristic of β is better in the range of β described above, and that the response is better than that of the cyclic filter represented by the equation (1).

Furthermore, as a constant β that simplifies the actual circuit configuration, Equation (10) when set to is as follows.

FIG. 13 shows the frequency characteristic of the non-recursive digital filter having the characteristic shown in equation (11). As in Fig. 7, the frequency ratio is 0.25.
Since the gain drops at 0.5 and 0.5, it is possible to eliminate the repeating error every four times.

The following equation (12) can be obtained by expressing equation (11) as a difference equation.

FIG. 14 is a diagram showing a configuration example of a digital filter having the characteristic of the expression (11).

Also, FIG. 15 is a block diagram of a circuit embodying the configuration example of FIG. 14, and in comparison with the circuit shown in FIG. 9, the register 27 is removed and the register 23 is added, which will be described later. As described above, the operation of the mode generation circuit 26 is slightly different.

The operation of the circuit shown in FIG. 15 will be described with reference to the flowchart of FIG. First, in step 1, an up pulse or a down pulse is input to the mode generation circuit 26. This causes the mode generating circuit 26 to start operating, clears the accumulator 25a of the arithmetic unit 25 in step 2, selects the value of the register 23 in step 3, and selects the value of the register 23 via the multiplexer 24 in the accumulator 25 of the arithmetic unit 25.
Type in a and subtract. This gives a value of (-xn _-4 ). Similarly, in step 4, the value of the register 22 is added to the accumulator 25a to obtain (x _n-3 −x _n-4 ) and the value of the register 22 is shifted to the register 23. In step 5, the value of the register 21 is changed. Add to accumulator 25a, (x
_n-2 + x _n-3 -x _n-4 ) is obtained and the value of the register 21 is shifted to the register 22, and in step 6, the value of the register 20 is added to the accumulator 25a to obtain (x _n-1 + x _n-2 ＋ x _n-3 −
x _n-4 ), the value of the register 20 is shifted to the register 21, and in step 7, the operation of adding the value of the pulse interval counting circuit 9 to the accumulator 25a is performed twice, (2x _n + x
_n-1 + x _n-2 + x _n-3 -x _n-4 ) is obtained and the value of the pulse interval counting circuit 9 is shifted to the register 20, the pulse interval counting circuit 9 is reset in step 8, and the clock pulse is again supplied. The control unit 1 counts the pulse interval and shifts the value of the accumulator 25a to the lower 2 bits.
Output to. As a result, the value output to the control unit 1 is (1
It becomes y _{n in the} formula 2). After this, the process returns to step 1 and the above-mentioned operation is repeated.

In the circuit shown in FIG. 15, all the outputs of the pulse interval counting circuit 9 and the registers 20 to 23 are input to the multiplexer 24. However, as shown in FIG. 17, the outputs from the pulse interval counting circuit 9 and the register 21 are output. Can be operated as shown in the flowchart of FIG. 18 instead of directly inputting to the multiplexer 24 to simplify the circuit. This simplifies the circuit configuration of the multiplexer 24.

That is, as shown in step 5 and step 6 in FIG. 18 from step 5 in FIG. 16, the value of the register 21 is shifted to the register 22 and then input to the accumulator 25a via the multiplexer 24. Further, as shown in steps 18 to 11 in FIG. 18 in steps 7 and 8 in FIG. 16, the value of the pulse interval counting circuit 9 is shifted to the register 20 and then input to the accumulator 25a via the multiplexer 24. There is.

The present invention is not limited to the above embodiment, and various modifications can be made.

For example, when detecting the front and back edges of the k-phase encoder, it includes a repetitive error every 2k times, so the cyclic digital filter having the characteristic shown in the following equation (13) or the non-repetitive characteristic shown in the equation (14). The error can be removed by the cyclic digital filter.

However, | α | <1, k = 1,2,3 ... However, β is a constant of | β | <1, k = 1,2,3 ... Also, using a known calculation method, for example, instead of adding twice, the value is shifted up by 1 bit, or conversely subtracted twice. Instead of this, the value can be shifted to the lower bit by 1 bit to change the operation method, or the operation order can be changed. These selections simplify the circuit configuration and
It also changes depending on which of the higher calculation speeds is important. For example, if it is acceptable to reduce the calculation speed, a part of the RAM built in the control unit 1 is used as a substitute for the registers 20, 21, 22, 23 and 27, and a microcomputer is used as an arithmetic unit.
It is also possible to substitute the multiplexer or the like and store the arithmetic program in the ROM to execute the above-mentioned arithmetic. However, in many fields of use, it is necessary to complete these operations at high speed. Therefore, it is more practical to rely on the circuits shown in the block diagrams of FIGS. 2, 9, 15, and 17. Is. Moreover, since the versatility is high, it becomes easier to use the correction circuit as an integrated circuit.

The choice of mounting the encoder directly on the driven body or on the motor is a matter of design.

Furthermore, if the output of the k-phase encoder is pulsed and only one of the leading edge and the trailing edge is used, an error that repeats every k times occurs, so 2k in equations (13) and (14) is generated. It is necessary to replace k with k.
Furthermore, in FIG. 11, the count 1/4 is inserted between the adder and the output, but it may be inserted between the input and the first unit delay element. It is also possible to configure as shown in FIG. May be inserted immediately after the input as in the case of FIG.

Furthermore, the load terminals of the registers 20, 21, 22, and 23 are commonly connected to form a series of shift registers, and the contents of the shift register and the pulse interval counting circuit 9 are sequentially input to the accumulator through the multiplexer by the mode generating circuit, and the next step The method of shifting the contents of the shift register by one stage at the same time can be similarly performed. Furthermore, although the edge / direction detection circuit 8 also detects the direction, the movement direction (including the rotation direction) is one direction, and when the speed is controlled, only the edge need be detected.

(Effect of the invention) Since the present invention removes the cyclic error contained in the encoder output by the cyclic filter or the non-cyclic filter, accurate moving time data can be obtained even if a low-precision encoder is used. You can Moreover, since it is not necessary to adjust the threshold level of the waveform shaping circuit, it goes without saying that a threshold level adjuster is not necessary, but encoders and waveforms may change due to temperature changes, aging changes, and circuit constant variations. Even if the characteristic of the shaping circuit becomes different from the reference value, it is not affected, so that accurate moving time data can be easily and stably obtained. Then, the count value is simply averaged by a calculation by a digital filter to eliminate a repeating error caused by each slit, and a coefficient is multiplied to improve the responsiveness. Further, since correction can be performed for each edge, optimum control can be performed by using the servo system in the acceleration / deceleration region where the movement amount or the rotation amount changes with time.

Further, the present invention is not limited to the control of the printer carriage described above, but can be applied to a control system of a drive motor for a robot, VTR, floppy disk or the like.

[Brief description of drawings]

FIG. 1 is a block diagram showing a conventional carriage control circuit,
2 is a block diagram showing an embodiment of the present invention, FIG. 3 is a waveform diagram showing the input / output characteristics of the waveform shaping circuit, and FIG. 4 is the input / output characteristics of the edge / direction detection circuit 8 for various input pulses. 5 is a characteristic chart showing the n-th and n + 1-th pulse intervals of the output pulse shown in FIG. 4 (d), and FIG. 6 is the response of the cyclic digital filter shown in equation (1). To show the characteristics, the horizontal axis is α, and the vertical axis is the characteristic diagram in which the sum of squares of input and output errors is plotted. In FIG. 7, the frequency characteristic of the cyclic digital filter having the characteristic shown in the equation (5) is plotted on the horizontal axis. Ratio / o (o is the sampling frequency), a characteristic diagram showing the vertical axis as a gain, FIG. 8 is a diagram showing a configuration example of a cyclic digital filter having the characteristic shown in equation (5), and FIG. 9 is shown in FIG. Block diagram of a circuit embodying the configuration example shown,
FIG. 10 is a flow chart showing the operation of the circuit shown in FIG. 9, FIG. 11 is a diagram showing a configuration example of an acyclic digital filter having the characteristic shown in equation (10), and FIG. 12 is shown in equation (10). Characteristic diagram in which β is plotted on the horizontal axis and the sum of squares of input and output errors is plotted on the vertical axis to show the response of the acyclic digital filter.
The figure shows the frequency characteristics of a non-recursive digital filter with the characteristics shown in equation (11), where the horizontal axis is the frequency ratio / o (o is the sampling frequency) and the vertical axis is the gain. Figure 14 shows equation (11). Fig. 15 is a diagram showing a configuration example of a non-recursive digital filter having the characteristics shown in Fig. 15, Fig. 15 is a block diagram of a circuit example embodying the configuration example shown in Fig. 14, and Fig. 16 is an operation of the circuit shown in Fig. 15. FIG. 17 is a block diagram of another circuit example embodying the configuration example shown in FIG. 14, and FIG.
FIG. 19 is a flowchart showing the operation of the circuit shown in FIG. 19, and FIG. 19 is a diagram showing another configuration example of the non-recursive digital filter having the characteristic shown in equation (10). 1 ... Control unit, 2 ... Motor drive circuit, 3 ... Servo motor, 4 ... Carriage unit, 5 ... Encoder,
6 ... Waveform shaping circuit, 7 ... Adjuster, 8 ... Edge / direction detection circuit, 8a ... OR circuit, 9 ... Pulse interval counting circuit, 10 ... Correction circuit, 11 ... Unit delay element, 12 ...... Multiplier, 13 …… Adder, 20,21,22,23 and 27 …… Register, 24 …… Multiplexer, 25 …… Operation unit, 25a ……
Accumulator, 26 …… Mode generation circuit.

 ─────────────────────────────────────────────────── --- Continuation of the front page (56) References JP-A-54-17051 (JP, A) JP-A-54-143679 (JP, A) JP-A-58-109854 (JP, A)

Claims (2)

[Claims] 1. An encoder for detecting a movement amount or a rotation amount of an object to be controlled, a waveform shaping circuit for pulsing the output of the encoder, and a periodical temporal variation output from the wave shaping circuit. An edge detection circuit that detects the edges of the pulse that has become abnormal, a counting circuit that counts the time between each edge detected by the edge detection circuit, and the k count values x _n , x _n-1 , x of the counting circuit. _n-2 , ... X _{n-k + 1} , and a digital filter for executing the following calculation from the moving time data y _n-1 before one sampling and outputting the corrected moving time data y _n Travel time detection circuit. However, α is a constant in the range of -1 <α <0, and k is k ≧ 2.
And 2. An encoder for detecting a movement amount or a rotation amount of a controlled object, a waveform shaping circuit for pulsing the output of the encoder, and a periodical temporal variation output from the waveform shaping circuit. An edge detection circuit that detects the edges of the pulse that has been detected, a counting circuit that counts the time between each edge detected by the edge detection circuit, and k + 1 count values x _n , x _n-1 , x of the counting circuit _n-2 , ...
A movement time detection circuit including a digital filter that executes the following calculation from x _{n-k + 1} , x _nk and outputs corrected movement time data y _n . However, β is a constant in the range of −0.73 <β <0, and k is k ≧
Set to 2.