JPH074991A - Method for detecting one-rotation signal of encoder and encoder - Google Patents

Abstract

PURPOSE:To avoid erroneous detection of a reference position due to edge cracking of a one-rotation signal pulse without signal delay by selecting one of a rising edge and a falling edge corresponding to a relation between an edge on a specific side and an encoder rotating/moving direction. CONSTITUTION:A-phase and B-phase pulse signals PA, PB are input to an up/down detection circuit 1, while a Z-phase signal (one-rotation signal) PZ is input to a Z-phase signal edge detection circuit 4. An up signal D+ and a down signal D- indicating a CW direction and a CCW direction rotation of an encoder based on the signals PA, PB are output and input to addition side and subtraction side ends of a CW direction counter 2 and a CCW direction counter 3. The circuit 4 detects all rising/falling edges included in the Z-phase signal and outputs an edge signal Z. Then, a determining process including AND and OR circuits eliminates erroneous detection due to edge cracking occurring in a rising/falling part of a rectangular pulse.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting one rotation signal of an encoder used for detecting the rotation position of a motor or the position of a table of a machine tool, and a method for detecting the one rotation signal. An encoder with means.

[0002]

2. Description of the Related Art A plurality of sets of a plurality of rectangular pulse signals (for example, A phase and B phase) having a phase difference as a position signal, which are attached to a rotor shaft of a motor, and a code signal representing an electric angle of the motor are included in each rotation. An encoder that generates one rectangular pulse signal (Z-phase signal) is already known.

By the way, the rectangular pulse signal forming one rotation signal in such an encoder is a detection signal pulse (rising / rise) from a light detecting means or a magnetic detecting means provided in the vicinity of a rotating code plate constituting the encoder. A sine / cosine function-like waveform whose falling edge is not sharp.) Is converted into a rectangular pulse whose rising / falling is clear through a waveform shaping circuit.

In general, since the waveform shaping circuit includes a comparator element, if the detection signal from the photodetection means or the magnetic detection means includes fluctuation components due to various noises, it is rectangular. A so-called edge crack may occur in a portion very close to the rise or fall of the pulse.

When such edge breakage occurs, if the rising / falling edge of the specific side of the rectangular pulse is detected in accordance with a conventionally used method, detection of one rotation signal is hindered and an incorrect position is detected. Detection will be performed. The reason will be described with reference to FIGS. 1 (1) and 1 (2). FIG. 1 (1) shows a situation in which normal rectangular pulses PL1 and PL2 (pulses for two rotations) without edge cracks are obtained. In this case, in order to obtain the detection signal that determines the reference position for each rotation, the specific-side edges EG1 and EG2 are moved in the direction of rotation of the rotary shaft to which the encoder is attached or the code plate (hereinafter, simply referred to as the direction of movement). Also, it is detected based on the determination result of rising / falling.

The determination of the moving direction is made by, for example, processing an A-phase signal and a B-phase signal by a circuit using a clock pulse generator and a data flip-flop (D-FF). This is performed by identifying and detecting the down count signal D-.

In the example shown here, when the moving direction is clockwise CW, the rising edge is detected twice in total in the order of EG1 → EG2, and when it is counterclockwise CCW, Falling edges are detected twice in total in the order of EG2 → EG1. In either case, since there is no edge crack, the intended one-rotation signal detection is executed corresponding to each reference position passage, and the correct reference position signal is obtained.

On the other hand, as shown in FIG. 1B, when the above detection method is directly applied to the rectangular pulse having the edge crack, EG1 is detected in the edge detection process for two clockwise CW rotations. → EG1 '→ EG2 → EG
2 ′ in total, four times in total, EG2 ′ → EG2 → EG1 ′ → EG in the edge detection process for counterclockwise CCW2 rotation.
The edge detection is performed four times in the order of 1. That is, originally, the point where the specific side edge for two rotations in the CW direction or the CCW direction should be detected is
Due to the edge crack, the result of performing the reference position detection for a total of four rotations is brought about.

The one-rotation signal erroneous detection operation including the duplicate detection of the specific side edge naturally lowers the position detection accuracy of the encoder. In particular, when the edge cracking phenomenon frequently occurs, an extremely large position measurement error may occur. Therefore, in order to maintain the reliability of the encoder,
There has been a strong demand for solving the problems caused by the edge cracks.

Conventionally, as a technique for preventing erroneous detection of one-rotation signal of the type described above, a method of masking an edge crack of a rectangular pulse by a filter circuit process has been adopted, but a signal delay due to the filter process cannot be avoided. However, it was difficult to say that it was a satisfactory solution.

[0011]

DISCLOSURE OF THE INVENTION As described above,
In the prior art, the method of detecting the specific side edge of the one rotation signal pulse of the encoder based on only the moving direction and the instantaneous pulse rising / falling edge discrimination result is adopted, and therefore, the reference is determined by edge cracking. The number of times of position detection may deviate significantly from the original value, and a decrease in position detection accuracy due to erroneous detection of one rotation signal cannot be avoided. Further, when the filter processing is performed to mask the edge crack, there is a problem of signal delay.

Therefore, an object of the present invention is to provide a method for detecting a single rotation signal in an encoder which can avoid erroneous detection of a reference position due to edge breakage of a single rotation signal pulse without performing a filtering process with a signal delay. An object of the present invention is to provide an encoder having means for executing such a one-rotation signal detection method.

[0013]

DISCLOSURE OF THE INVENTION The present invention is an encoder for detecting a one-rotation reference position of the encoder by detecting a specific side edge of a rectangular pulse generated every one rotation of a rotary shaft to which the encoder is attached. 1 rotation signal detection method; edge rising / falling selection condition for selecting one of rising edge / falling edge according to the relationship between the specific side edge and the encoder rotation movement direction, and the 1 rotation signal at the time of edge detection Only when both of the history conditions regarding the level state are satisfied, a process of determining the detected edge as the specific side edge is included, and the history condition is such that an edge crack occurs near the end of the rectangular pulse. In this case, by setting the condition to suppress the specific side edge duplication detection operation, Those given basic solution to problems arising in connection with Tsu di cracking phenomenon. (Structure according to claim 1).

Further, the present invention is an encoder provided with means for executing such a one-rotation signal detecting method,
Means for generating one rotation signal including one rectangular pulse for each rotation, means for generating a signal including a large number of pulses within one rotation, and detecting a specific side edge of the rectangular pulse included in one rotation signal And a means for outputting the specific side edge detection output; the rectangular pulse specific side edge detection means discriminates the rotation direction of the encoder based on a signal including a large number of pulses generated within one rotation. Rotation direction discrimination means, edge selection detection means for selectively detecting the rising edge or falling edge of the rectangular pulse corresponding to the rotation direction discriminated by the rotation direction discrimination means, and edge selection detection by the edge selection detection means Occasionally, means for detecting the level state history of one rotation signal within a period going back a predetermined stroke from the detection time point, and detection output of the level state history detecting means Based on includes a means for generating a particular side edge detection output, retrospectively predetermined stroke, 1
The present invention provides an encoder that is set in correspondence with an edge crack occurrence section that is assumed for a rectangular pulse included in a rotation signal.

[0015]

According to the present invention, the specific side edge of the one-rotation signal pulse of the encoder is not detected only based on the moving direction and the instantaneous rising / falling edge discrimination result, but the rising / falling edge of the rectangular pulse is detected. By introducing a decision process that includes logic that eliminates erroneous detection due to edge cracks that occur in the falling part, it is possible to prevent duplicate detection of specific-side edges, which reduces encoder position detection accuracy and reliability. Certainly prevented.
The essential points of the judgment standard peculiar to the present invention will be described below with reference to FIG.
A case of detecting an edge-breaking rectangular pulse similar to the above will be described with reference to FIG.

As described in the related explanation part of FIG. 1 (2), when edge breakage occurs in the two rectangular pulses PL1 and PL2 at the positions shown in the figure, when viewed in the CW / CCW direction, a total of four are generated. Rising / falling edge EG1, EG
1 ', EG2, EG2' will be generated. In the present invention, in order to prevent erroneous detection of the reference position edge, the state of the detection signal (Z-phase signal) forming one rotation signal (H; high / L; low state, hereinafter simply referred to as “H” state, ” L state) is continuously monitored, erroneous detection is suppressed in consideration of the state of the Z-phase signal and the movement section, and the original number of reference position detection times is substantially matched.

That is, the edge crack has a property that it is generated in the vicinity of both ends of the rectangular pulse, as expected from the cause thereof, and therefore, in order to avoid the erroneous detection operation, these adjacent edges are not included. Paying attention to the fact that it is only necessary to suppress the individual continuous detection of, the method of limiting the detection of the edge at the specific side end of the rectangular pulse to only once in any movement state in the CW direction and the CCW direction is adopted. Is.

Specifically, as a judgment criterion for avoiding an erroneous specific side edge detection operation at the time of one rotation signal erroneous detection,
The following [I] to [III] are introduced. [I] A predetermined movement section in which the movement direction is CW and the "L" state of the Z-phase signal exceeds the edge cracking time (β described later)
When the rising edge is detected after the above, it is determined that the edge detection is not an erroneous detection. [II] The moving direction is CCW and the Z-phase signal is "H".
When the trailing edge is detected after a predetermined movement section in which the state exceeds the edge cracking time (here, β is the same as the condition [I], but not necessarily the same), the edge detection is erroneous. Judge that it is not detected. [III] The rising edges and falling edges detected in cases other than the above [I] and [II] are
All are judged not to be on the specific side edge.

This condition will be described with reference to the example of FIG. First, if the section where the edge crack is expected to be expressed is expressed by the encoder moving distance measured from the end of each rectangular pulse, and the section set with some margin is represented by β, the edge in each rectangular pulse Regarding the sizes α1 and α2 of the cracked section, β> α1 and β> α2 are established. That is, β> α with respect to the generally assumed maximum edge cracking interval α.
The value of β is set so that However, it should be noted that if the value of β is set too large, the tendency to exclude even the edge that should originally be detected as the specific side edge becomes stronger. Generally, a sufficiently small value is selected as compared with the rectangular pulse width w and the rectangular pulse interval γ. That is,
w >> β and γ >> β.

Now, considering the case where the edge detection is started from the position Q1 shown in the figure and the edge moves in the CW direction, the first rising edge that arrives is EG1 '. In the conventional method, the edge EG1 'is mistakenly recognized as the specific side edge, but the edge cracked portion "
The size of the L ″ state duration section is obviously β or less, and in the method of the present invention, the above conditions [I] and [III] work, and this edge EG1 ′ is not judged to be the specific side edge.

Then, when the movement in the CW direction is further continued and the edge EG2 arrives, since γ >> β, "
The condition [I] is satisfied by the history of the L ″ state being maintained, and it is determined that the edge is the specific side. Further, immediately after EG2 (after α2 movement), a rising edge EG2 ′ arrives, but α2 < From β, neither of the conditions [I] and [II] is satisfied, and it is not determined that the edge is on the specific side according to the condition [III] .In the end, duplicate detection due to edge cracking is avoided for both rectangular pulses PL1 and PL2.

Similarly, considering the case of starting edge detection from the CCW position and moving in the CCW direction from Q2 in the figure, the first arriving falling edge is EG2 '. Unless Q2 is close to the edge EG2 '(less than or equal to β), the edge E is determined by the history of the "H" state being maintained.
G2 ′ is determined to be the specific side edge. Then, immediately after this (after the movement of α2), the falling edge EG2 arrives, but since α2 <β, neither of the conditions [I] and [II] is satisfied, and according to the condition [III], the specific side edge is Not judged. Therefore, the specific-side edge detection is counted only once for the rectangular pulse PL2.

Further, when the falling edge EG1 'of the edge-breaking portion of the rectangular pulse PL1 arrives, the "H" state is maintained for the duration of β or less, so it is not judged as the specific side edge. Then, when the original falling edge EG1 arrives, the condition [II] that the history of the “H” state is β or more is satisfied, and this edge EG1 is determined to be the specific side edge.

Here, EG2 ', not edge EG2, is judged to be the specific side edge, and the encoder generates the reference position signal, so a position detection error of at most α occurs, but α Since the value of is usually extremely small, its adverse effect is generally small enough to be ignored as compared with duplicate detection. (Originally, it is not always possible to uniquely determine which of EG2 and EG2 'truly corresponds to the reference position). Further, when the encoder makes a reciprocating motion near the edge EG1 or EG2, there is a possibility of failing to detect what should originally be determined / detected as the specific side edge, but α is a rectangular pulse. Since the width w and the rectangular pulse interval γ are usually extremely small, by setting β sufficiently small, the probability of causing this situation approaches 0, and the influence on the position detection accuracy causes a practical problem. It is relatively easy to suppress it to the extent that it does not exist.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 3 illustrates, in the form of a principal part block diagram, the configuration of a signal processing circuit of an encoder provided with means for executing the one-rotation signal detection method of the present invention.

At the input portion of the signal processing circuit shown in FIG. 3 as a whole, A-phase and B-phase signals PA and PB consisting of two sets of rectangular pulse trains having a phase difference of 90 degrees, and FIG. The Z-phase signal PZ including one rectangular pulse as shown in FIG. 2 is input for each rotation of the rotor or the encoder code plate. Such a configuration is common to known typical encoders.

The signal input unit is an up / down detection circuit 1
And a Z-phase signal (1 rotation signal) edge detection circuit 4, and the up / down detection circuit 1 receives the A-phase and B-phase pulse signals PA and PB, and the Z-phase signal (1 rotation signal). The Z-phase signal (one rotation signal) PZ is applied to the edge detection circuit 4.
Is entered.

The up / down detection circuit 1 is of a usual type incorporating a D flip-flop, a clock pulse generator, etc., and moves (rotates) in the CW direction of the encoder based on the A phase and B phase signals PA and PB. ) Is output, and a down signal D- corresponding to CCW direction movement (rotation) is output.

Up signal D of the up / down detection circuit 1
+ Is input to the addition side terminal of the CW direction counter 2 and the subtraction side terminal of the CCW direction counter 3 in the next stage. On the other hand, the down signal D- is input to the subtraction side terminal of the CW direction counter 2 and the addition side terminal of the CCW direction counter 3.

On the other hand, the Z-phase signal PZ is input to the Z-phase signal (one rotation signal) edge detection circuit 4. The Z-phase signal (single rotation signal) edge detection circuit 4 detects all edge rising / falling edges included in the Z-phase signal and outputs the edge detection signal Z, and distinguishes between rising / falling edges. The rising edge detection signal Zr or the falling edge detection signal Zf is selectively output according to the above.

The Zr signal is input to one terminal of the AND gate AND1 and the Zf signal is input to the AND gate AND2.
Input to one side terminal. Further, the Z signal is input to the counter clear control circuit 5, and the counter clear control circuit 5 immediately after receiving the Z signal (1 clock later).
The counter clear signal CLR is sent to the count clear terminals of the CW direction counter 2 and the CCW direction counter 3.

The CW direction counter 2 is set with a full count value F1 and a lower limit count value of 0, and unless the counter clear signal CLR is received, the lower limit value 0 to
The up-down counting operation is continued within the full count value F1. That is, +1 is counted (added) upon receiving the signal D + from the up / down detection circuit 1, and -1 is counted (subtracted) at the input of the signal D-, but the subtraction is stopped when the count value becomes 0. When the count value reaches F1, the addition operation is stopped and "1" is output as a full count signal to one input terminal of the AND gate AND1.

However, at any time, when the count clear signal CLR is received from the counter clear control circuit 5, the count value is immediately reset to 0. Here, 15 is set as the full count value. This full count value is set for the time being to determine that the encoder is moving in the CW direction, but in the present embodiment, as described later, the overlap detection suppression section length β is set. This is the amount that is also used as the corresponding amount.

Similarly, in the CCW direction counter 3, F2 = 15 which is equal to F1 is set as the full count value F2, and 0 is set as the lower limit count value.
Then, as long as the counter clear signal is not received, the up-down counting operation is continued between the lower limit value 0 and the full count value 15. However, in response to the fact that this counter detects the movement in the CCW direction, it receives the signal D + from the up / down detection circuit 1 and counts (subtracts) -1 to obtain the signal D- Count +1 on input (addition)
To do. When the count value becomes 0, the subtraction is stopped, and when the count value becomes F1, the addition operation is stopped and "1" is output as a full count signal to one input terminal of the AND gate AND2.

Similar to the CW direction counter 2, the count value is immediately reset to 0 when the count clear signal CLR is received from the counter clear control circuit 5 at any time. The full count value F2 = 15 of the CCW direction counter 3 is also set for the time being to determine that the encoder is moving in the CCW direction, but like the full count value F1 = 15 of the CW direction counter 2, Here, it also has a meaning as an amount corresponding to the overlap detection suppression section length β (details will be described later).

Now, as shown, the AND gate AN
In D1, "1" is output to the OR gate OR1 only when the rising edge is detected (Zr input) while the full count signal "1" is output from the CW direction counter 2, and Gate AND2
, The falling edge is detected while the CCW direction counter 3 outputs "1" (Zf input)
Only when this is done, "1" is output to the OR gate OR1. When any one of these conditions is satisfied, the OR gate OR1 outputs the reference position detection output signal (specific side edge detection signal) ZEG as if the specific side edge of the rectangular pulse included in the Z-phase signal is detected. Is output.

Therefore, if the function of the counter clear control circuit 5 is excluded, the signal processing section shown in FIG. 2 functions as a circuit for carrying out the one-rotation signal detection method according to the prior art. Therefore, the counter clear control circuit 5 is not necessary if there is no possibility of edge cracks occurring in the rectangular pulse of the Z-phase signal.

However, the rectangular pulse shown in FIG.
When edge cracks such as those shown in are generated, as described above,
Overlap counting of rising edges (EG
1 and EGE1 ′; EG2 and EG2 ′), and the falling edge is counted twice while moving in the CCW direction (EG2 ′ and EGE).
2; EG1 'and EG1). The reason why the duplicate detection is suppressed by the configuration of this embodiment will be described below with reference to the time charts of FIGS. 4 and 5 in addition to the reference diagrams.

FIG. 4 is a time chart showing the transition of each signal when the encoder moves in the CW direction from the position Q1 shown in FIG. First, when the movement in the CW direction is started from the position Q1, a large number of D + signal pulses representing the displacement in the CW direction are continuously input to the CW direction counter 2. C
The W direction counter 2 counts this pulse and reaches the full count value F1 = 15 within a short time. In this state,
One input terminal of the AND gate 1 becomes "1".

When the encoder moves in the CW direction and reaches the falling edge eg1, the Z signal and the Zf signal change to Z
The phase signal (one rotation signal) is output from the edge detection circuit 4. The Z signal activates the counter clear control circuit 5,
The count value of the CW direction counter 2 is cleared 1 count after receiving the Zf signal, and the count value is changed from 15 to 0. The count value of the CCW direction counter 3 is 0 all the time, and there is no change due to the count clear. On the other hand, the Zf signal is sent to the AND gate AND2, but since the CCW direction counter 3 does not output the full count signal,
The output of the AND gate AND2 never becomes "1". Therefore, the specific side edge detection signal ZEG is not output from the OR circuit OR1.

After passing the falling edge eg1, the encoder reaches the rising edge EG1 'in a short time. At this point in time, it is considered that the count value of the CW direction counter 2 has increased corresponding to the number of D + signal pulse inputs after the count clearing due to the passage of the edge eg1. However, this count value does not exceed the value corresponding to the edge cracking section α1 (for example, about 10 pulses), and naturally, the value corresponding to the maximum value α assumed as the edge cracking section (for example, about 12 pulses). It does not exceed.
The full count value F1 of the CW direction counter 2 is
Since it is set to a value (here, F1 = 15) corresponding to β which is not smaller than this α (usually slightly larger),
The full count signal is not sent from the CW direction counter 2 to the AND gate AND1 when the edge EG1 'is detected. Therefore, the edge EG1 'is prevented from being detected as the specific side edge.

Upon detection of the edge EG1 ', a counter clear signal is sent from the counter clear control circuit 5 to the CW direction counter 2 and the CCW direction counter 3, and the edge eg1' is detected before and after the count is cleared. Although detected, this edge cannot be detected as the specific side edge because it is a falling edge.

Immediately after the detection of the falling edge eg1 ', the counting of the CW direction counter 2 is started, reaching 15 counts at the time when a slight distance corresponding to the distance β is passed, and the AND gate A is reached.
The full count signal is output to ND1. This state is held until the counter clear control circuit 5 outputs a count clear signal to the CW direction counter 2.

When the encoder continues to move and reaches the rising edge EG2 of the next rectangular pulse PL2, the Z-phase signal (one rotation signal) edge detection circuit 4 sends the rising edge detection signal Zr to the AND gate 1 and
The Z signal is sent to the counter clear control circuit 5. The counter clear control circuit 5 clears the count value of the CW direction counter 2 after one count, but immediately before that, the state in which both the input sides of the AND gate AND1 are in the "1" state is realized, so that the "1" signal is output. , OR gate OR1, and the specific side edge detection signal ZEG is output.

After detecting the edge EG2, a short section (α2
In the following), the falling edge eg2 and the rising edge EG2 'are detected by the Z-phase signal (one rotation signal) edge detection circuit 4, and Zf, Z signals and Zr, Z signals are AND gates AND1, 2 and. Although it is output to the counter clear control circuit 5, neither the CW direction counter 2 nor the CCW direction counter 3 outputs the full count signal, so that the ZEG signal is not generated.

Since the edge eg2 'detected after the movement of approximately one rectangular pulse is the falling edge,
It cannot be detected as a specific side edge.

As described above, even if an edge crack occurs in the Z phase signal (one rotation signal) of the encoder in the CW moving state, erroneous detection due to duplicate detection of the specific side edge does not occur. The circuit operation described above is basically the same even if the starting point of edge detection is other than Q1 shown in the figure.

Next, FIG. 5 is a time chart showing the transition of each signal when the encoder moves in the CCW direction from the position Q2 shown in FIG. However, for the sake of convenience, the time traveling direction is set to the direction from the right side to the left side on the paper in accordance with the WCCW direction.

When the movement in the CCW direction starts from the illustrated position Q2, a large number of D- signal pulses representing the displacement in the CCW direction are continuously input to the CCW direction counter 3. The CCW direction counter 3 counts this pulse,
The full count value F2 = 15 is reached within a short time. In this state, one input terminal of the AND gate 2 is in the "1" state.

When the encoder moves in the CCW direction and reaches the falling edge EG2 ', the Z signal and Zf signal are output from the Z phase signal (one rotation signal) edge detection circuit 4, and the Z signal is the counter clear control circuit. Activate 5
The count value of the CCW direction counter 3 is cleared from 15 to 0 after 1 count after receiving the Z signal (the count value of the CW direction counter 2 is 0 from the beginning to the end).

On the other hand, the Zf signal is sent to the AND gate AND2, but at that time, the CCW direction counter 3 is in the state immediately before the count clear (for one count) and outputs the full count signal, so the AND gate AND2.
Receives the Zf signal and sets "1" to the OR gate OR1
Output to. As a result, the specific side edge detection signal (reference position detection signal) ZEG is output.

The encoder reaches the rising edge eg2 in a short time after detecting the falling edge EG2 '.
Since the count value of the CW direction counter 2 is 0 all the time,
There is no room for detection of the rising edge eg2 as the specific side edge.

Upon detection of the edge eg2, a counter clear signal is sent from the counter clear control circuit 5 to the CW direction counter 2 and the CCW direction counter 3, and the edge EG2 is detected before and after the count is cleared. It At this point, the CCW corresponding to the number of D- signal pulse inputs after the count is cleared accompanying the detection of the edge eg2
It is considered that the count value of the direction counter 3 is increasing.

However, this count value must exceed a value corresponding to the edge cracking section α 2 (for example, about 10 pulses) or a value corresponding to the maximum value α assumed as the edge cracking section (for example, about 12 pulses). There is no. And
The full count value F2 of the CCW direction counter 3 is also this α
The value corresponding to β (usually slightly larger) (here, F2 = 15) is set, so that the full count signal is sent from the CCW direction counter 3 to the AND gate AND2 when the edge EG2 is detected. There is no. Therefore, the edge GE2 is not detected as the specific side edge, and overlapping detection of the specific side edge of the rectangular pulse PL2 is avoided.

Immediately after the detection of the falling edge EG2, the CCW direction counter 3 starts counting, reaches 15 counts at the time when a slight distance corresponding to the distance β has passed, and the AND gate A
The full count signal is output to ND2. This state is held until the counter clear control circuit 5 outputs a count clear signal to the CCW direction counter 3.

When the encoder continues to move, the next rising edge eg1 'of the rectangular pulse PL1 is changed to the Z-phase signal (1
Rotation signal) detected by the edge detection circuit 4, but CW
Since the count value of the direction counter 2 is always 0, even if the rising edge detection signal Zr is sent from the Z-phase signal (one rotation signal) edge detection circuit 4 to the AND gate 1, the CW
Since the full count signal of the direction counter 2 is not input, the "1" signal is not sent from the AND gate AND1 to the OR gate OR1, and the specific side edge detection signal ZEG is not output.

After detecting the edge eg1 ', a short section (α
(1 or less), the falling edge EG1 'and the rising edge eg1 are detected by the Z-phase signal (one rotation signal) edge detection circuit 4, and Zf, Z signals and Zr, Z signals are AND gates AND1, 2 respectively. And the counter clear control circuit 5, the CW direction counter 2, CCW
Since no full count signal can be output from any of the direction counters 3, no ZEG signal is generated.

Then, after detecting the edge eg1, the interval distance β
At the time of movement, the D- signal pulse reaches the full count of 15,
One input of the AND gate AND2 is set to the "1" state. In this state, the edge EG1 detected by the encoder after the movement of approximately one rectangular pulse is the falling edge,
The Z signal and the Zf signal are output from the Z-phase signal (one rotation signal) edge detection circuit 4, and the Z signal activates the counter clear control circuit 5 so that the CCW direction counter 3 of the CCW direction counter 3 is counted one count after receiving the Z signal. The count value is cleared from 15 to 0 (the count value of the CW direction counter 2 remains 0 and does not change).

On the other hand, the Zf signal is sent to the AND gate AND2, but at that time, the CCW direction counter 3 is in the state immediately before the count clear (one count) and is outputting the full count signal.
2 receives "1" by receiving the Zf signal or OR gate OR1
Output to. As a result, the specific side edge detection signal (reference position detection signal) ZEG is output.

As is apparent from the above description, even when the encoder is in the CCW moving state, erroneous detection due to duplicate detection of the specific side edge does not occur due to edge cracking of the Z phase signal (one rotation signal). Even if the start point of edge detection is other than Q2 shown, the circuit operation described above is basically the same as in the case where the encoder moves in the CW direction.

Further, the specific numerical values such as the full count values F1 and F2 shown in the above description are merely examples, and it is not intended to limit the invention of the present application. You won't need it. For example, when the code plate of the encoder is densely coded, the pulse intervals of the A phase, the B phase, etc. are narrowed. Therefore, in general, the pulse full count numbers F1, F corresponding to the section β are used.
The value of 2 will also be chosen large.

F1 corresponding to the section β where the edge crack occurs,
When actually determining the value of F2, switch F1, F2 to 1, 2, 3 ..., and so on, and rotate the encoder in the CW and CCW directions under the conditions as close as possible to the actual usage conditions. A method is conceivable in which the number of output pulses or the waveform of the ZEG signal is measured or observed by a measuring instrument to search for a value at which the detection of the specific side edge does not occur.

As a special case, when the encoder makes a reciprocal motion in the vicinity of the edge EG1 or EG2 (in the vicinity of β or less), it is necessary to detect what should originally be determined / detected as the specific side edge. However, since α is usually much smaller than the rectangular pulse width w and the rectangular pulse interval γ, β (F1, F2)
It is preferable that the probability that this situation will occur be close to 0 by setting the value to be sufficiently small. In the above embodiment, the counting operation of the CW direction counter 2 or the CCW direction counter 3 is not stopped even if the ZP signal is in the H state / L state, but the ZP signal is in the H state. Even when a configuration is adopted in which the counting operation of the CW direction counter 2 is stopped during the certain period and the counting operation of the CCW direction counter 3 is stopped during the L state,
It is clear from the description of the above embodiment that there is no problem in executing the detection operation of the one-rotation signal according to the technical idea of the present invention.

[0064]

According to the invention of the present application, even if an edge crack phenomenon occurs in one rotation signal in the encoder, erroneous detection of the specific-side edge is suppressed, so that the position accuracy of the encoder is prevented from lowering. Particularly, even when edge cracks frequently occur, it is possible to avoid causing a serious position measurement error, and therefore reliability of the encoder is ensured.

Further, when a filter circuit is used for edge suppression, the adverse effect of edge cracking can be eliminated without accompanying the inevitable signal delay, so that the responsiveness of the encoder can be kept high from the viewpoint of the present invention. Has great advantages. Therefore, when the encoder or one-rotation signal detection method of the present invention is incorporated as position detection means in a servo control system, it is possible to avoid a decrease in position measurement accuracy due to edge cracking without impairing the responsiveness of the servo control system. .

[Brief description of drawings]

FIG. 1A is a diagram for explaining detection of a specific-side edge of a rectangular pulse for two normal rotations without edge cracking, and FIG. 1B is a diagram for explaining edge cracking in these rectangular pulses. It is a figure for explaining the situation in a case.

FIG. 2 is a diagram for explaining the erroneous detection avoidance principle of the present invention in the case of edge-breaking rectangular pulse detection similar to that of FIG. 1 (2).

FIG. 3 is a principal block diagram illustrating the configuration of a signal processing unit of an encoder including means for executing the one-rotation signal detection method of the present invention.

FIG. 4 is a time chart showing the transition of each signal when the encoder moves from the position Q1 shown in FIG. 2 in the CW direction.

5 is a time chart showing the transition of each signal when the encoder moves in the CCW direction from the position Q2 shown in FIG.

[Explanation of symbols]

1 Up / Down detection circuit 2 CW direction counter 3 CCW direction counter 4 Z-phase signal (1 rotation signal) edge detection circuit 4 5 Counter clear control circuit PLl, PL2 Rectangular pulse EG1, EG1 ', EG2, EG2', eg1, eg
1 ', eg2, eg2' rising / falling edge

Claims (2)

[Claims] 1. A method for detecting one rotation signal of an encoder, which detects a one rotation reference position of the encoder by detecting a specific side edge of a rectangular pulse generated for each rotation of a rotary shaft to which an encoder is attached; Both the edge rising / falling selection condition that selects one of the rising edge / falling edge according to the relationship between the specific side edge and the encoder rotation movement direction, and the history condition regarding the level state of the one rotation signal at the time of edge detection. The process includes determining a detected edge as the specific side edge only when the following condition is satisfied:
The one-rotation signal detection method for the encoder, which includes a condition for suppressing a specific-side edge overlap detection operation when an edge crack occurs near an end of the rectangular pulse. 2. A means for generating one rotation signal including one rectangular pulse for each rotation; a means for generating a signal including a large number of pulses within one rotation; a rectangular pulse included in the one rotation signal. Of the specific side edge of the rectangular pulse specific side edge detection means;
A rotation direction discriminating means for discriminating the rotation direction of the encoder based on a signal containing a large number of pulses generated during rotation; a rising edge of the rectangular pulse corresponding to the rotation direction discriminated by the rotation direction discriminating means. Alternatively, an edge selection detecting means for selectively detecting a falling edge; and a means for detecting a level state history of the one rotation signal within a period going back a predetermined process from the detection time point when the edge selection detecting means detects the edge selection. And; means for generating the specific side edge detection output based on the detection output of the level state history detection means,
The encoder is characterized in that the retrospective predetermined process is set in correspondence with an edge crack occurrence section assumed for a rectangular pulse included in the one rotation signal.