JP5580185B2 - Rotary encoder signal processing method - Google Patents

Description

  The present invention relates to a method for processing a pulse signal output from a rotary encoder.

  A rotary encoder used for observing the rotation state (rotation speed, rotation angle position, rotation direction, etc.) of various rotating bodies is, for example, as shown in Patent Document 1, the rotating body rotates by a predetermined angle. Each time, a pulse signal whose level value (voltage value) is switched is generated and output. In addition, the thing of patent document 1 is a contact-type rotary encoder which performs the contact conduction and the interruption | blocking with a sliding member and a metal for every predetermined rotation angle according to rotation of a rotary body. A rotary encoder is also generally known.

  A pulse signal output from this type of rotary encoder may cause chattering, bounce, or instantaneous noise in the vicinity of the level value switching. Here, more specifically, the “chattering phenomenon” is a phenomenon in which an ON contact repeatedly opens and closes due to an external cause. The “bounce phenomenon” is an undesirable intermittent switching phenomenon that occurs when a contact is turned on or off. In the present specification, unless otherwise specified, the term “chattering phenomenon” is used for the sake of convenience to include a bounce phenomenon in addition to the chattering phenomenon in its original meaning.

  When the chattering phenomenon or instantaneous noise occurs as described above, the level value of the pulse signal is switched multiple times within a short time even if the rotational angle position of the rotating body has hardly changed. As a result, the correspondence between the switching of the level value of the pulse signal and the actual rotational angle position of the rotating body is destroyed.

  For this reason, if the rotational state of the rotating body is grasped directly from the pulse signal output from the rotary encoder by an arithmetic processing unit such as a CPU, the rotational state cannot be correctly recognized.

  Therefore, conventionally, when the pulse signal output from the rotary encoder is input to an arithmetic processing device such as a CPU and the rotational state of the rotating body is grasped by the arithmetic processing device, the rotary encoder rotates the rotating body. Among the switching of the level value of the pulse signal to be output in response to the switching, the switching with the switching for returning the level value within a predetermined chattering exclusion time period set in advance after the switching (chattering exclusion time Switching of one or a plurality of pulses having a short pulse width included in the period of time) is excluded as switching to be excluded due to chattering phenomenon or instantaneous noise. The pulse output by the rotary encoder by executing the switching detection process that detects the occurrence of switching of the detection target in the arithmetic processing unit. So as to recognize the pulse signal waveform in the form of removal of the excluded switching despite the items, from the pulse signal waveform, it has been conducted to ascertain the rotational state of the rotating body.

JP 2005-195579 A

  By the way, when the rotational speed of the rotating body becomes high, the switching of the level value of the pulse signal generated depending on the rotational angle position of the rotating body (the switching of the detection target) is repeated at short time intervals. Become.

  On the other hand, in the conventional switching detection process, the chattering exclusion time is set to a constant value (fixed value). This constant value is set to a longer time so that the duration of the chattering phenomenon that can actually occur is surely within the period of the chattering exclusion time.

  For this reason, in the conventional switching detection process, when the rotational speed of the rotating body becomes high, the original switching of the level value of the pulse signal to be detected as the detection target switching (the rotational angle position of the rotating body is the rotary There is a case in which the switching to be generated at a predetermined angular position defined by the structure of the encoder is repeated within the chattering exclusion time period. In such a case, the original switching is excluded as an exclusion target switching in the same manner as the chattering phenomenon or temporary noise.

  Therefore, conventionally, when the rotating body rotates at high speed, the pulse signal waveform corresponding to the actual rotating state of the rotating body may not be correctly recognized, and as a result, the rotating state of the rotating body may be correctly grasped. There was a case that could not be.

  The present invention has been made in view of such a background, and when the rotational state of the rotating body is grasped based on the pulse signal output from the rotary encoder according to the rotating state of the rotating body, the rotational speed of the rotating body is determined. In a wide range of speeds, including high speeds, it is possible to detect changes in the level value of the pulse signal that matches the actual rotating state of the rotating body, and thus, it is highly reliable to know the rotating state of the rotating body. An object of the present invention is to provide a signal processing method that can be performed by the above method.

In order to achieve this object, the rotary encoder signal processing method of the present invention is set in advance after the switching of the level values of the pulse signal output by the rotary encoder according to the rotation of the rotating body. The switching of the level value is performed by excluding the switching accompanied by the switching to restore the level value within the predetermined chattering exclusion time period as the switching to be excluded due to chattering phenomenon or instantaneous noise. A signal processing method for recognizing a waveform of the pulse signal used for grasping a rotation state of the rotating body while executing a switching detection process for detecting occurrence of switching of a detection target,
A switching detection step of performing the switching detection process using the chattering exclusion time while observing the pulse signal output from the rotary encoder;
A time measuring step of measuring an elapsed time from detection of the occurrence of the detection target switching to detection of the next detection target switching;
A chattering exclusion time setting step for variably setting the chattering exclusion time used for the subsequent switching detection processing every time the occurrence of the detection target switching is detected;
Each time the occurrence of switching of the detection target is detected, the set value of the chattering exclusion time used in the switching detection process when the occurrence of the current detection target switching is detected, and the previous detection target switching Depending on the deviation from the set value of the chattering exclusion time used in the switching detection process when an occurrence is detected, the occurrence of the current detection target switching is detected after the detection of the previous detection target switching is detected. A pulse width estimating step for estimating the pulse width of the pulse signal immediately before the occurrence of the current detection target switching by correcting the elapsed time measured in the time measuring step until detection,
In the chattering exclusion time setting step, each time the occurrence of switching of the detection target is detected, it is estimated in the pulse width estimation step so as to shorten the chattering exclusion time as the rotational speed of the rotating body increases. The chattering exclusion time is variably set according to at least the latest value of the estimated pulse widths (first invention).

  According to the first aspect of the present invention, the chattering exclusion time is variably set by the chattering exclusion time setting step. In the switching detection step, the set chattering elimination time is used, and among the switching of the level value of the pulse signal output by the rotary encoder according to the rotation of the rotating body, the switching detection step sets in advance after the switching. The switching of the level value is performed by excluding the switching accompanied by the switching for restoring the level value within the predetermined chattering exclusion time period as the switching of the exclusion target due to chattering phenomenon or instantaneous noise. The switching detection process for detecting the occurrence of switching of the detection target is executed.

  Note that the detection target switching detected by this switching detection process is the original pulse signal that should be output from the rotary encoder when there is no chattering phenomenon or noise (level value only at the predetermined rotation angle position of the rotating body). This means that the level value of the pulse signal is switched.

  In addition, as described above, the “chattering phenomenon” in the present invention includes a bounce phenomenon in addition to the chattering phenomenon in the original meaning.

  Here, according to the knowledge based on various experiments and examinations of the inventors of the present application, the chattering phenomenon and instantaneous noise of the pulse signal are generally determined by the actual rotation angle position of the rotating body being the level value of the original pulse signal. Therefore, the time width during which the chattering phenomenon and the like can occur becomes shorter as the rotational speed of the rotating body becomes higher.

  Therefore, in the first invention, in the chattering exclusion time setting step, the basic guideline for setting the chattering exclusion time is to set the chattering exclusion time to be shorter as the rotational speed of the rotating body is faster. And

  In this case, the higher the rotational speed of the rotating body, the shorter the pulse width of the original pulse signal (the time width between the detection target switching of the level value of the pulse signal and the next detection target switching).

  The pulse width can be estimated by the process of the pulse width estimation step every time the occurrence of switching of the detection target is detected. That is, the set value of the chattering exclusion time used in the switching detection process when the occurrence of the current detection target switching is detected, and the switching detection process when the previous detection target switching is detected. According to the deviation from the set value of the chattering exclusion time used in the above, the time measured in the timing step from the detection of the previous detection target switching to the detection of the current detection target switching By correcting the elapsed time, it is possible to estimate the pulse width of the pulse signal immediately before the occurrence of the current detection target switching.

  In this case, more specifically, the setting value of the chattering exclusion time used in the switching detection process when the occurrence of the current detection target switching is detected is Tc1, and the previous occurrence of the switching of the detection target is detected. When the setting value of the chattering exclusion time used in the switching detection process at this time is Tc0 and the deviation is (Tc0-Tc1), this deviation (Tc0-Tc1) is used as the previous detection target switching. Addition to the elapsed time measured in the timing step from detection of occurrence to detection of occurrence of the current detection target switching (in other words, subtract (Tc1-Tc0) from the elapsed time) Thus, the pulse width can be estimated.

  It should be noted that the correction of the elapsed time measured in the timing step as described above is that the detection of the occurrence of switching of the detection target this time and the previous time is different from the switching of the actual level value of the pulse signal, respectively. This is because the chattering exclusion time is set at a timing delayed by the set values Tc0 and Tc1.

  Therefore, in the first invention, the chattering exclusion time is variably set according to at least the latest value of the estimated pulse width values estimated in the pulse width estimating step. In this case, the latest value of the estimated values of the pulse width estimated in the pulse width estimating step is the estimated value of the pulse width of the pulse signal immediately before the occurrence of the current detection target switching, so the rotation It reflects the current rotation speed of the body.

  Therefore, the chattering exclusion time is variably set according to at least the latest value of the estimated pulse widths estimated in the pulse width estimation step, so that the higher the rotational speed of the rotating body, the higher the chattering speed. Setting the chattering exclusion time so as to shorten the exclusion time can be appropriately realized.

  Note that the estimated value of the pulse width to be reflected in the setting of the chattering exclusion time may be only the latest value. However, as will be described later, according to the latest value and one or more estimated values before that (in other words, For example, the chattering exclusion time may be set by reflecting the history of the estimated value of the pulse width.

  In the first invention, since the switching detection process is performed using the chattering exclusion time set as described above, the rotating body rotates at a relatively high speed and the pulse width (detection of the original pulse signal) is detected. Even if the time width between the target switching and the next detection target switching) is relatively short, the level value of the pulse signal can be switched due to chattering or instantaneous noise. Thus, it is possible to detect the switching of the original pulse signal level value as the detection target switching while preventing the detection target switching.

  In other words, it is possible to increase the upper limit of the rotational speed of the rotating body that can be normally detected by switching the level value of the original pulse signal as the detection target switching.

  In addition, even when the rotating body rotates at a relatively low speed, the switching of the level value of the pulse signal due to chattering phenomenon or instantaneous noise is erroneously detected as the switching of the detection target. It is possible to detect the switching of the original level value of the pulse signal as the detection target switching.

  Therefore, according to the first aspect of the present invention, the level value switching of the pulse signal matched with the actual rotation state of the rotating body (switching of the detection target) is performed in a wide speed range including the case where the rotating speed of the rotating body is high. Can be detected, and as a result, the rotation state of the rotating body can be grasped with high reliability.

  The switching detection process executed in the switching detection step includes switching of level values in one or a plurality of pulses having a short pulse width that are included in the chattering exclusion time. Any method may be used as long as it is a method that can detect the switching of the detection target, but as a typical method, for example, the following method can be adopted. .

  In the one method, the switching detection step includes a first step of detecting switching of an actual level value of a pulse signal output from the rotary encoder, and a period of the chattering exclusion time has elapsed after the detection. A second step of observing the actual level value of the pulse signal immediately after, and the level value observed in the second step is the actual level value of the pulse signal immediately before the switching detection in the first step. It is determined whether or not the level value is different from the above, and when the determination result is affirmative, the occurrence of switching of the detection target is detected (second invention).

  Alternatively, in another method, in the switching detection step, the level value corresponding to the chattering exclusion time period is obtained by sequentially observing an actual level value of the pulse signal output from the rotary encoder at a predetermined period. The process of determining whether or not a plurality of observed values are the same as each other is sequentially executed, and when the determination result changes from a negative situation to a positive situation, the detection target switching is performed. Is detected (third invention).

  According to the second invention or the third invention, the switching detection process can be appropriately executed by a simple method while preventing the exclusion target switching from being detected as the detection target switching.

  In the above first to third inventions, the following aspects can be adopted as an example of a more specific aspect of the chattering exclusion time setting step. That is, in the chattering exclusion time setting step, for example, as the rotational speed of the rotating body indicated by the latest value of the estimated pulse widths estimated in the pulse width estimation step increases, the chattering exclusion time decreases. Thus, the chattering exclusion time is set according to the latest value (fourth invention).

  According to the fourth aspect of the invention, since the chattering exclusion time is set according to the latest value of the estimated pulse width values estimated in the pulse width estimation step, the chattering exclusion that matches the current rotational speed of the rotating body You can set the time. Further, since only the latest value of the estimated pulse width is used, the chattering exclusion time setting process can be easily performed.

  In the first to third aspects of the invention, as a more preferable example of the chattering exclusion time setting step, it is preferable to employ the following aspects. That is, in the chattering exclusion time setting step, the current detection target switching is changed based on the latest value of the estimated pulse width values estimated in the pulse width estimation step and at least one past value before that. Predicting the future rotational speed of the rotating body after the detection of the occurrence of, or the pulse width of the pulse signal corresponding to the rotational speed, or the parameter value having a certain correlation with the rotational speed or pulse width, The chattering exclusion time is set according to the prediction value so as to shorten the chattering exclusion time as the rotation speed of the rotating body indicated by the prediction value increases (fifth invention).

  As the parameter value, typically, a parameter value proportional to the rotational speed or the pulse width can be used.

  According to the fifth aspect of the present invention, based on the latest value of the pulse width estimation values estimated in the pulse width estimation step and at least one past value before that, the predicted future rotational speed of the rotating body is determined. Or the chattering exclusion time is set in accordance with a pulse width of the pulse signal corresponding to the rotational speed or a predicted value of a parameter value having a certain correlation with the rotational speed or pulse width. The chattering elimination time used in the future switching detection process after the detection of the occurrence of the target switching can be set to match the future rotational speed of the rotating body.

  For this reason, even when the rotational speed of the rotating body increases rapidly, the actual rotational speed of the rotating body in the vicinity of the generation time of the detection target switching of the level value of the pulse signal has already been set. It can be prevented that the chattering exclusion time becomes too long and the detection target switching is detected and thus excluded.

  Therefore, according to the fifth aspect of the present invention, it is possible to detect switching of the level value of the pulse signal (detection target switching) that matches the actual rotation state of the rotating body in a wider speed range of the rotating speed of the rotating body. it can. As a result, the rotation state of the rotating body can be grasped with higher reliability.

  In the first to fifth inventions described above, is the rotational speed of the rotating body indicated by the latest value of the estimated values of the pulse width estimated in the pulse width estimating step higher than a predetermined value? A step of determining whether or not the rotation is based on a waveform of a pulse signal recognized by reflecting the detection target switching detected in the switching detection step when the determination result is affirmative. You may make it prohibit the process which grasps | ascertains the rotation state of a body (6th invention).

  That is, when the rotational speed of the rotating body becomes very high, the pulse width between the switching of the original detection target of the pulse signal and the next switching of the detection target becomes very small. It becomes difficult to distinguish from pulses caused by noise.

  Therefore, in the sixth aspect of the invention, when the result of determination as to whether or not the rotational speed of the rotating body is higher than a predetermined value is affirmative, the detection object switching detected in the switching detection step is reflected. Thus, the process of grasping the rotation state of the rotating body based on the waveform of the pulse signal recognized is prohibited. Thereby, it is possible to prevent the rotating state of the rotating body from being erroneously grasped.

The figure which shows the circuit structure of the rotary encoder in embodiment of this invention. The figure which shows the outline of the mechanical structure of the rotary encoder regarding the switch shown in FIG. The flowchart which shows the arithmetic processing in 1st Embodiment of CPU shown in FIG. The figure for demonstrating the 1st example of the process of STEP1,2 of FIG. The figure for demonstrating the 2nd example of the process of STEP1,2 of FIG. The figure for demonstrating the process of STEP7 of FIG. The figure which shows the example of a setting of chattering exclusion time in STEP29 of FIG. The flowchart which shows the arithmetic processing in 1st Embodiment of CPU shown in FIG. The figure which shows the example of a setting of the chattering exclusion time in STEP29 of FIG.

[First Embodiment]
A first embodiment of the present invention will be described below with reference to FIGS.

  First, a schematic configuration of the rotary encoder 1 in the present embodiment will be described with reference to FIGS. 1 and 2. The rotary encoder 1 is a three-phase contact rotary encoder, and includes a switch mechanism unit 2 that constitutes switches 3a, 3b, and 3c for generating pulse signals for three phases of A phase, B phase, and C phase. Prepare.

  Each switch 3a, 3b, 3c of the switch mechanism 2 (hereinafter collectively referred to as the switch 3 when it is not necessary to distinguish between them) is rotated by the rotating body 4 by the mechanical structure schematically shown in FIG. This is a switch that is electrically connected and disconnected according to the condition.

  FIG. 2 representatively shows the mechanical structure of the switch 3 of one phase. In FIG. 2, for the convenience of illustration, the former is represented by a broken line and the latter is represented by a solid line in order to distinguish between the rotating body 4 and a portion that rotates integrally with the rotating body 4 and a fixed portion that does not rotate.

  As shown in FIG. 2, each switch 3 includes a slider 5 made of a conductive material such as a metal attached to the rotating body 4, and a rotating shaft 4 a of the rotating body 4 as a portion that makes the sliding contact with the slider 5. And an annular portion 7 formed on the fixed portion 6 on the same axis, and the slider 5 rotates integrally with the rotating body 4 with respect to the fixed portion 6 while being in sliding contact with the annular portion 7. ing.

  In the annular portion 7, a plurality of arc-shaped conductor portions 8 (portions made of a conductor material such as metal) having a predetermined arc length are arranged at regular intervals in the circumferential direction, and adjacent conductor portions 8, A portion 9 between 8 is an insulator portion (a portion made of an insulator material such as resin).

  For this reason, the slider 5 of each switch 3 is placed on the conductor portion 8 and the insulator portion 9 at a constant rotation angle period when the rotating body 4 rotates in any one of the clockwise direction and the counterclockwise direction. It comes in sliding contact with each other. As a result, when the rotating body 4 rotates in either the clockwise direction or the counterclockwise direction, the electrical conduction of each switch 3 (contact between the slider 5 and the conductor portion 8) and the electrical disconnection are achieved. (Contact between the slider 5 and the insulator portion 9) is performed alternately.

  At this time, the electrical continuity of each switch 3 is such that the rotational angle position of the slider 5 that rotates integrally with the rotating body 4 depends on the arrangement position of the conductor portion 8 in the circumferential direction of the annular portion 7 and This is performed when a predetermined angle range defined by the arc length of the conductor portion 8 is present. Similarly, the electrical disconnection of each switch 3 is performed when the rotational angle position of the slider 5 is set at the annular portion 7. This is performed when the angle is within a predetermined angular range defined by the arrangement position of the insulator portion 9 in the circumferential direction and the arc length of the insulator portion 9.

  In addition, the arc length of the conductor portion 8 of each switch 3 or its arrangement position, or the arrangement position of the slider 5 (arrangement position in the circumferential direction with respect to the rotating body 4) is between the switches 3a, 3b, 3c. It is different. For this reason, the range of the rotational angle position of the rotating body 4 where the electrical continuity and interruption of each switch 3 are performed or the width of the range are different among the switches 3a, 3b, 3c. .

  In FIG. 2, for convenience of illustration, the arc lengths of the conductor portion 8 and the insulator portion 9 of each switch 3 are described as being long, but in reality, these arc lengths are shorter.

  Further, when the rotation angle range of the rotating body 4 is limited to a predetermined range, the arrangement of the conductor portion 8 and the insulator portion 9 is provided only in a predetermined angle range region of the annular portion 7. May be formed. In this case, a ground electrode or the like may be formed in a region other than the region in the predetermined angular range of the annular portion 7.

  Returning to FIG. 1, one of the slider 5 and the conductor portion 8 of each switch 3 (one end of the switch 3) is grounded, and the other (the other end of the switch 3) is a DC power source (battery or the like) not shown. ) To a power supply voltage input unit 10 to which a DC voltage is applied, via a parallel circuit of a resistor 11 and a capacitor 12. Thus, when the rotating body 4 rotates in one direction with the DC voltage applied to the power supply voltage input unit 10 and the switches 3 are turned on and off alternately, the parallel circuit and the switches 3 The signal output unit 13 between the two outputs a pulse signal of high and low binary levels.

  In the illustrated example, the level value of the pulse signal output from the signal output unit 13 corresponding to each switch 3 is low level (ground level) when the switch 3 is in a conductive state (ON state), and the switch 3 When in the shut-off state (OFF state), the level is high. However, the level value of the pulse signal when the switch 3 is in the conductive state (ON state) is high, and the level value of the pulse signal when the switch 3 is in the cutoff state (OFF state) is low. It may be. Further, the circuit for generating the pulse signal is not limited to the circuit configuration shown in FIG. 1, and for example, the capacitor 12 can be omitted.

  The above is the schematic configuration of the rotary encoder 1 in the present embodiment.

  In this embodiment, the signal output unit 13 for each phase of the rotary encoder 1 is connected to a CPU 20 as an arithmetic processing unit that performs arithmetic processing for grasping the rotation state (rotational speed and direction) of the rotating body 4. The pulse signal output from the signal output unit 13 is input to the CPU 20.

  The CPU 20 executes arithmetic processing to which the present invention is applied in accordance with a program mounted in advance in a ROM (not shown).

  Hereinafter, the arithmetic processing of the CPU 20 will be described in detail.

  In the present embodiment, the CPU 20 sequentially executes the processing shown in the flowchart of FIG. 3 at a predetermined arithmetic processing cycle while observing (sampling) the level value of the pulse signal of each phase.

  The CPU 20 executes the switching detection process for detecting the occurrence of switching of the detection target in STEP 1 for each phase pulse signal input from the rotary encoder 1. Then, based on the result of the switching detection process, the CPU 20 determines the level value of the pulse signal (level value used for grasping the rotation state of the rotating body 4) in STEP2, and forms the waveform of the pulse signal. Recognize

  Here, the detection target switching is the sliding specified in accordance with the rotation angle position of the rotating body 4 among the switching of the level value of the actual pulse signal for each phase input from the rotary encoder 1 to the CPU 20. The contact position of the moving element 5 with respect to the annular portion 7 means that the chattering phenomenon (including the bounce phenomenon) and the exclusion target switching as switching due to instantaneous noise are excluded.

  In other words, this detection target switching is a change in the rotation angle position of the rotating body 4 (change in the contact position of the slider 5 with respect to the annular portion 7) when no chattering phenomenon or instantaneous noise occurs. Depending on the pulse signal to be output from the rotary encoder 1 (that is, a pulse whose level value is switched only at the moment when the contact position of the slider 5 shifts from one of the conductor portion 8 and the insulator portion 9 to the other). Signal) level value switching. The waveform of the pulse signal is a waveform of a pulse signal (a pulse signal obtained by removing the exclusion target switching from the actual pulse signal) in which the switching of the level value of the waveform is only the detection target switching. means.

  The switching detection process of STEP1 can be executed by various methods. In the present embodiment, two typical examples will be described below together with the process of STEP2.

  In the first example of the switching detection process of STEP 1, the CPU 20 detects an edge (rising edge or falling edge of the actual pulse signal waveform) as a step-like switching of the actual level value of the input pulse signal. In response to the edge detection, the following interrupt processing is executed to detect the occurrence of switching of the detection target.

  That is, after detecting the edge, the CPU 20 observes (samples) the actual level value of the pulse signal immediately after the currently set chattering exclusion time has elapsed. The currently set chattering exclusion time is set by the process described later when the CPU 20 detects the occurrence of switching of the detection target in the switching detection process already executed before the current switching detection process. It is the latest value of the recorded time. Basically, the chattering exclusion time is such that one or a plurality of continuous pulses (pulses having a short width) generated due to chattering phenomenon or instantaneous noise are within the time width of the chattering exclusion time. Is set.

  Then, in the switching detection process of STEP 1, the observed value of the level value of the pulse signal acquired immediately after the chattering exclusion time has elapsed is the same as the level value immediately before the edge detection (in other words, For example, when the level value is different from the level value immediately after the edge detection), the switching of the level value of the pulse signal generated within the period from the detection of the edge until the chattering exclusion time elapses is detected. It is assumed that the exclusion target switching due to chattering phenomenon or instantaneous noise is included, including the edge, and that the detection target switching has not occurred.

  In this case, the CPU 20 determines in STEP 2 that the level value of the pulse signal within the chattering elimination time period is held at the same level value as the level value immediately before the edge detection. Recognize the shaped waveform.

  For example, with reference to FIG. 4, a case is assumed where a pulse signal having a waveform as shown in the upper part of FIG. Then, the edge of this pulse signal is detected at time t10, and then the level value of the actual pulse signal is switched at time t11 within the period until chattering exclusion time elapses (the level value is the level immediately before time t10). Switching to return to the value) occurs, and the level value of the pulse signal observed at time t12 immediately after the chattering exclusion time has elapsed is the same as the level value immediately before time t10 at which the edge was detected. .

  In this case, in the switching detection process of STEP 1, the CPU 20 determines whether the level value switching (edge) of the pulse signal at time t 10 and the level value switching of the pulse signal at time t 11 are excluded. It is considered that there is no change in the detection target within the chattering exclusion time period from time t10 to t12 (including t10).

  In this case, in the processing of STEP 2, the CPU 20 maintains the level value of the pulse signal in the period from time t10 to time t12 immediately after the chattering exclusion time has passed, to the same level value as the level value immediately before time t10. As shown in the lower part of FIG. 4, the shaped waveform of the pulse signal is recognized.

  On the other hand, if the observed value of the level value of the pulse signal acquired immediately after the currently set chattering exclusion time has elapsed differs from the level value immediately before edge detection (in other words, the level immediately after edge detection) If the same value), the CPU 20 considers that the switching of the level value of the pulse signal is not switched to the exclusion target, that is, the detection target switching has occurred (detection of detection target switching). .

  For example, referring to FIG. 4, the edge value is detected at time t13, and then the level value of the pulse signal observed at time t14 immediately after the chattering exclusion time has elapsed is the level value immediately before time t13 when the edge is detected. Assume a different case. In this case, in the switching detection process of STEP1, the CPU 20 detects that the detection target switching has occurred at time t14.

  In this case, in step 2, the CPU 20 recognizes the shaped waveform of the pulse signal on the assumption that the level value is switched at time t14 as shown in the lower part of FIG.

  In the first example of the switching detection process, the above-described processing causes a switching with a switching for returning the level value within the chattering exclusion time period after the actual level value switching, that is, chattering phenomenon or Switching due to a pulse having a short width that occurs due to instantaneous noise is not detected as switching of the detection target, and the original switching depending on the rotational angle position of the rotating body 4 (slider) 5), which occurs only at the moment when the contact position 5 shifts from one side of the conductor portion 8 and the insulator portion 9 to the other side), can be detected as the detection target switching.

  Note that the switching detection process of STEP 1 in the first example is performed in a period from when an edge is detected until immediately after the chattering exclusion time has elapsed, and is omitted in other periods.

  Next, in the second example of the switching detection process, the CPU 20 sequentially observes (samples) the actual level value of the input pulse signal at a predetermined calculation processing cycle, and uses a polling format as described below. The occurrence of switching of the detection target is detected by the process.

  That is, in the switching detection process of STEP 1, the CPU 20 stores and holds a plurality of observed values of the level values of the pulse signal for the currently set chattering exclusion time period while sequentially updating each calculation processing cycle. Then, it is determined whether all of the plurality of observed values of the level value for the chattering exclusion time period are the same level value, and if this determination result is negative, the chattering exclusion time period It is assumed that there is no switching of the detection target.

  Further, in this case, the CPU 20 holds the level value of the pulse signal within the chattering exclusion time period at STEP 2 at the same level value as the level value immediately before the chattering exclusion time period. As such, the shaped waveform of the pulse signal is recognized.

  In addition, when the determination result changes from negative to positive, the CPU 20 regards that the detection target switching has occurred (detects the detection target switching).

  For example, assuming that the waveform of the actual pulse signal input to the CPU 20 changes as shown in the upper part of FIG. 5, the currently set chattering exclusion time is 2 of the sequential observation period (sampling period) of the CPU 20. It is assumed that the time is a period. In this case, the number of observed values of the pulse signal for the chattering exclusion time period is three. Note that times t20 to t25 in FIG. 5 indicate the observation time of the level value for each calculation processing cycle, and the time width between two adjacent times of those times t20 to t25 is the above calculation processing cycle. It is.

  In this case, in each period from time t20 to t22, from t21 to t23, and from t22 to t24 (period of chattering exclusion time), the three observed values of the level value of the pulse signal in the period are the same. Since it is not a level value, until the time t24, in the switching detection process of STEP1, the CPU 20 regards that the detection target switching has not occurred. In this case, the CPU 20 assumes that the level value of the pulse signal is maintained at the same level value as that immediately before time t20 in the processing of STEP2, and the pulse signal as shown in the lower part of FIG. Recognize the shaped waveform.

  Further, since the three observed values of the level value of the pulse signal in the period from time t23 to t25 become the same level value, the CPU 20 performs the switching detection process of STEP1 in the calculation processing cycle at time t25. Detects occurrence of switching of detection target. In this case, the CPU 20 recognizes the shaped waveform of the pulse signal as a pulse signal whose level value switches at time t25 as shown in the lower part of FIG.

  In the second example of the switching detection process, the switching described in the above-described process is performed to restore the level value within the chattering exclusion time period after the actual level value is switched, as in the case of the first example. That is, switching due to a pulse having a short pulse width that occurs due to chattering phenomenon or instantaneous noise is not detected as detection object switching, and the rotational angle position of the rotating body 4 It is possible to detect the original switching depending on the detection target switching.

  Returning to the description of the flowchart of FIG. 3, the CPU 20 determines whether or not the switching of the level value of the shaping waveform of the pulse signal has occurred in STEP 3 following the processing of STEP 2. This determination result is affirmative when the detection switching of the detection target is detected in the switching detection process of STPE1, and is negative when the occurrence of the switching of the detection target is not detected.

  When the determination result of STEP 3 is negative, the CPU 20 measures the elapsed time (time width) from when the determination result of STEP 3 becomes positive until it becomes positive next. Counting up the count value of the width counter by a predetermined step value (for example, “1”) is executed in STEP 10, and the processing of the current arithmetic processing cycle is terminated (waiting for the next arithmetic processing cycle).

  If the determination result in STEP 3 is affirmative, the CPU 20 determines that the pulse immediately before switching of the actual pulse signal corresponding to the detection target switching detected in STEP 1 based on the current count value of the width counter. After the processing for estimating the pulse width of the signal is executed in STEP 4, the count value of the width counter is initialized (reset) to “0” in STEP 5.

  Here, the estimation process of STEP4 will be described with reference to FIG. Assume that the actual pulse signal is a pulse signal in which level value switching occurs at times t30 and t32 as shown in the upper part of FIG. Here, for convenience of understanding, it is assumed that the actual pulse signal does not include the switching of the exclusion target, and the level value switching in the pulse signal is all the detection target switching. To do.

  It is assumed that chattering exclusion times set at times t30 and t32 are Tc0 and Tc1, respectively (Tc0 <Tc1 in the illustrated example).

  In this case, the waveform of the pulse signal having the pulse width recognized from the count value of the width counter is as shown in the lower part of FIG. That is, the actual pulse signal detection target switching at times t30 and t32 is detected at times t31 and t33, respectively, and the waveforms that cause the level value switching at times t31 and t33 are the count values of the width counter. It is the waveform of the pulse signal of the pulse width recognized from.

  In such a situation, the case where the estimation process of STEP4 is performed in the calculation process cycle at time t33 when the detection target switching is detected will be described as an example.

  In the calculation processing cycle at time t33, the pulse width to be estimated by the CPU 20 in the estimation processing in STEP 4 is the time t30 detected in the calculation processing cycle at time t31 in the actual level change of the pulse signal. Between the detection object switching (previous detection object switching) and the detection object switching (current detection object switching) at time t32 detected in the calculation processing cycle at time t33, that is, This is the pulse width Wd of FIG.

  Also, the count value used in the estimation process of STEP 4 is from the time t31 of the arithmetic processing cycle when the occurrence of the previous detection target switching is detected to the time t33 of the arithmetic processing cycle when the occurrence of the current detection target switching is detected. Therefore, the time width Wdc from time t31 to time t33 is indicated. The time width Wdc is obtained by multiplying the count value used in the estimation process of STEP 4 by the time of the arithmetic processing period when the increment of the count up of the width counter is “1”.

  Here, the detection target switching on the start end side of the pulse width Wd in the actual pulse signal is detected at the calculation processing period from the generation time t30 to the time t31 immediately after the chattering exclusion time Tc0 has elapsed, and the pulse width Wd The switching of the detection object on the end side of is detected at the calculation processing cycle (current calculation processing cycle) at time t33 immediately after the chattering exclusion time Tc1 elapses from the occurrence time t32.

  In this case, when the chattering exclusion times Tc0 and Tc1 are the same, the pulse width Wd matches the time width Wdc from time t31 to time t33. Therefore, in this case, the count value of the width counter in the current calculation processing cycle indicates the pulse width Wd.

  However, in the present embodiment, since the chattering exclusion time is variably set by the process described later every time the occurrence of switching of the detection target is detected, Tc0 and Tc1 are not necessarily the same. .

  When Tc0 and Tc1 are not the same as described above, the time width Wdc from time t31 to time t33 is equal to Tc0 and Tc1 as shown in FIG. 6 with respect to the pulse width Wd of the actual pulse signal. Deviation of the deviation is generated.

  Therefore, in the estimation process of STEP 4, the CPU 20 detects the chattering exclusion time Tc1 used in the switching detection process that has detected the occurrence of the current detection target switching, and the switching detection process that has detected the previous detection target switching. By correcting the time width Wdc indicated by the count value of the width counter in the current calculation processing period by the deviation from the chattering exclusion time Tc0 used in the above, the pulse width Wd in the actual pulse signal (the latest detection target switching The pulse width between the change and the previous detection target change) is estimated.

  More specifically, the CPU 20 estimates Wd as the pulse width of the actual pulse signal by correcting Wdc according to the deviation between Tc0 and Tc1 according to the following equation (1).

Wd = Wdc + (Tc0-Tc1) (1)

The above is the details of the estimation process of STEP4.

  In the present embodiment, the pulse width Wd is estimated as described above. However, as another form, the following may be used. That is, in STEP5 following STEP4, the initial value of the width counter is set to a count value corresponding to the chattering exclusion time used in the switching detection process in which occurrence of the current detection target switching is detected (the chattering exclusion time is calculated as a processing cycle). (Value divided by).

  In such a case, the count value of the width counter used in STEP 4 in the current calculation processing cycle indicates the time width of (Wdc + Tc0) in the equation (1). Therefore, the pulse width Wd of the actual pulse signal is estimated by subtracting the chattering exclusion time Tc1 used in the switching detection process for detecting the occurrence of the current detection target switching from the time width (Wdc + Tc0) indicated by the count value. A value may be obtained.

  After executing the processing of STEPs 4 and 5 as described above, the CPU 20 determines in STEP 6 whether or not a change in the rotational speed of the rotating body 4 has occurred.

  Specifically, the CPU 20 compares the rotation speed of the rotating body 4 indicated by the estimated value of the pulse width calculated this time with the rotation speed of the rotating body 4 indicated by the estimated value of the pulse width calculated last time. If the former rotational speed is different from the latter rotational speed, the CPU 20 determines that the determination result in STEP 6 is positive, and if both are the same, the determination result in STEP 6 is negative. To do. Whether or not the rotational speeds of the two are the same is determined by whether or not the absolute value of the difference between the rotational speeds of the two is smaller than a predetermined value (> 0) in the vicinity of “0” determined in advance. It may be.

  Here, even in a situation where the rotation speed of the rotating body 4 is kept constant, the pulse width of the period in which the level value of the pulse signal of each phase is high, and the pulse width of the period in which the level is low Is generally different. Therefore, in the determination process of STEP 6, the CPU 20 determines the current pulse width estimated value (or a count value obtained by dividing the pulse width by the calculation processing period) and the previous pulse width estimated value (or the calculated value in the calculation processing period). The calculation value or map (pulse width and rotation of the rotating body 4) set in advance depending on whether the level value corresponding to the pulse width is a low level or a high level. An arithmetic expression or map representing a relationship (inverse proportional relationship) with the speed is converted into the rotation speed of the rotating body 4 and the converted rotation speeds are compared.

  However, in a situation where the rotational speed of the rotating body 4 is kept constant, the pulse width during the period when the level value of the pulse signal of any phase is high and the pulse width during the period when it is low are the same. When the rotary encoder 1 is configured so that the arc length of the conductor portion 8 of the annular portion 7 and the arc length of the insulator portion 9 are set equal to each other, In the judgment process of STEP 6 for the pulse signal of this time, the estimated value of the current pulse width (or the count value obtained by dividing this by the calculation processing period) and the estimated value of the previous pulse width (or this is calculated by the calculation processing period). A count value obtained by division may be compared.

  In a situation where the rotational speed of the rotating body 4 is kept constant, one pulse width corresponds to the other pulse width according to the ratio of the pulse width during the low level period to the pulse width during the high level period. It is also possible to compare the converted value of one pulse width with the other pulse width.

  When the determination result in STEP 6 is affirmative, the CPU 20 executes the process of STEP 8 after executing the process of updating the setting value of the chattering exclusion time in STEP 7. If the determination result in STEP 6 is negative, the CPU 20 executes the process in STEP 8 without updating the set value for the chattering exclusion time (while maintaining the current value). In this embodiment, the processing of STEP 7 (update processing of the setting value for chattering exclusion time) is performed only when the determination result of STEP 6 is affirmative, but the determination processing of STEP 6 is omitted (the rotation of the rotating body 4). The processing of STEP 7 may be executed regardless of whether or not a speed change has occurred.

  The process of updating the set value of chattering exclusion time in STEP 7 is performed as follows.

  That is, as a basic guideline, the CPU 20 sets the chattering exclusion time so that the chattering exclusion time is shortened as the rotational speed of the rotating body 4 indicated by the current pulse width estimated value (latest estimated value) increases. To do.

  More specifically, the CPU 20 calculates the estimated value of the current pulse width (or a count value obtained by dividing the estimated value by the calculation processing cycle) according to the predetermined calculation formula or map as described with respect to STEP 6. Convert to the rotation speed of. Then, the CPU 20 calculates a new chattering exclusion time from the converted rotation speed using a predetermined arithmetic expression or map (an arithmetic expression or map representing a relationship between the rotation speed and the chattering exclusion time) set in advance. decide.

  In this case, the arithmetic expression or map used to determine the chattering exclusion time is set so that the chattering exclusion time becomes shorter in inverse proportion to the increase in the rotational speed.

  Alternatively, the chattering exclusion time may be set as follows. That is, the CPU 20 calculates a predetermined arithmetic expression or map (the pulse width or a count value corresponding to the pulse width or the corresponding value) from the estimated value of the current pulse width (or a count value obtained by dividing the estimated value by the arithmetic processing cycle). A new chattering exclusion time is determined using an arithmetic expression or map representing a relationship between the chattering exclusion time.

  In this case, the above calculation formula or map shows the pulse width during the period when the level value of the pulse signal is high and the period when the level is low, in a situation where the rotation speed of the rotating body 4 is kept constant. When the pulse width is different, the pulse width is set separately depending on whether the pulse width is a low level pulse width or a high level pulse width. Alternatively, for example, an estimated pulse width is set by setting only an arithmetic expression or a map corresponding to either the case where the pulse width is a low level pulse width or the case where the pulse width is a high level pulse width. Is the pulse width of the other level value, the estimated value of the pulse width is converted into a pulse width equivalent to the pulse width of one level value, and new chattering is eliminated using the above equation or map. It is also possible to determine the time.

  As described above, by determining the chattering elimination time, the chattering elimination time is shortened in inverse proportion to the higher rotational speed of the rotating body 4 indicated by the current pulse width estimation value (latest estimation value). Thus, the chattering exclusion time is set.

  For example, it is assumed that an actual pulse signal having a waveform as shown in the upper part of FIG. Here, for convenience of understanding, it is assumed that the actual pulse signal does not include the exclusion target switching.

  In the illustrated example, it is assumed that the rotational speed of the rotating body 4 increases during the period from time t42 to time t46. Further, in this example, for convenience of understanding, it is assumed that the pulse width when the rotational speed of the rotating body 4 is constant is the same when the level value is high and when it is low.

  In this case, the pulse width Wd0 of the actual pulse signal in the period from time t40 to t42, the pulse width Wd1 (<Wd0) of the actual pulse signal in the period from time t42 to t44, and from time t44 to t46 The pulse width Wd2 (<Wd1) of the actual pulse signal in the period of time is estimated in the calculation processing period at time t43, the calculation processing period at time t45, and the calculation processing period at time t47, respectively. At this time, the rotational speed of the rotating body 4 indicated by the estimated values of the respective pulse widths Wd0, Wd1, and Wd2 is ω0, ω1, and ω2 (ω0 <ω1 <ω2) as shown in the lower part of FIG.

  Then, the chattering exclusion time is updated in each of the calculation processing period at time t43, the calculation processing period at time t45, and the calculation processing period at time t47 in which the pulse widths Wd0, Wd1, and Wd2 are estimated.

  In this case, the chattering exclusion time after the update in the calculation processing cycle at time t43 is Tc2 of the chattering exclusion times Tc0 to Tc4 shown in FIG. 7 (chattering exclusion used for detection of switching of the detection target at time t44). Time), the chattering elimination time after update in the computation processing cycle at time t45 is Tc3 in FIG. 7 (chattering elimination time used for detection of switching of the detection target at time t46), and the computation processing cycle at time t47. The chattering elimination time after updating is Tc4 in FIG. 7 (the chattering elimination time used for detecting the detection target switching at time t48).

  In this case, the chattering exclusion times Tc2, Tc3, and Tc4 are set to values corresponding to the rotational speeds ω0, ω1, and ω2, respectively. The faster the rotational speeds ω0, ω1, and ω2, the chattering exclusion times Tc2, Tc3 The chattering elimination times Tc2, Tc3, and Tc4 are determined so that Tc4 becomes shorter (Tc2> Tc3> Tc4).

  Returning to the description of FIG. 3, next, in STEP 8, the CPU 20 determines whether or not the rotation speed of the rotating body 4 is high-speed rotation. In this case, the CPU 20 determines that the estimated value of the current pulse width (or a count value obtained by dividing the estimated value by the calculation processing period) is a predetermined value set in advance for each of the low level pulse width and the high level pulse width. When the rotational speed of the rotating body 4 is smaller than the lower limit value or when the rotational speed of the rotating body 4 converted from the estimated value of the pulse width is larger than a predetermined upper limit value set in advance, the rotational speed of the rotating body 4 is high-speed rotation. Judge that there is.

  Here, when the rotational speed of the rotating body 4 is very high, the pulse width of the time interval of occurrence of switching of the detection target becomes very small. Therefore, the pulse width and chattering phenomenon or instantaneous It is difficult to estimate the pulse width of the time interval of occurrence of switching of the detection target while distinguishing the pulse due to noise with high reliability.

  For this reason, in this embodiment, when the determination result of STEP 8 is affirmative, the final grasp of the rotation state (rotation speed and rotation direction) of the rotating body 4 is omitted, and the processing of the current calculation processing cycle is performed. Exit. In this case, the final grasp of the rotation state of the rotating body 4 based on the shaped waveform of the pulse signal recognized in STEP 2 in the current arithmetic processing cycle is prohibited. The current rotation state of the rotator 4 is considered to be maintained, for example, in the state of the previous calculation processing cycle.

  On the other hand, if the determination result in STEP 8 is negative, the CPU 20 finally grasps (specifies) the rotation state such as the rotation speed and the rotation direction of the rotating body 4 in STEP 9 and determines the current calculation processing cycle. End the process. In this case, the rotation direction is specified based on the order in which the detection target switching of the pulse signal of each phase occurs between the phases.

  The above is the processing executed by the CPU 20 in the present embodiment.

  According to the embodiment described above, the chattering exclusion time is set while being updated as needed to shorten the chattering exclusion time as the rotational speed of the rotating body 4 increases. The occurrence of a change (detection target switching) can be detected with high reliability. As a result, the rotation state of the rotating body 4 can be grasped with high reliability.

  Here, the correspondence between the present embodiment and the present invention will be supplemented. In the present embodiment, the switching detection step, the pulse width estimation step, the chattering exclusion time setting step, and the timing step in the present invention are realized by the processing of STEP1, STEP4, STEP7, and STEP10, respectively. Moreover, the step of determining whether or not the rotational speed of the rotating body is high is realized by the processing of STEP8.

[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference to FIGS. Note that this embodiment is different from the first embodiment only in a part of the arithmetic processing of the CPU 20. For this reason, in the description of the present embodiment, the same reference numerals as those in the first embodiment are used for the same configurations as those in the first embodiment, and detailed description thereof is omitted.

  In the first embodiment, when the rotational speed of the rotating body 4 rapidly increases, the chattering exclusion time set at the time when the detection target switching of the level value of the pulse signal occurs is the actual rotation around the time. It takes a longer time than the rotational speed of the body 4, and there is a possibility that the detection target switching is excluded from the detection as the exclusion target switching.

  For example, when the rotational speed of the rotating body 4 after time t42 in FIG. 7 suddenly increases, the pulse width (Wd1, Wd2, etc.) of the pulse signal after time t42 is the generation time of detection object switching on the start end side of the pulse width. In some cases, the chattering time (Tc1, Tc2, etc.) already set becomes shorter.

  Therefore, in the present embodiment, every time the occurrence of switching of the detection target of the level value of the pulse signal is detected, an estimated value of a plurality of nearest pulse widths among the pulse widths already estimated in the past, for example, the latest pulse Using the two estimated values of the estimated width and the previous estimated pulse width, the rotational speed of the rotating body 4 in the future (near the time when the next detection target switching occurs) is calculated. Predict and update / set chattering exclusion time according to the predicted value.

  In order to update / set the chattering exclusion time in this way, in this embodiment, the CPU 20 performs the arithmetic processing shown in the flowchart of FIG. 8 instead of the arithmetic processing shown in the flowchart of FIG. 3 described in the first embodiment. Are sequentially executed in a predetermined calculation processing cycle.

  Explained below, the CPU 20 executes the same processing (switching detection processing etc.) as the processing of STEP 1 to STEP 3 of FIG. If the determination result in STEP 23 is negative, the CPU 20 executes the same processing as STEP 10 in STEP 3 in STEP 32 and ends the arithmetic processing in the current arithmetic processing cycle.

  On the other hand, in the present embodiment, when the determination result in STEP 23 is affirmative, the CPU 20 stores and holds (backs up) the previous estimated value of the pulse width in the backup memory (not shown) in STEP 24. To do. Here, the “previously estimated value of the pulse width” is a calculation process at the time when the occurrence of the previous detection target switching (the detection target switching immediately before the current detection target switching) is detected. It is the pulse width estimated by the processing of STEP25 in the cycle.

  Next, in the present embodiment, the CPU 20 executes the same process as STEP 4 in FIG. 3 in STEP 25, and estimates the pulse width of the pulse signal immediately before the detection target switching detected in the current arithmetic processing cycle. Further, the CPU 20 executes the same processing as STEP 5 of FIG. 3 in STEP 26 to reset the count value of the width counter, and then executes the processing of STEP 27 to predict the future rotational speed of the rotating body 4.

  This prediction is performed as follows. In other words, the CPU 20 calculates the previous estimated value of the pulse width backed up in STEP 24 (or a count value obtained by dividing the pulse width by the calculation processing period) and the current estimated value of the pulse width estimated in STEP 25 (or calculates this). The count value obtained by dividing by the period) is converted into the rotation speed of the rotating body 4, respectively. This conversion is performed using a predetermined arithmetic expression or map, as in the case described with respect to the processing of STEP 6 in the first embodiment.

  Then, the CPU 20 converts the rotational speed converted from the previous estimated value of the pulse width (hereinafter referred to as ωa) and the rotational speed converted from the estimated value of the current pulse width (hereinafter referred to as ωb). Therefore, the future rotational speed of the rotator 4 near the time when the next detection target switching is detected is predicted by a linear interpolation method (hereinafter, this will be referred to as ωc).

  Specifically, a predicted value of the future rotational speed ωc is calculated by the following equation (2).

ωc = ωb + (ωb−ωa) × α (2)

Α in the equation (2) is a coefficient set in advance as a ratio that takes into account the acceleration of the rotating body 4.

  The predicted value of the rotational speed ωc calculated in this way is faster than the rotational speed ωb indicated by the estimated value of the current pulse width in a situation where the rotational speed of the rotating body 4 is increasing (ωb> ωa). In a situation where the rotational speed of the rotating body 4 is decelerating (ωb <ωa), the rotational speed is slower than the rotational speed ωb indicated by the estimated value of the current pulse width. In a situation where the rotational speed of the rotating body 4 is kept constant (ωb = ωa), the predicted value of the rotational speed ωc is the constant rotational speed (= ωb = ωa).

  After predicting the future rotational speed ωc as described above, the CPU 20 next determines in STEP 28 whether or not the rotational speed of the rotating body 4 predicted after the current time changes.

  Specifically, the CPU 20 compares the predicted value of the future rotational speed ωc calculated in STEP 27 with the rotational speed ωb of the rotating body 4 converted from the estimated value of the current pulse width. If the former rotational speed is different from the latter rotational speed, the CPU 20 determines that the determination result of STEP 28 is positive, and if both are the same, the determination result of STEP 28 is negative. To do. Whether or not the rotational speeds of the two are the same is determined by whether or not the absolute value of the difference between the rotational speeds of the two is smaller than a predetermined value (> 0) in the vicinity of “0” determined in advance. It may be.

  Supplementally, in a situation where the rotational speed of the rotating body 4 is kept constant, the pulse width of the period in which the level value of the pulse signal of any phase is high and the pulse width of the period in which the level is low are When the rotary encoder 1 is configured to be the same, the right side ωb and ωa of the above equation (2) are set to the current pulse width in the processing of STEP 27 for the pulse signal of the phase. Formula replaced with estimated value and estimated value of previous pulse width (or replaced with count value obtained by dividing each of estimated value of current pulse width and estimated value of previous pulse width by calculation processing cycle ) May be used to predict a future pulse width (or a count value obtained by dividing the pulse width by an arithmetic processing cycle).

  Further, the determination process in STEP 28 is performed in the same manner as described in connection with the process in STEP 6 in FIG. 3, and the pulse width corresponding to the predicted value of the future rotation speed (or the count value obtained by dividing this by the calculation processing period) Alternatively, this may be performed by comparison with an estimated value of the current pulse width (or a count value obtained by dividing this by the calculation processing period).

  If the determination result in STEP 28 is affirmative, the CPU 20 executes the process of STEP 30 after executing the process of updating the chattering exclusion time setting value in STEP 29. On the other hand, if the determination result in STEP 28 is negative, the CPU 20 executes the process in STEP 30 without updating the set value for the chattering exclusion time (while maintaining the current value). In this embodiment, the process of STEP 29 (the process of updating the set value for chattering exclusion time) is performed only when the determination result of STEP 28 is positive, but the determination process of STEP 28 is omitted (the rotation of the rotating body 4). The process of STEP 29 may be executed regardless of whether or not a change in speed occurs.

  The processing for updating the set value of chattering exclusion time in STEP 29 is performed as follows.

  That is, the CPU 20 sets the chattering exclusion time so that the chattering exclusion time becomes shorter as the predicted value of the future rotational speed ωc of the rotating body 4 obtained in STEP 27 is faster.

  More specifically, the CPU 20 uses a predetermined arithmetic expression or map (an arithmetic expression or map representing a relationship between the rotational speed and the chattering exclusion time) set in advance from the predicted value of the future rotational speed ωc. To determine a new chattering exclusion time. The arithmetic expression or map used in this case may be the same as that used in STEP 7 of FIG.

  Note that in STEP 27, a predicted value of a pulse width (low level or high level pulse width) corresponding to a predicted value of the future rotational speed ωc in the same manner as described in regard to the processing of STEP 7 in FIG. A predetermined arithmetic expression or map (an arithmetic expression or map representing the relationship between the pulse width or count value and the chattering exclusion time) set in advance is used from the predicted value of the count value obtained by dividing this by the arithmetic processing cycle. Thus, the chattering exclusion time may be determined.

  As described above, by determining the chattering exclusion time, the rotational speed ωb indicated by the estimated value of the current pulse width (the latest estimated value of the pulse width of the pulse signal) and the rotational speed indicated by the estimated value of the previous pulse width As the predicted value of the future rotational speed ωc predicted from ωa is faster, the chattering exclusion time is set so as to shorten the chattering exclusion time in inverse proportion to the predicted value.

  For example, it is assumed that an actual pulse signal having a waveform as shown in the upper part of FIG. Here, for convenience of understanding, it is assumed that the actual pulse signal does not include the exclusion target switching.

  In the illustrated example, it is assumed that the rotational speed of the rotating body 4 has increased from time t52. Further, in this example, for convenience of understanding, it is assumed that the pulse width when the rotational speed of the rotating body 4 is constant is the same when the level value is high and when it is low.

  In this case, the pulse width Wd0 of the actual pulse signal in the period from time t50 to t52, the pulse width Wd1 (<Wd0) of the actual pulse signal in the period from time t52 to t54, and from time t54 to t56 The pulse width Wd2 (<Wd1) of the actual pulse signal in the period is estimated at the calculation processing period at time t53, the calculation processing period at time t55, and the calculation processing period at time t57, respectively. At this time, the rotational speed of the rotating body 4 indicated by the estimated values of the respective pulse widths Wd0, Wd1, and Wd2 is ω0, ω1, and ω2 (ω0 <ω1 <ω2) as shown in the lower part of FIG. In the example of FIG. 9, the pulse width of the actual pulse signal before time t50 is assumed to be the same as the pulse width Wd0.

  Then, predicted values of the future rotational speed of the rotating body 4 are calculated in the calculation processing cycle at time t53, the calculation processing cycle at time t55, and the calculation processing cycle at time t57 at which the pulse widths Wd0, Wd1, and Wd2 are estimated. The

  In this case, in the calculation processing cycle at time t53, the rotation speed ω0 converted from the estimated value of the pulse width Wd0 calculated in this calculation processing cycle and the rotation speed converted from the estimated value of the previous pulse width (= Wd0). A predicted value ωc0 of the future rotational speed ωc shown in the middle part of FIG. 9 is calculated by an arithmetic expression formed by substituting (= ω0) into ωb and ωa on the right side of the equation (2). This predicted value ωc0 matches ω0 in the example of FIG. Then, Tc2 of the chattering exclusion times Tc0 to Tc4 shown in FIG. 9 is set according to the predicted value ωc0.

  In the calculation processing cycle at time t55, the rotation speed ω1 converted from the estimated value of the pulse width Wd1 calculated in the calculation processing cycle and the rotation speed ω1 converted from the estimated value of the previous pulse width Wd0 are respectively set. The predicted value ωc1 of the future rotational speed ωc is calculated by an arithmetic expression formed by substituting ωb and ωa on the right side of the equation (2). The predicted value ωc1 is ωc1> ω1 in the example of FIG. Then, the chattering exclusion time Tc3 shown in FIG. 9 is set according to the predicted value ωc1.

  Further, in the calculation processing cycle at time t57, the rotation speed ω2 converted from the estimated value of the pulse width Wd2 calculated in the calculation processing cycle and the rotation speed ω2 converted from the estimated value of the previous pulse width Wd1 are respectively set. The predicted value ωc2 of the future rotational speed ωc is calculated by an arithmetic expression formed by substituting ωb and ωa on the right side of the equation (2). The predicted value ωc2 is ωc2> ω2 in the example of FIG. Then, the chattering exclusion time Tc4 shown in FIG. 9 is set according to the predicted value ωc2.

  In this case, the chattering exclusion times Tc2, Tc3, and Tc4 set as described above are such that the chattering exclusion times Tc2, Tc3, and Tc4 are shorter as the predicted rotational speed values ωc0, ωc1, and ωc2 are faster. (Tc2> Tc3> Tc4).

  Returning to the description of FIG. 8, next, in STEP 30, the CPU 20 executes the same determination process as in STEP 8 of FIG. 3, and determines whether or not the rotational speed of the rotating body 4 is high speed rotation. If the determination result in STEP 30 is affirmative, the final grasp of the rotation state (rotation speed and rotation direction) of the rotating body 4 is omitted, and the processing of the current calculation processing cycle is terminated.

  On the other hand, if the determination result in STEP 30 is negative, the CPU 20 executes the same processing as STEP 9 in FIG. 3 in STEP 31 and ends the processing in the current arithmetic processing cycle.

  The above is the processing executed by the CPU 20 in the present embodiment.

  According to the embodiment described above, every time the detection target switching of the level value of the actual pulse signal is detected, the future rotational speed of the rotating body 4 is predicted, and the faster the predicted value, the more the chattering exclusion time. Since the chattering exclusion time is set while being updated as needed so as to shorten it, even if the rotational speed of the rotating body 4 increases rapidly, the original switching of the level value of the pulse signal (detection target) It is possible to prevent the switching) from being detected as the exclusion target switching. For this reason, the occurrence of the original switching of the level value of the pulse signal can be detected with higher reliability. As a result, the rotation state of the rotating body 4 can be grasped with higher reliability.

  Here, the correspondence between the present embodiment and the present invention will be supplemented. In the present embodiment, the switching detection step, the pulse width estimation step, the chattering exclusion time setting step, and the time measurement step in the present invention are realized by the processing of STEP 21, STEP 25, STEP 29, and STEP 32, respectively. Further, the step 30 determines whether or not the rotational speed of the rotating body is high. The count value corresponding to the pulse width corresponds to the parameter value in the present invention.

  In the present embodiment, the rotational speed ωb of the rotating body 4 indicated by the latest value of the estimated pulse width values, and the rotational speed ωa of the rotating body 4 indicated by the estimated value of the previous pulse width, The future rotation speed ωc is predicted from the above, but the future rotation speed ωc (or the corresponding pulse width or count value) is predicted based on the estimated values of three or more pulse widths on the nearest side. You may make it do. In this case, the future rotation speed ωc (or the corresponding pulse width or count value) may be predicted with a higher-order function pattern than a linear function (linear function) such as a quadratic function. Good.

  In each of the above embodiments, the case where the rotary encoder 1 is a rotary encoder that generates a three-phase pulse signal has been described as an example. However, the rotary encoder according to the present invention generates, for example, a one-phase or two-phase pulse signal. It may be a rotary encoder.

  In each of the above embodiments, the case where the rotary encoder 1 is a contact type rotary encoder has been described as an example. However, the present invention can be applied to an optical rotary encoder.

  DESCRIPTION OF SYMBOLS 1 ... Rotary encoder, STEP1, 21 ... Switching detection step, STEP4, 25 ... Pulse width estimation step, STEP7, 29 ... Chattering exclusion time setting step, STEP10, 32 ... Timing step.

Claims (6)

Of the switching of the level value of the pulse signal output by the rotary encoder in accordance with the rotation of the rotating body, the switching of returning the level value within a predetermined chattering exclusion time period set in advance after the switching. While performing the switching detection process for detecting the occurrence of switching of the detection target that is the switching of the level value, which is excluded as the switching of the exclusion target due to chattering phenomenon or instantaneous noise, A signal processing method for recognizing a waveform of the pulse signal used for grasping a rotation state of the rotating body,
A switching detection step of performing the switching detection process using the chattering exclusion time while observing the pulse signal output from the rotary encoder;
A time measuring step of measuring an elapsed time from detection of the occurrence of the detection target switching to detection of the next detection target switching;
A chattering exclusion time setting step for variably setting the chattering exclusion time used for the subsequent switching detection processing every time the occurrence of the detection target switching is detected;
Each time the occurrence of switching of the detection target is detected, the set value of the chattering exclusion time used in the switching detection process when the occurrence of the current detection target switching is detected, and the previous detection target switching Depending on the deviation from the set value of the chattering exclusion time used in the switching detection process when an occurrence is detected, the occurrence of the current detection target switching is detected after the detection of the previous detection target switching is detected. A pulse width estimating step for estimating the pulse width of the pulse signal immediately before the occurrence of the current detection target switching by correcting the elapsed time measured in the time measuring step until detection,
In the chattering exclusion time setting step, each time the occurrence of switching of the detection target is detected, it is estimated in the pulse width estimation step so as to shorten the chattering exclusion time as the rotational speed of the rotating body increases. A signal processing method for a rotary encoder, characterized in that the chattering exclusion time is variably set according to at least the latest value of estimated pulse widths. In the signal processing method of the rotary encoder according to claim 1,
The switching detection step includes a first step of detecting switching of an actual level value of a pulse signal output from the rotary encoder, and a pulse signal immediately after the chattering exclusion time period has elapsed after the detection. A second step of observing an actual level value, and the level value observed in the second step is a level value different from the actual level value of the pulse signal immediately before detection of switching in the first step. A signal processing method for a rotary encoder, wherein the occurrence of switching of the detection target is detected when the determination result is affirmative. In the signal processing method of the rotary encoder according to claim 1,
In the switching detection step, the actual level value of the pulse signal output from the rotary encoder is sequentially observed at a predetermined cycle, and a plurality of observed values of the level value for the chattering exclusion time period are the same. The process of sequentially determining whether or not the value is a value is detected, and when the determination result changes from a negative situation to a positive situation, the occurrence of the detection target switching is detected. Signal processing method for a rotary encoder. In the signal processing method of the rotary encoder according to any one of claims 1 to 3,
In the chattering exclusion time setting step, the chattering exclusion time is shortened as the rotational speed of the rotating body indicated by the latest value of the pulse width estimation values estimated in the pulse width estimation step increases. A signal processing method for a rotary encoder, wherein the chattering exclusion time is set according to the latest value. In the signal processing method of the rotary encoder according to any one of claims 1 to 3,
In the chattering exclusion time setting step, the occurrence of the current detection object switching is generated based on the latest value of the estimated pulse width values estimated in the pulse width estimation step and at least one past value before that. Predicting the future rotational speed of the rotating body after detection of the above, the pulse width of the pulse signal corresponding to the rotational speed, or the parameter value having a certain correlation with the rotational speed or pulse width, The chattering exclusion time is set according to the predicted value so that the chattering exclusion time is shortened as the rotational speed of the rotating body indicated by is higher. In the rotary encoder signal processing method according to claim 4 or 5,
A step of determining whether the rotational speed of the rotating body indicated by the latest value of the estimated values of the pulse width estimated in the pulse width estimating step is a high speed equal to or higher than a predetermined value; Is positive, prohibiting the process of grasping the rotation state of the rotating body based on the waveform of the pulse signal recognized by reflecting the detection target switching detected in the switching detection step. A rotary encoder signal processing method.

Images