JP3166496B2 - Eccentric error correction device for rotary encoder - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a rotary encoder (pulse encoder) which is mounted coaxially on a rotating shaft of a large rotating body such as a rotor of a generator and outputs a pulse signal proportional to the rotation angle of the rotating body. More specifically, when an output signal that is correct in time phase cannot be obtained due to the eccentricity of the rotating disk that constitutes the rotary encoder, an output signal corrected for this is created. The present invention relates to an eccentricity error correction device for a rotary encoder that outputs an output signal.

[0002] In the drawings, the same reference numerals indicate the same or corresponding parts.

[0003]

2. Description of the Related Art A pulse encoder is coaxially mounted on a rotating shaft of a large rotating body such as a rotor of a generator, and the number of rotations of the rotating body is determined from a pulse train output from the pulse encoder according to the rotation of the rotating body. It is practiced to know the rotation speed and control the rotation of the rotating body accordingly. If the rotating body is a generator rotor, the rotation of the rotor will be controlled using the output of the pulse encoder.
Eventually, the output voltage and output current from the generator will be controlled.

FIG. 8 is a conceptual diagram for explaining the operation principle of such a pulse encoder. In the figure, 1 is a rotating disk, c is its rotation center, 2a to 2f are projections,
3a and 3b are pickups (also referred to as pulse sensors). The projections 2a and 2b are located at a position 180 degrees apart from each other when viewed from the rotation center c (in other words, the projection 2a, the rotation center c, and the projection 2b are on a straight line). For the set of projections 2c and 2d,
The same is true for the pair of and 2f. The pickups 3a and 3b are located at certain fixed positions away from the disk 1, but their relative arrangement is still the pickups 3a and 3b.
The rotation center c and the pickup 3b are arranged so as to be aligned on a straight line.

When the disc 1 is mounted coaxially with a rotating body (not shown) and rotates about the center c, the projection 2
a to 2f sequentially pass near the pickups 3a and 3b, and the pickups 3a and 3b output a pulse each time they are detected. The relationship between the pickup and the protrusion may be such that the protrusion is a magnet and the pickup is a coil, or the protrusion is a light reflector and the pickup is a photoelectric conversion element. The light may be reflected to be incident on the pickup.

Now, considering the set of the projections 2c and 2d, when the projection 2c approaches the pickup 3a and the pickup 3a outputs a pulse, the projection 2d approaches the pickup 3b and the pickup 3b at the same timing. Output pulse. A set of projections 2a and 2b,
The same is true for the set of projections 2e and 2f. In this manner, the pickups 3a and 3b output pulses at the same timing, and the rotation speed and the number of rotations of the rotating body can be known from the pulse train.

In the above-described pulse encoder, a problem arises when the rotation center of the disk is eccentric. FIG. 9 is an explanatory diagram showing the state of eccentricity of the rotation center in the pulse encoder. As can be seen from the figure, although the physical center of the disk 1 is c, the eccentricity occurs for some reason, and the actual rotation center c ′ is eccentric and deviates from the disk center c. I do. The shifted range 5 is a range in which whirling occurs.

FIG. 10 shows a pickup 3a,
It is a timing chart of the pulse which 3b outputs. When the disk 1 is not eccentric and its center of rotation c 'coincides with the physical center c, the timing of the pulse a output by the pickup 3a and the timing of the pulse b output by the pickup 3b are: As can be seen from FIG.

However, if the center of rotation c 'is deviated from the physical center c due to eccentricity, the pulse b output by the pickup 3b will be smaller than that of the pickup 3 depending on the degree of the deviation.
The pulse a is advanced by the time Δt as shown in FIG. 10B or delayed by the time Δt as shown in FIG. As described above, the pulse output from the pickup 3a and the pulse output from the pickup 3b match in terms of timing if the rotating disk does not have eccentricity, but are temporally shifted due to the eccentricity. Whichever pulse is employed, the timing is not accurate. That is, considering the case where the rotating body is rotating at a constant angular velocity, if there is no eccentricity, the pulses output from the pickup 3a (or 3b) are always output at the same timing with a constant interval, but there is eccentricity. When the output pulse of either of the pickups 3a and 3b is viewed, there is a delay and advance, and it is not possible to represent a situation where the rotating body is rotating at a constant angular velocity.

As a solution to this problem, Japanese Patent Application Laid-Open No. 1-113614, filed by the present applicant, discloses a method in which the pulse output from the pickup 3a and the pulse output from the pickup 3b are positioned exactly in the middle. An apparatus has been proposed which artificially generates a corresponding pulse and outputs the pulse as a truly corresponding pulse.

[0011]

However, the above-mentioned prior art device can correct only the eccentricity error based on the primary rotational vibration (1 cycle / rotation) of the shaft, and for example, the higher order vibration of the shaft based on modern control theory. However, there remains a problem that the method cannot be applied to high-precision control of a generator in consideration of the above. Accordingly, an object of the present invention is to provide an eccentricity error correction device for a rotary encoder that can correct an eccentricity error of a high-order rotational vibration (multi-cycle / rotation) of a shaft and can detect a rotation angle of the shaft with high accuracy. I do.

[0012]

According to a first aspect of the present invention, there is provided an eccentricity error correcting device according to the first aspect of the present invention, wherein a rotating disk (1 or the like) that can rotate eccentrically rotates. Position identifiers (protrusions 2a, 2a) provided at equal intervals on the circumference
, etc.) pass through the vicinity and generate three or more (pickups 3-1, 3-2, 3-3, etc.) for generating position detection signals (pickup output pulses 20, etc.). The position detection signals (21, 2) generated by the detectors at fixed positions near the periphery of the rotating disk, at substantially equal intervals when viewed from the center of the rotating disk, and when there is no eccentricity of the rotating disk.
2, 23, etc.), and the total number of the position identifiers is an integral multiple of the total number of the detectors. (Phase difference times t1, t2, t3), a position detection signal after eccentricity error correction (hereinafter referred to as a corrected position detection signal, such as an average pulse output 30) is generated from the output position detection signal, and the rotation encode output is generated. Eccentricity error correction device (6 etc.)

Each time the corrected position detection signal is output, the number of occurrences (the number N of pick-up output pulses) is reset to a predetermined original clock frequency for timing each time a position detection signal is generated for each detector. A clock frequency (fc) obtained by dividing (fc) by the total number (n) of the detectors.
/ N) to generate a variable clock (43a or the like) having a frequency (fc × N / n) multiplied by the frequency multiplication means (latch 4).
1, adder 42, programmable counter 43, etc.)
The variable clock, which is reset every time the corrected position detection signal is output and whose frequency is sequentially multiplied by the frequency multiplying means, is counted and integrated. Means (fc × T) for outputting the corrected position detection signal at a point in time when the product coincides with the product (fc × T) of a predetermined delay time (T) smaller than the period of the position detection signal output from one detector. Counter 44, monostable FF 45, etc.).

An eccentricity error correcting device according to a second aspect of the present invention provides a position identifier (protrusion 2) provided at equal intervals on the circumference of the rotating disk with the rotation of the rotating disk (1 or the like) which can rotate eccentrically.
a, 2b,...) pass through the vicinity and generate position detection signals (such as the pickup output pulse 20).
-3) at a fixed position near the periphery of the rotating disk, at substantially equal intervals as viewed from the center of the rotating disk, and when there is no eccentricity of the rotating disk, the position detection signal ( 2
1, 2, 23, etc.)
In addition, the total number of the position identifiers is set to an integral multiple of the total number of the detectors, and each detector outputs a phase shift (phase difference time t1, t2, t3) due to the eccentricity of the rotating disk. An eccentricity error correction device (6 or the like) for a rotary encoder that generates a position detection signal after eccentricity error correction (average pulse output 30 or the like, hereinafter referred to as a corrected position detection signal) from the obtained position detection signal and outputs the rotation encoded output. Each pulse of a predetermined original clock frequency (fc) for timekeeping is applied to all the detectors assigned in a predetermined order in a one-to-one manner.
Means for cyclically assigning the end of the detector order to the head (1 / n counter 51, decoder 52, etc.); resetting each time the correction position detection signal is output; After each successive position detection signal is generated, means for validating each pulse of the original clock frequency assigned to the corresponding detector (latch 41, AND
Gates 61 to 63, etc.), and also resets each time the corrected position detection signal is output, counts and integrates only the pulses of the original clock frequency that are validated as described above, and this integrated value is compared with the original clock frequency. Means (OR gate 53, fc × T) for outputting the corrected position detection signal at a time when the product coincides with a product (fc × T) of a predetermined delay time smaller than the cycle of the position detection signal output from one detector. Counter 44, monostable FF 45, etc.).

[0015]

In a pulse encoder having three or more projections at equal intervals on the circumference, at a fixed position opposed to the circumference of the rotary disk, substantially at equal intervals along the circumference, and when the rotary disk is not eccentric, 3 so that the phases of the pickup output pulses match.
A plurality of pickups are provided, and the average timing of the phase-shifted pulses output from the respective pickups due to the eccentricity of the rotating disk is defined as the true pulse timing.

FIG. 2 shows the rotation angle θ of the disk of the rotary encoder mounted on the eccentric shaft of the primary vibration and the rotation angle error detected by the pickup when three pickups 3-1 and 3-3 are used. An example of the relationship of Δθ is shown. As described above, the detected angle error Δθ becomes like a three-phase alternating current.
It can be seen that the average value of the angle errors of the two pickups is zero.

However, the true average timing is at least halfway between the earliest detected pick-up output pulse detected earlier and the latest detected pick-up output pulse detected later. Cannot be detected. For this reason, a point in time which is delayed from the true average timing by a fixed delay time T (provided that the time is sufficiently shorter than the period of the output pulse of one pickup) and after the generation of the pickup output pulse detected latest. , A pulse as a rotation encode output is generated, and is used instead of the pulse at the true average timing.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a diagram showing the entire configuration of the principle of an embodiment of the present invention. In the present invention, three or more pickups are provided at substantially equal intervals along the circumference of the rotating disk 1 to enhance the effect of pulse averaging. In this example, 3
An example in which three pickups (3-1, 3-2, 3-3) are provided is shown. Projection 2 provided on the circumference of rotating disk 1
are arranged at equal intervals, and when the center c of the disk and the center of rotation c 'coincide with each other, each pickup 3-
The pickup output pulses 20 (21, 22, 23) of the same phase are always generated from 1, 3-2 and 3-3. The eccentricity error correction device 6 is delayed by a fixed delay time T from the average timing (true pulse positions) of the pickup output pulses 21 to 23 out of phase due to the mismatch between the disk center c and the rotation center c ′. A pulse (referred to as an average pulse output) 30 is generated and output.

FIG. 3 is a time chart for explaining the operation of FIG. In FIG. 3, t ₁ true time difference between the output pulses 21, 22, 23 of each of the pick-up 3-1, 3-2 and 3-3 with respect to the pulse position 32 (called retardation time) t, t _2, t ₃ (However, in this example, t ₁ , t ₃ > 0, t
_{If 2} <0), the following equation (1) is established.

[0020]

T ₁ + t ₂ + t ₃ = 0 (1) That is, the output pulses 21, 22, 22 of each pickup are always obtained.
The time point at which the sum of the arrival time differences (phase difference times) of the 23 becomes zero is the true pulse position 32. Then, the eccentric error correction device 6 generates an average pulse output 30 at a position 33 which is delayed from the true pulse position 32 by a certain delay time T. However,
It is assumed that the delay time T is sufficiently shorter than the output pulse period from one pickup.

The equation (1) can be modified as the following equation (2).

[0022]

## EQU2 ## t ₁ + t ₂ + t ₃ + 3T = 3T∴ (T + t ₁ ) / 3 + (T + t ₂ ) / 3 + (T + t ₃ ) / 3 = T (2) Equation (2) is the time of the hatched portion in FIG. Length, that is, the elapsed time (T + t ₁ ) after the generation of each pulse 21, 22, 23,
The position where the average of (T + t ₂ ) and (T + t ₃ ) is the delay time T is the average pulse position 33. In other words, this 2 (Expression) shows that each pickup 3-1 and
When the elapsed time after the arrival of the pulse is counted at a clock advance of 1/3 for each of the channels from 3 to 3-3, the point where the sum of the counted values coincides with the delay time T is the time T from the true pulse position 32. This indicates that the average pulse position 33 is delayed.

[0023] Generally as n-channel from where (2) both sides multiplied by a predetermined clock frequency f _c for counting the of, and three pickup 3 (3 channels) (2)
By rewriting the equation, the following equation (3) is obtained.

[0024]

Equation 3] _{(T + t 1) f c} / n + (T + t 2) f c / n + (T + t 3) f c / n + ··· + (T + t n) f c / n = T × f c ··· (3 FIG. 4 is an explanatory view of the calculation of the equation (3). The area of the step-like figure F in FIG. 4 is the value on the left side of the equation (3), that is, the generation of the pickup output pulses 21, 22,. The elapsed time (T + t ₁ ), (T +
_{t 2), ··· (T +} t each the sum of the clock count value which is the product of the clock division frequency f _c / n of _n) (which indicates the total number of clocks) 31, the total number of clocks 31 value It is equal to the value T × f _c of the right side of equation (3).

FIG. 5 is a block circuit diagram showing the configuration of the eccentricity error correcting device 6 according to the first embodiment of the present invention which outputs the average pulse 30 based on the principle of equation (3). In FIG. 5, a latch 41 detects a pickup output pulse 2
When the pulse edges of 1, 22,... 2n arrive, each of the incoming pulses is held, and the pulse latch outputs 81 to 8n as output signals are enabled one by one for each holding. The adder 42 adds the number of valid output lines of the latch 41, outputs the arrival number N of pickup output pulses as the added value, and gives the number to the programmable counter 43 as a program input. Programmable counter 43 as a programmable divider counter to the clock frequency f _c and the original input clock frequency obtained by dividing the clock frequency f _c in accordance with the value of a given pick-up pulses arriving number N f _c A variable clock 43a of × N / n is output.
f _c × T counter 44 counts the variable clock 43a, the count value is output a count-up signal 44a reaches the f _c × T, a monostable flip-flop (flip-flops abbreviated as FF) given to the 45 . The monostable FF 45 outputs the count-up signal 44
After the input of a, an average pulse output 30 as an H level output signal is generated for a certain period of time. This average pulse output 30
Reset these be given to the latch 41, f _c × T counter 44.

As a result, the same operation as described above is repeated for pickup pulses 21 to 2n arriving next time. The practice in the circuit of FIG. 5, it is difficult to increase the clock frequency f _c from the limit of response speed of the programmable counter 43 increases the number of the pickup 3, it is difficult to carry out more accurate rotation angle measurements.

FIG. 6 shows a circuit configuration of an eccentricity error correcting device 6 according to a second embodiment of the present invention which solves this problem. In FIG. 6, instead of the adder 42 and the programmable counter 43 of FIG.
AND gates 61, 62, 63,..., OR gate 5
3 is used. However, the _fc × T counter 4 shown in FIG.
4 counts the clock f _c only input period of the counter enable signal 90.

FIG. 7 is a time chart for explaining the operation of FIG. Next, the operation of FIG. 6 will be described with reference to FIG.
1 / n counter is a counter which divides the clock frequency f _c to 1 / n, in the form of the frequency f _c are counted and n a (= total number of the pickup 3) the count-up value, 1
To n are repeated, and the counted value is output as an _fc / n code.

The decoder 52 decodes the f _c / n code, at a frequency of f _c / n, and the 1 / n period increments order in one cycle such that H level, as a signal whose phase is shifted in the order f _c / n clock phase signals 71, 72,
, And AND gates 61, 62, 63,
Give to ... On the other hand, each AND gate 61, 62, 6
6n (where 6n is not shown), signals (pulse latch outputs) 81, 82 which hold and output the H level when pickup pulses 21, 22, 23,. , 83,... 8n. Therefore, the AND gates 61, 62, 63,...
_{The c} / n clock phase signals 71, 72, 73,.
Only during the level period, the H-level signals 61a, 62a, 6
, And the OR gate 53 outputs the input signal 61
a, 62a, 63a,... are at the H level while the counter enable signal 9 is at the H level.
Outputs 0. Thus f _c × T counter 44 counter enable signal 90 counts the clock f _c only period is at H level, and outputs a count-up signal 44a when the count value reaches the f _c × T.

That is, in FIG. 6, when counting the value of (T + t ₁ ) × f _c / n, for example, on the left side of equation (3),
Instead of directly counting without dividing the clock f _c, and counting only the period corresponding to the f assigned by _c / n clock phase signal _{71 (T + t 1) /} n, the (T ₁ + t ₁₎ This means that the period of / n is made by time division.

[0031]

According to the first aspect of the present invention, the number of pickups 3 for detecting the passage of the projections 2 of the rotating disk 1 is three or more, and the programmable counter 43 for variably dividing the clock frequency fc is used. Each time an output pulse arrives, the programmable counter 43 outputs a variable clock 43a having a frequency fc × N / n (n is the total number of pickups) proportional to the number N of arrivals.
Since the counter 44 counts the variable clock 43a to generate the average pulse output 30, the following effects can be obtained.

(1) An eccentric error can be canceled by a rotary encoder having no pitch error. (2) The measurement error caused by the pitch error of the rotary encoder or the installation error of the pickup (pulse sensor) is
It is averaged to 1 / n for one location. (3) Even if the rotary encoder itself is huge and generates dynamic distortion, if the total number n of pulse sensors is increased, the averaging effect can be obtained similarly. (4) The number of channels of the pickup can be increased, and the clock frequency can be varied by dividing the frequency of the programmable counter in a circuit having a small number of input channels, so that the circuit can be downsized.

According to the invention of claim 2, each pulse of the clock frequency fc is assigned to each pickup by the counter 51 for dividing the clock frequency fc by 1 / n and the decoder 52 for decoding the output of the counter. The elapsed time after the arrival of the pulse is counted by counting the allocated pulses via the fc × T counter 44, that is, by counting the time division of the clock frequency fc, and an average pulse output is obtained. In addition to easily obtaining a pulse detection channel, the circuit of the eccentricity error correction device having a large number of detection channels does not use a programmable counter that is difficult to cascade with a high-frequency clock, so that circuit design can be facilitated.

[0034]

[Brief description of the drawings]

FIG. 1 is a diagram showing the entire configuration of a principle as an embodiment of the present invention.

FIG. 2 is a diagram showing an example of the relationship between the rotation angle of the disk shown in FIG. 1 and an angle error detected by a pickup;

FIG. 3 is a time chart showing the relationship between the pickup output pulse of FIG. 1 and the average pulse output by the eccentricity error correction device.

FIG. 4 is a time chart for explaining the operation of the invention according to claim 2;

FIG. 5 is a block circuit diagram showing a configuration of the invention according to claim 2;

FIG. 6 is a block circuit diagram showing a configuration of the invention according to claim 3;

FIG. 7 is a time chart for explaining the operation of FIG. 6;

FIG. 8 is an explanatory diagram of the operating principle of the pulse encoder.

FIG. 9 is an explanatory view of an eccentric state of a rotating disk.

FIG. 10 is a timing chart of a pickup output pulse in FIG. 9;

[Explanation of symbols]

1 Rotating disk 2 (2a, 2b, ...) Projection 3 (3-1, 3-2, 3-3) Pickup 5 Whirling range 6 Eccentric error correction device 20 (21, 22, 23, ...) 2n) pickup output pulse 30 average pulse output 31 total clock number 32 true pulse position 33 average pulse position T delay time f _c clock frequency _{(clock) t (t 1, t 2} , t 3) the phase difference time 41 latch 42 added vessel 43 programmable counter 43a variable clock 44 f _c × T counter 45 monostable FF n pickup output pulses arriving number n detectors of the total number 51 1 / n counter 52 decoder 53 OR gates 61, 62, 63, · · · the AND gate 71 _{, 72,73, ··· 7n f c /} n clock phase signals 81, 82, 83, · · · 8n pulse latch Force 90 counter enable signal

Claims (2)

(57) [Claims] 1. A position detection signal is generated by detecting that position identifiers provided at equal intervals on the circumference of the rotating disk pass close to the rotating disk as the rotating disk rotates eccentrically. Three or more detectors at fixed positions near the periphery of the rotating disk, at substantially equal intervals when viewed from the center of the rotating disk, and the position detection generated by each detector when there is no eccentricity of the rotating disk. The signals are arranged so as to have the same phase, and the total number of the position identifiers is an integral multiple of the total number of the detectors. Each of the detectors is output with a phase shift due to the eccentricity of the rotating plate. An eccentricity error correction device for a rotary encoder that generates a position detection signal after eccentricity error correction (hereinafter referred to as a corrected position detection signal) from the obtained position detection signal and outputs the rotation encoded output. Reset, every detector Frequency multiplying means for generating a variable clock having a frequency obtained by multiplying the number of occurrences of each position detection signal by a clock frequency obtained by dividing a predetermined original clock frequency for timing by the total number of the detectors, Similarly, the variable clock, which is reset every time the corrected position detection signal is output and whose frequency is sequentially multiplied by the frequency multiplying means, is counted and integrated, and the integrated value is output from the original clock frequency and one detector. And a means for outputting the corrected position detection signal at a time when a product of a predetermined delay time smaller than a cycle of the position detection signal is output. 2. A position detection signal is generated by detecting that a position identifier provided at equal intervals on the circumference of the rotating disk passes nearby in accordance with the rotation of the rotating disk that can rotate eccentrically. Three or more detectors at fixed positions near the periphery of the rotating disk, at substantially equal intervals when viewed from the center of the rotating disk, and the position detection generated by each detector when there is no eccentricity of the rotating disk. The signals are arranged so as to have the same phase, and the total number of the position identifiers is an integral multiple of the total number of the detectors. Each of the detectors is output with a phase shift due to the eccentricity of the rotating plate. An eccentricity error correction device for a rotary encoder that generates a position detection signal after eccentricity error correction (hereinafter, referred to as a corrected position detection signal) from the detected position detection signal and outputs the rotation encoded signal. Each pulse is assigned in a predetermined order. All of the detector of the pair to the order 1
And means for cyclically allocating such that the end of the order of the detectors follows the head, and resetting each time the corrected position detection signal is output, and after generating one position detection signal for each of the detectors, Means for validating each pulse of the original clock frequency assigned to the detector; and similarly counting only the pulses which are reset for each output of the corrected position detection signal and are valid as described above of the original clock frequency. A means for outputting the corrected position detection signal when the integrated value coincides with the product of the original clock frequency and a predetermined delay time smaller than the period of the position detection signal output from one detector; An eccentricity error correcting device for a rotary encoder, comprising: