JP2719480B2 - Multi-rotation position detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improved multi-rotational position detecting device capable of detecting a rotational position of one or more rotations. More specifically, the present invention relates to a mechanism for correcting a count error caused by a shift between a reference point of a rotation detector and a reference point of a position detector.

[0002]

2. Description of the Related Art In recent years, a multi-rotational position detecting device has been used for position control in various robots and machine tools.
The multi-rotation position detection device has two detectors, a detection plate for detecting a position within one rotation and a detection plate for detecting the number of rotations (hereinafter, the former is referred to as a position detector, and the latter is referred to as a rotation detector).
It has. The position detector and the rotation detector are each provided with a reference point that is the origin position.To synchronize the count of the rotation detector with the origin output of the position detector and display the absolute position of the multi-rotation angle space , These reference points must be exactly matched. Therefore, it is necessary to precisely align these reference points during assembly.

However, it is very complicated and difficult to accurately align the reference point of the rotation detector with the reference point of the position detector. In addition, when the positions of both reference points are adjusted after assembly, the operation itself is not easy. Further, even when the two reference points are accurately matched at the time of assembling, the detection plate itself or a sensor for detecting the reference points may be shifted due to a subsequent disturbance, and the two reference points may not match.

The disagreement or close arrangement of these reference points makes it impossible to display the absolute position in the multi-rotation angle space. The rotation detector 1 and the position detector 2 were arranged coaxially. A description will be given based on the schematic diagrams of FIGS. FIG. 5 shows a case where the mounting of the detection plate itself is shifted.
The shift in this case is a unique shift that the position detector 2 can have in the manufacturing process. In the case of FIG. 5, the counterclockwise rotation direction (hereinafter C
CW), the reference point 2s of the position detector 2
Is earlier than the reference point 1 s of the rotation detector 1, so that even if the output of the position detector 2 passes the reference point, the number of rotations output by the rotation detector 1 is not added at the hatched position. Conversely, when rotating in the clockwise direction (hereinafter referred to as CW), the rotation number output of the rotation detector 1 is larger than the position detector 2.
Is subtracted before the output of.

On the other hand, FIG. 6 shows a case where a sensor or the like is shifted due to a disturbance such as a temperature change or vibration. In such a case, in the worst case, the reference point output 1s of the rotation detector 1 precedes or delays the reference point output 2s of the position detector 2 for each rotation. In the example of FIG. 6, in the portion indicated by hatching of the position detector 2, the rotation speed output of the rotation detector 1 is synchronized with the rotation speed of the position detector 2, that is, the reference point (origin) of the position detector 2. The displayed rotation speed will not be displayed accurately.

If considered in more detail, the electrical reference point may inevitably deviate due to system factors. The above shift may occur as a sum of a mechanical shift and an electrical shift. These inconsistencies arise because no constraint is imposed on the output values of the rotation detector 1 and the position detector 2.

Therefore, the conventional multi-rotational position detector divides the position information held by the position detector into four output ranges and performs correction based on the relationship with the output of the rotation detector (particularly). No. 62-261016), the rotation counter was reset by the reference position signal of the position detector, and the rotation was counted twice during one rotation, and the drift of the reference position was corrected depending on whether the rotation number was even or odd. (Japanese Unexamined Patent Publication No. 63-3217
issue).

[0008]

However, in these conventional correction methods, the number of rotations is doubled and the number of rotations is reduced by half.
There is a problem that a processing system such as processing becomes complicated. If the processing system becomes complicated, the number of components increases and the cost increases. Furthermore, although the conventional multi-rotational position detector is corrected, it is necessary to match the reference point of the rotation detector and the reference point of the position detector with a certain degree of accuracy, and the production efficiency is not good.

An object of the present invention is to provide a multi-rotational position detector capable of accurately counting the number of rotations with a simple configuration.

[0010]

In order to achieve the above object, the present invention relates to a multi-rotation position detecting device having a rotation detector and a position detector, wherein a reference point of the rotation detector and a position of the position detector are determined. The reference point and the uncertainty area near these reference points are arranged so as to be shifted from each other so as not to overlap with each other, and the rotation detector and the position detector are divided into two areas, respectively, and these areas are unique. Means for transmitting two outputs, and determining means for determining the logical product of the output of the rotation detector and the output of the position detector at the time of rotation detection, based on a determination result of the determining means, It corrects the count output of the rotation detector and counts the number of accurate rotations.

[0011]

Therefore, by arranging the reference point of the rotation detector and the reference point of the position detector so that their uncertainties do not overlap, the distance between the reference point of the rotation detector and the reference point of the position detector is reduced. Is set to a mismatch area. This non-coincidence area is an area where the number of rotations actually counted by the rotation detector does not match the count number which should be originally synchronized with the reference point of the position detector. The non-coincidence area is recognized by the logical detector of the rotation detector and the position detector obtaining the logical product of the respective outputs that make each of the two divided areas unique. Then, the count output of the rotation detector is corrected using the result of the logical product.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The construction of the present invention will be described below in detail with reference to the embodiments shown in the drawings.

FIGS. 1A and 1B show the principle of the present invention. As is clear from this principle diagram, this multi-rotation position detecting device is composed of a reference point 1 s of the rotation detector 1 and a position detector 2.
Are arranged so that the uncertainty areas near both reference points 1s and 2s do not overlap, and an unmatched area 12 is set between them. This mismatch area 12
This is an area requiring correction in which the number of rotations actually counted by the rotation detector 1 does not match the number of counts that should be originally synchronized with the reference point 2s of the position detector 2. The outputs of the position detector 2 and the rotation detector 1 are each divided into two regions, and an E signal (signal values E1, E2) and an R signal (signal values R1, R2) for distinguishing the respective regions are output to respective outputs. Means for synchronous transmission are provided. Alternatively, part of the output of each detector is quoted for this function. The counting of the number of rotations is performed exclusively by the rotation detector 1. However, since the original rotational speed must be synchronized with the reference point 2s of the position detector 2,
There is an angle region where the count of the rotation detector 1 needs to be corrected, that is, a non-coincidence region 12. This region 12 is composed of the regions R1 and R2 of the rotation detector 1 and the region E of the position detector 2.
It is determined as a product set area of E1 and E2. Area distinction is R
Since the signal and the E signal are used, the area determination is obtained by the logical product of the E signal and the R signal.

More specifically, since the problem in the multi-rotation position detector is the relationship between the reference point 1s of the rotation detector 1 and the reference point 2s of the position detector 2, at least each of the reference points 1s, The information obtained from each of the detectors 1 and 2 needs to be classified into two types by 2s.
That is, each of the detectors 1 and 2 is divided into two regions. Mechanically, it can be imagined that the detection plate is divided into two areas in this way, but in order to provide generality, the state of each detection system including the circuit is divided into two areas. It is better to fix it. In the case of this embodiment, each of the detectors 1 and 2 is formed by a magnet, and one half is an N pole and the other half is an S pole, so that two regions R1 and R2 and E1 and E2 are formed. Has been split.

The two regions of the rotation detector 1 have R1 and R2
And the two areas of the position detector 2 have E
The symbols 1, E2 will be added. Here, suppose that the respective reference points 1s, 2 of the rotation detector 1 and the position detector 2
If we consider a system that matches s,
The ideal relationship between R1, R2 and E1, E2 is R1-E1,
Assume that it must be R2-E2. That is, R1 and E
1, R2 and E2 have an equal relationship. This relationship is shown in Table 1.
Shown in

[0016]

[Table 1] However, since the reference point generally deviates for the reason described above, there is also a region where the relationship is R1-E2 and R2-E1. The hatched portions in FIGS. 5 and 6 correspond to this. Now, this "illegal" area of the position detector is E1 'or E2'. The following two types of misalignment can be considered as shown in Tables 2 and 3.

[0017]

[Table 2]

[0018]

[Table 3] Here, the correction relational expression in case A shown in Table 2 is:

[0019]

## EQU1 ## Nc = N + 1 (when R2∩E1 is satisfied; only when R1 is 1 under the AND condition of R2 and E1), and the correction relational expression in case B shown in Table 3 is

[0020]

## EQU2 ## Nc = N-1 (when R1∩E2 holds; only when R1 is 1 under the AND condition of R1 and E2). Here, Nc is the rotational speed after the correction. Equations 1 and 2 are contradictory. Therefore, as shown in FIG. 2, when the front-back relationship between the reference points 1s and 2s of the rotation detector 1 and the position detector 2 is not constant, these relational expressions cannot be used.

However, if it is possible to set the positional relationship between the rotation detector 1 and the position detector 2 so that this relationship is constant in any situation, Equation 1 or Equation 2
The rotation speed can be corrected by using one of the relational expressions.

Therefore, according to the present invention, the reference point 1 of the rotation detector 1 is
By not shifting the reference point s and the reference point 2s of the position detector 2 from the beginning, rather by shifting the reference points 1s and 2s largely so that the uncertain regions near the reference points 1s and 2s do not overlap with each other,
The mismatch area 12 is set between the two reference points 1s and 2s. For example, E1 and E1 ', E2 and E2
It is ideal if the rotation detector and the position detector can be shifted so that the area of the region 'is approximately the same. However, the amount of this shift does not need to be constant if there is a margin to the extent that cases A and B do not coexist. Here, the number of rotations is counted by the count output of the rotation detector 1, and the correction is performed on the counted number.

In FIG. 1, the tips of the arrows are the reference points 1s and 2s of the rotation detector 1 and the position detector 2, respectively. This rotation detector 1 is considered to be on the detection plate recognized by the circuit. The reference points 1 s and 2 s are about 9 as described above.
They are arranged at an angle of about 0 °. In this figure, the area where the value held by the rotation counter of the rotation detector 1 is to be corrected is the area of R2∩E1. In this region, if the rotation is CW, as shown in Table 2, the rotation speed that should be counted by the position detector 2 is counted up before the rotation speed that the rotation detector 1 is counting. If the rotation is CCW, the rotation detector 1 counts down first. Therefore, in this region, the rotation detector 1 holds a rotation number smaller by one than the rotation number that the position detector 2 originally counts.

Here, it is not considered that the number of rotations is counted using the reference point 2s of the position detector 2. The counting of the number of rotations is left to the rotation detector 1 to the last. Whether or not this value should be corrected is determined by whether or not the relationship of R2１E1 holds. The contents of the counter of the rotation detector 1 should not be modified with the information of the reference point 2s of the position detector 2. This is because the relationship between the number of rotations that should be synchronized with the reference point 2s of the position detector 2 and the number of rotations counted by the rotation detector 1 is disturbed.

Further, the operation will be described with reference to the time chart of FIG. In FIG. 2, the upper part is the output of the position detector 2, and the lower part is the output of the rotation detector 1. The portion indicated by hatching is a portion that needs to be corrected in the count of the rotation detector 1, that is, the mismatch region 12. Ideally, the number of revolutions should be counted at the position where the value of the position detector 2 is reset. However, in a situation where the absolute position cannot be detected for some reason, it is needless to say that It is also impossible to detect the reference point 2s of the position detector 2. On the other hand, if the rotation detector 1 is operable at all times, the rotation detector 1 only needs to detect only the rotation speed. Once the absolute position can be detected,
At that time, it is determined whether or not the output of the rotation detector 1 should be corrected using the relationship between the position information and the state of the rotation detector 1, that is, whether or not the relationship of R2∩E1 is satisfied in this case. I just need.

It is better to perform the correction each time the number of rotations is required. As a specific situation, consider a situation at the time of a power failure and a situation at the time of power recovery. If the number of rotations is not required at the time of a power failure, it is sufficient that only the rotation detector 1 operates at the time of the power failure, whether or not there is an incorrect area. Since the position information can be obtained when the power is restored, R2∩E1
It can be checked immediately whether or not holds, and if it holds, it can be corrected immediately. Of course, a delay corresponding to the delay time of the circuit is inevitable. Since the counting of the number of revolutions is left to the rotation detector 1, the exact number of revolutions is determined in each case as to whether or not correction is necessary in response to a request from outside the position sensor, and may be corrected if necessary.

The problem here is that if the reference point 2s detected by the position detector 2 is shifted at random, whether the correction method can be accurately obtained by the above-described method. Alternatively, there is also a problem as to when the reference point 1s of the rotation detector 1 is shifted at random. In particular, the latter problem corresponds to a situation in which when the rotation detector 1 is driven by a pulse and the relationship between the pulse interval and the rotation speed of the detection plate is not constant, the electrical reference point greatly varies.

However, looking at FIG.
If the relationship between the reference point 2s and the reference point 1s of the rotation detector 1 does not change, only the area requiring correction, that is, the area of the mismatched area changes, and the correction method itself does not need to be changed.

Since the reference point 2s of the position detector 2 is shifted at random from the reference point of the mechanical system, the problem that the rotation speed is inaccurate during that time differs in the dimension of the problem. This problem is a problem such as accuracy and reproducibility of the position detector 2. That is, one edge of the area requiring correction is synchronized with the best reference point that can be recognized by the position detector, and the other edge is synchronized with the best reference point that can be recognized by the rotation detector 1. Is an important point.

The next problems are R1 and R2, and E1 and E
The question is how to distinguish the two. When the rotation is detected, for example, a rectangular wave having a duty ratio of 50% and the phase of one cycle of one rotation is shifted by 90 ° is used. One of them may be used for identifying R1 and R2. In the present embodiment, the output of the position detector 2 has a sawtooth shape, but this output is actually digitized by an A / D converter (not shown). Therefore, the output of the position detector 2 has a step shape. In other words, if E1 and E2 are assigned by the address ²ⁿ in the rotation and are assigned by binary integers or gray codes with weights, the most significant bit (MSB) outputs 0 in a half turn and 1 in the other half turn. Can be. In the case of BCD code, it is slightly complicated, but the maximum value is 1
A comparator that compares the value of / 2 may be used, or a combinational logic of bits may be used.

Next, an embodiment in which a binary integer is used for the output from the position detector 2 will be described with reference to FIGS. Here, the operation in the case of rotating in the CW direction will be described. The A-phase signal and the B-phase signal are signals recorded on the rotation detector 1, and may be considered as divided regions R1 and R2. In this embodiment, the number of rotations of the rotation detector 1 is counted by the counter 3 based on the number of rotations. That is, when the rising edge of the A-phase signal is input to the rotation speed counter 3, the rotation speed counter 3 counts up. This is because the A-phase signal is generated once in one rotation. In this embodiment, the configuration is such that the rising edge of the A-phase signal is counted. However, the configuration may be such that the rotation speed counter 3 counts up at the rising edge of the B-phase signal. The rotation speed counter 3 is connected to a full adder (FA) 7 via a bus line 8 and a three-slatch buffer 6. The position detector 2 is configured to be converted into a binary integer and the MSB thereof is input to the AND gate 4 via the inverter 5. The A-phase signal is input to another terminal of the AND gate 4. The logical product of the inverted MSB of the A-phase signal and the A-phase signal is the A-value.
The circuit is made as taken by ND gate 4. The output of the AND gate 4 is input to the full adder (FA) 7 via the least significant bit (LSB) of the bus line 8. When the LSB of the bus line 8, that is, the output of the AND gate 4, is 0, the count output of the rotation counter 3 is output from the full adder (FA) 7 as it is. When the LSB of the bus line 8 is 1, the count output of the rotation counter 3 is corrected by (+1) and output. The full adder (FA) 7 measures the carry bit output, and has a configuration in which the carry bit is set when carry over occurs. In this embodiment, when the logical product of the inverted bit of the MSB output of the position detector 2 and the output of the A-phase signal is 1, the area becomes E1 'in FIG.
Is performed. When considering a power failure, backup up to the rotation counter 3 may be performed using a battery or the like. Further, when the circuit is configured by a device such as a CMOS or the like, the three-slot buffer as the bus controller may be omitted. Further, in this embodiment, a full adder (FA)
7 is used, there is no problem in terms of speed even if a normal adder is used, but if a high-speed processing, that is, a high rotation is required, a carry look-head type adder can be used. good.

Next, performs countdown rotation counter 3 at the falling edge of No. A Aisin in the case of counterclockwise rotation (CCW), rotation number counter when the output of the AND gate 4 is 1 The correction may be performed by counting +1 from the count value of 3. And again, AND gate 4
Is 0, the count value of the rotation counter 3 may be output as it is. In the present embodiment, the circuit is configured by using the device to perform the correction, but it is needless to say that the configuration can be dealt with by software.

Next, a second embodiment will be described with reference to the timing chart of FIG. In this embodiment, the rotation detection system is pulse-driven in order to suppress the average current consumption during a power failure. The A-phase signal and the B-phase signal in the first and second stages are signals recorded in the rotation detector 1. This is sampled by the clock BuCk to obtain a BA signal (corresponding to the A-phase signal) and a BB signal (corresponding to the B-phase signal). The clock BuCk is different from the system clock that controls the entire multi-rotational position detector, but is synchronized with anything.
The rotation speed is counted and corrected based on the BA signal and the BB signal. The operation requiring N + 1 may be performed in a region that satisfies E1∩R2 without changing the region requiring correction at all.

The same circuit as in the first embodiment can be used. However, it is necessary to input the BA signal and the BB signal instead of the A-phase signal and the B-phase signal. Here, the problem is that the rising edge rE of the BA signal that defines the boundary of the correction area with respect to the change in the rotation speed of the detector 1 deviates from the E1 area to E
It is possible to invade two areas. A change in the rotation speed can be reconsidered as a change in the clock frequency. That is, a state in which the rotation speed becomes slow can be considered as a case where the clock frequency becomes high. Conversely, a state where the rotation speed is high can be considered as a case where the clock frequency is low.

Now, reviewing FIG. 4 under such an idea. Actually, the state of FIG. 4 is in an extreme state, and the frequency of the clock cannot be further reduced. That is, four pulses are required for one rotation. If the clock is further delayed, for example, if there are only three pulses in one rotation, the phase relationship between the A-phase signal and the B-phase signal cannot be reflected on the BA signal / BB signal. If so, the correction area cannot be defined.

On the other hand, consider the case where the clock frequency is increased. In this case, since the sampling problem becomes narrower, BA
The waveforms of the signal and the BB signal approach the A-phase signal and the B-phase signal.
In other words, the rise edge rE of the BA signal defining the non-coincidence area, which is the correction area, approaches the time indicated by the dashed line in FIG.
However, the reference points of the position detector 2 and the rotation detector 1 are about 90
If there is a deviation, it will not go into the E2 region beyond this. Even if it approaches the edge of the A-phase signal, it may not penetrate into the E2 region. However, since both ends of the edge of the A-phase signal are easily uncertain due to disturbance, it should be avoided to approach the edge extremely. This margin is 5 °
Position detector 2 within ± 85 °
It is only necessary that the reference position is within. This indicates that it is not necessary to perform strict alignment in the assembly process. As described above, correction can be performed on the sum of both the mechanical system deviation and the electric system deviation, so that the position accuracy and reproducibility of the reference position depend only on the accuracy of the position detector 2.

As described above, even if the rotation speed is increased to a rotation speed at which a four-pulse clock can be secured during at least one rotation, it is considered that normal operation is performed. However, if the edge of the clock is racing with the edge of the A-phase signal and the edge of the B-phase signal, the pulse function is lost, so four pulses are dangerous. In the worst case, the number of pulses should preferably be 8 or more, since there is a possibility that four pulses will be racing at the edge of each of the A-phase signal and the B-phase signal.

Therefore, if the function is compensated up to Nrpm, the clock frequency f should be equal to or higher than the frequency represented by the following equation.

[0039]

F = (2/15) N For example, at 3000 rpm, it is necessary to drive at 400 Hz or more. However, the configuration may be such that the frequency of the clock BuCk changes with the rotation of the rotation detector 1.

Although the above description has been made on the configuration using the relational expression of Expression 1, a configuration using the relational expression of Expression 2 is also possible.
When the relational expression of Expression 2 is used, the position of the reference point of the position detection plate with respect to the rotation detection plate may be set to a position shifted from the above position by 180 °. In addition, the circuit may perform the subtraction. However, considering that the subtraction is generally performed by inputting a two's complement to the full adder (FA) 7,
In the circuit diagram shown in FIG.
dd and the AND gate may be replaced by a NAND gate. The discussion in FIG. 6 holds true if the position information is delayed by 180 °.

The above embodiment is one preferred embodiment of the present invention, but the present invention is not limited thereto, and various modifications can be made without departing from the spirit of the present invention. For example, in this embodiment, a multi-rotational position detector using an encoder method is mainly described, but the present invention is not particularly limited to this, and the present invention can also be applied to a multi-rotational position detector using a resolver. Rather, application to resolvers whose reference points are not clear is effective. Further, in the present embodiment, rotation and position are detected by using magnetism.However, the present invention is not particularly limited to this, and a reference point is formed using a slit, a reflector, or the like. Optical detection may be used.

[0042]

As is apparent from the above description, the multi-rotational position detecting device of the present invention is characterized in that the reference point of the rotation detector and the reference point of the position detector have an uncertain area near these reference points. The rotation detector and the position detector are arranged so as not to be overlapped with each other, the rotation detector and the position detector are each divided into two regions, and two output units are provided to make the regions unique. Determining means for determining the logical product of the output of the rotation detector and the output of the position detector; correcting the count output of the rotation detector based on the determination result of the determination means to obtain an accurate rotation number; Since the counting is performed, the accurate rotation speed is always detected by simply arranging the reference position of the rotation detector and the position detector so that the uncertainty areas near the reference position of both the position detector and rotation detector do not overlap. At the reference position of the container It can be obtained by synchronize. That is, characteristics such as position accuracy and reproducibility of the reference position depend only on the accuracy of the position detector.

Further, in the multi-rotation position detecting device of the present invention, the correction circuit can be basically constituted by, for example, an AND gate (in some cases, an inverter is necessary) and a one-count interleaver, so that the number of parts is small and the cost is reduced. And current consumption can be reduced.

Further, in the multi-rotational position detecting device of the present invention, since the uncertainty area near the reference position is several minutes in angle, the positional relationship between the position detector and the rotation detector is extremely small. Thus, it is not necessary to strictly manage these positional relations in the assembling process of the position detecting device, and the production efficiency can be improved.

[Brief description of the drawings]

FIGS. 1A and 1B are explanatory views showing the principle of the present invention, in which FIG. 1A is a perspective view, and FIG. 1B is a view seen from the direction of the arrow in FIG.

FIG. 2 is a time chart in one embodiment of the present invention.

FIG. 3 is a circuit diagram according to an embodiment of the present invention.

FIG. 4 is a time chart showing another embodiment of the present invention.

FIG. 5 is an explanatory diagram showing a state where the detection plate itself is displaced.

FIG. 6 is an explanatory diagram showing a state where a sensor or the like is displaced by disturbance.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Rotation detector 2 Position detector 1s Reference point of rotation detector 2s Reference point of position detector 12 Non-coincidence area R1, R2 Area that divides rotation detector into two E1, E2 Area that divides position detector into two

Claims (1)

(57) [Claims] 1. A multi-rotational position detecting device having a rotation detector and a position detector, wherein a reference point of the rotation detector and a reference point of the position detector overlap an uncertain region near the reference point. The rotation detector and the position detector are each divided into two regions and output two outputs to make the regions unique, and the rotation detection unit detects the rotation when the rotation is detected. Determining means for determining the logical product of the output of the detector and the output of the position detector; correcting the count output of the rotation detector based on the determination result of the determining means to count an accurate number of rotations A multi-rotational position detecting device.