JP4629458B2 - Linear encoder device - Google Patents

Description

  The present invention relates to a linear encoder used for position measurement of a machine tool such as a milling machine or a semiconductor manufacturing apparatus, and more particularly to alarm detection inside the linear encoder.

  A conventional linear encoder will be described with reference to FIG.

  FIG. 11 is a perspective view illustrating an example of a signal detection unit of a general optical linear encoder. A moving axis position detecting device such as a machine tool includes a scale 10 having a periodic main grid scale 16 and a slider 15. The slider 15 includes a light-transmitting index scale 13 in which four periodic sub-lattices corresponding to the main lattice scale 16 are formed in a square shape, and four light-receiving elements 14 corresponding to the four sub-lattices. Part, the light emitting element 12, and the collimator lens 11 which makes the light emitted from the light emitting element 12 parallel light.

  The light quantity change caused by the relative movement operation of the scale 10 and the slider 15 is photoelectrically converted by the light receiving element 14, whereby four signals a1, a2, b1, and b2 having the same period and different phases are obtained.

These four signals a1, a2, b1, b2 are connected to Sig. Assuming L, the position POS is derived as follows. First, the signal A having no offset and the signal B whose phase is advanced by 90 degrees with respect to the signal A are obtained by the calculation of A = a1-a2 and B = b1-b2. Then, it is derived by an interpolation operation of POS = P × tan ⁻¹ (A / B) / 2π. Here, P is the pitch of the main grid scale 16.

  In this conventional example, the case where the main grid scale (hereinafter referred to as scale) has one track is illustrated, but there may be a plurality of scales having different pitches (not shown).

  Next, a position detection method employed in the above-described linear encoder will be described. As a position detection method in a general linear encoder, an absolute method and an incremental method are known.

  The absolute method is a detection method that reads a plurality of scales with different pitches and has absolute positions at arbitrary positions. The incremental method is a detection method in which the grid pitch of the scale is counted based on the position when the power is turned on, and the position is calculated by integrating the pitch distance and the count value. However, even in the case of the absolute method, since the detection by the incremental method can be performed at high speed, the absolute position is usually established by the absolute method when the power is turned on, and thereafter, the detection by the incremental method is always performed, and the absolute method is appropriately used. A method is used in which the detection position by the incremental method is checked according to the detection position by the method, and a method for improving the reliability of position detection is adopted.

  FIG. 10 shows a schematic configuration diagram of the internal calculation unit of the linear encoder using both the absolute method and the incremental method, and the position detection will be described based on this.

  The signal detection unit 21 in FIG. 10 has the signal detection configuration shown in FIG. 11, but the scale has a two-track configuration of an upper digit scale and a lower digit scale. Two signals of L and Sig.H are obtained. Signal Sig. L is an electric signal detected from the lower digit scale, and Sig. H is an electrical signal detected from the upper digit scale.

  The interpolation calculation unit 22 receives the signal Sig. L, Sig. The above-described interpolation operation is performed by inputting H, and the position data POS. L, position data POS. H is calculated and output to the absolute position calculator 24.

  The absolute position calculation unit 24 calculates precise absolute position data PABS by combining the position data of each digit obtained by the interpolation calculation and outputs it to the position data inspection unit 25.

  Next, incremental position detection will be described.

  The signal detection unit 21 is configured to output the low-order electric signal Sig. L is output to the counter 26. The counter 26 is connected to the electric signal Sig. Based on L, the count value is counted at every pitch interval of the lower digits. For example, the electrical signal Sig. Each time L rises or falls, one counter is added or subtracted to obtain a count result CNT, which is output to the counter position calculator 23. The counter 26 is set to an initial value by a counter set signal CSET based on the data of the absolute position calculation unit 24 when the power is turned on. This initial value is a value corresponding to the upper digit of the absolute position PABS.

  The counter position calculation unit 23 includes the count CNT and the lower digit position data POS. The absolute position data PINC is calculated by adding L. The absolute position data PINC is calculated using the count value CNT that has started counting after the initial value is set by the counter set signal CSET, so that the alignment with the position data PABS from the absolute position calculation unit 24 is ensured. The

  Thus, the calculation results PINC and PABS obtained in the counter position calculation unit 23 and the absolute position calculation unit 24 are supplied to the position data inspection unit 25.

  The position data inspection unit 25 determines whether or not a position detection error has occurred by comparing the absolute position data PABS and the position data PINC. When a position detection error is detected in the position data inspection unit 25, an alarm signal PALM is generated and output to the control device 50.

  A linear encoder that performs such absolute and incremental position detection is disclosed in Patent Document 1 and the like. Further, alarm determination is disclosed in Patent Document 2 and the like.

JP2003-240605A Japanese Patent Laid-Open No. 8-76841

  However, the following problems occur in a linear encoder that employs a method for improving the reliability of position detection using the two methods as described above.

  The method of checking the detection value by the incremental method based on the detection value by the absolute method. Even if the position of the incremental method is accurate, if the detection value of the absolute method is incorrect, an alarm is generated.

  For example, if dirt is attached only to the upper digit scale, the absolute detection value is incorrect. In addition, even if the upper digit scale and the lower digit scale are evenly soiled, the position detection value from the upper digit scale with a wide pitch has the same resolution and accuracy as the detection principle and configuration. Inferior, the detection error is large on the upper digit scale side, so that the detection position is incorrect only on the upper digit scale side.

  For example, when the pitch interval of the lower digit scale is 80 μm and the pitch interval of the upper digit scale is 2 mm, the upper digit scale and the lower digit scale are evenly soiled and a 3% position error occurs. Then, a position error of 0.06 mm is obtained at the upper digit scale, and a position error of 2.4 um is obtained at the lower digit scale.

  At this time, the position POS output to the control device 50 is an incremental position calculated from the integrated value of the counter 26. If the position error of the lower digit scale is about 2.4 μm, the counter 26 does not malfunction. , Can output the exact position.

  On the other hand, as disclosed in Patent Document 2, when accurate position detection is performed, the alarm determination criterion is set to be less than ± half pitch of the lower digit scale (for example, ± 40 um in the case of the lower digit scale 80 um). A position error of 0.06 mm on the upper digit scale is determined as an alarm.

  In this way, the combination of the two methods is highly reliable, but in principle it detects an alarm even when it is not necessary to detect the alarm, so it detects the alarm excessively and increases the opportunity to stop the machine. There's a problem.

  Further, in order to prevent excessive detection of alarms, for example, when taking measures such as setting a large alarm judgment value in advance and lowering the detection sensitivity, there is a problem that detection is not possible even if the position is incorrect.

  An object of the present invention is to enable highly reliable alarm detection while improving the reliability of position detection.

The present invention provides a scale having a plurality of tracks having different pitches, a position obtained by performing an absolute method operation on a signal obtained from the upper and lower digit scales of the scale, and a lower digit scale. When the difference between the two obtained positions is greater than the alarm judgment value by comparing with the position obtained by performing an incremental calculation obtained from the count value obtained by counting the obtained signals A linear encoder device having a slider for detecting an alarm, further comprising a determination value calculation unit for changing the alarm determination value, and the determination value calculation unit counts signals obtained from the lower digit scale by an incremental method. When rewriting the initial value or offset value of the counter used when The alarm determination value is set to be smaller than ± half a pitch lower digits scale, the slider after moving a predetermined distance, and sets the determination value greater than ± half a pitch lower digits scale.

  The judgment value to be set should be less than ± half pitch of the lower digit scale when the alarm judgment value is set small, and the alarm judgment value should be ± 1 or more half pitch of the lower digit scale when extending the judgment criteria. It is preferable that the movement distance for changing the alarm determination value to be less than the pitch is 1 pitch or more of the upper digit scale.

  According to the present invention, it is possible to detect a position with high reliability without generating an excessive alarm by detecting an alarm state more accurately.

  An embodiment of the present invention will be described below with reference to the drawings. In the configuration of the linear encoder, the description of the same parts as those in the conventional example is omitted. Further, the description of the position calculation method when reading from the slider is the same as in the conventional example will be omitted.

  FIG. 1 is a linear encoder using the absolute position detection method and the incremental position detection method according to the present embodiment. The difference from the conventional example of FIG. 10 is that an alarm determination value ALVL is output to the position data inspection unit 25. This is a point where a judgment value calculation unit 60 is newly added.

  When the counter initial value is input by the counter set signal CSET, the determination value calculation unit 60 changes the alarm determination range to the alarm determination value ALVL having a narrow alarm determination range, and sets this alarm determination value in the position data inspection unit 25. The position data inspection unit 25 detects whether the error between the absolute position detection method and the incremental position detection method is within the alarm determination value ALVL, as in the conventional example. When the error exceeds the alarm determination value ALV, an alarm signal PALM is generated and supplied to the control device 50.

  Next, when the slider moves, the determination value calculation unit 60 monitors the position detection value PINC supplied from the counter position calculation unit 23, and when the slider moves more than a predetermined distance, an alarm in the position data inspection unit 25 is detected. The judgment value ALVL is set to a large value to widen the alarm judgment range. As a result, the position data inspection unit 25 increases the allowable range of errors between the absolute position detection method and the incremental position detection method, and as a result, can increase the allowable range of position errors that occur in the track of the absolute position detection method.

  The alarm judgment value ALVL is set to be less than ± half pitch of the lower digit scale in order to detect that the initial value is incorrect when the initial value is set in the counter by the counter set signal CSET.

  Whether the initial value of the counter is incorrect cannot be determined when the slider is at the same position, so it is necessary to continue to detect an alarm based on the same reference during a certain period of movement. For this reason, after moving a predetermined distance, the criterion is changed to less than ± 1 pitch of the lower digit scale.

  As a result, the alarm state can be detected more accurately, an excessive alarm is not generated, and a highly reliable position can be detected.

  Next, the above operation will be described with reference to FIGS.

  First of all, the problem with conventional technology is that even if the position of the incremental method is normal during normal operation, the position detection by the absolute method causes a position error. This can be reduced by expanding the slider to less than ± 1 pitch after the slider has moved a predetermined distance.

  FIG. 2 shows the relationship between the position error ERR obtained by subtracting the detection value by the incremental method from the detection value by the absolute method and the position POS on the scale. Here, a case is shown where the scale is dirty and the points B and C are not dirty near the point A of the scale position POS. The position error ERR shows a maximum of +60 μm (μm: hereinafter referred to as “um”) in the vicinity of the point A of the POS and becomes a normal position error in other places. Therefore, the slider is located at the points B and C. In the case of the conventional technology, an alarm is detected at a position where the position error ERR exceeds 40 μm near the position A.

  However, in this embodiment, since the alarm determination value is less than ± 80 μm, no alarm is detected at the position A.

  FIG. 3 shows a case where, in position detection by the incremental method, a count error occurs due to the above-described contamination of the scale, and a position error corresponding to one pitch (80 μm) of the lower digit occurs. Compared with FIG. 2, it can be seen that the position error ERR is reduced by −80 μm. In this case, if the alarm judgment criterion is set as less than ± 1 pitch of the lower digit track, the alarm cannot be detected from the position D to the point F in FIG. 3, but the alarm is detected at a position below −80 μm near the point G. It can be detected.

  That is, since the alarm can be detected by moving the slider, it indicates that there is no problem even if the alarm determination value is set to be less than ± 1 pitch.

  Next, the reason why it is necessary to set the alarm judgment value to be less than ± half pitch of the lower digit scale when the power is turned on and to continue the reference for a period of one cycle or more of the upper digit scale.

  Since the scale of interest in this embodiment has an upper digit scale and a lower digit scale in the scale, as shown in FIG. 6, the positions of the upper digit receiver and the lower digit receiver are separated by a distance X as shown in FIG. Yes. Further, the scale and the slider can be mounted in parallel in the normal normal state as shown in FIG. However, if the scale and slider are tilted as shown in FIG. 9, the positions of the upper digit receiver and lower digit receiver are also tilted as shown in FIG. An error represented by the distance Z occurs at the position of the part.

  Position detection in the above state will be described with reference to FIG.

  FIG. 4 shows that there is no dirt on the scale, but when the slider is attached with an inclination to the scale, the position error ERR obtained by subtracting the detection value by the incremental method from the detection value by the absolute method and the scale error Represents the relationship of the position POS.

  In this state, the error between the position detected by the absolute method and the position detected by the incremental method has an offset of error Z as shown in FIG. 4, but in this state, there is a slider at the position K point. When the power is turned on in the state, an error of 1 occurs in the counter initial value when the counter initial value is calculated.

  FIG. 5 shows an error when an error of 1 occurs in the counter initial value. Although the position of 80 μm is shifted from that in FIG. 4, an error of −80 μm is not reached at the maximum. That is, there is a problem that an alarm cannot be detected when the alarm determination value ALVL described in the above embodiment is less than ± 1 pitch (80 um) of the lower digit scale.

  In this embodiment, when the counter initial value is calculated, the alarm determination value ALVL is set to be less than ± half pitch.

  When the alarm judgment value ALVL is less than ± half pitch, the power is turned on at the point Q in FIG. 5, and if an error of 1 occurs in the counter initial value, the slider moves, and the position error ERR near the point P is − When it becomes 40 um or less, an alarm can be detected. Moreover, since errors are mostly generated in the upper digit scale, the error changing period is within a distance corresponding to one cycle of the upper digit scale. For this reason, the distance from the position Q point to the position P point is always less than one pitch of the upper digit scale. Therefore, at least a distance between one pitch of the upper digit scale in the interval where the alarm determination value ALVL is less than ± half pitch. That is all.

  In this embodiment, the timing for changing the alarm determination value is the timing for changing the initial value of the counter. However, the value is set not for the initial value of the counter but for the offset of the counter that performs the same function. The same effect can be obtained in terms of timing.

  Furthermore, the present invention can be similarly applied to a long linear encoder in which two scales described in detail in Patent Document 1 and the like are arranged apart from each other, and an equivalent effect can be obtained.

It is a figure which shows the structure of the linear encoder which concerns on one Embodiment of this invention. It is explanatory drawing which shows an example of the state of an error signal. It is explanatory drawing which shows an example of the state of an error signal. It is explanatory drawing which shows an example of the state of an error signal. It is explanatory drawing which shows an example of the state of an error signal. It is a figure which shows the installation state of a light-receiving part. It is a figure which shows the installation state of a light-receiving part. It is a figure which shows the installation state of the slider with respect to a scale. It is a figure which shows the installation state of the slider with respect to a scale. It is a figure which shows the structure of the linear encoder by a prior art. It is optical part explanatory drawing of the linear encoder by a prior art.

Explanation of symbols

  10 scale, 11 collimator lens, 12 light emitting section, 13 sub-lattice (index scale), 14 light receiving element, 15 slider, 16 main grid scale, 21 signal detection section, 22 interpolation calculation section, 25 position data inspection section, 26 counter , 23 counter position calculation unit, 40 position detection unit, 50 control device, 60 judgment value calculation unit.

Claims (2)

A scale with multiple tracks with different pitches,
From the position obtained by performing an absolute calculation on the signal obtained from the upper and lower digit scales of the scale and the count value obtained by counting the signals obtained from the lower digit scales A slider that detects an alarm when the difference between the two obtained positions is larger than the alarm judgment value by comparing the position obtained by performing the required incremental method calculation;
In a linear encoder device comprising:
Furthermore, a judgment value calculator for changing the alarm judgment value is provided,
In this judgment value calculation unit, when the initial value or offset value of the counter used when counting the signal obtained from the lower digit scale by the incremental method is rewritten, the alarm judgment value is calculated from ± half pitch of the lower digit scale. Set smaller,
A linear encoder device characterized in that, after the slider has moved a predetermined distance, the determination value is set larger than ± half pitch of the lower digit scale . The linear encoder device according to claim 1,
The judgment value to be set should be less than ± half pitch of the lower digit scale when setting the alarm judgment value small, and the alarm judgment value should be set to ± half pitch or more of less than ± 1 pitch of the lower digit scale when expanding the judgment criteria. Further, the linear encoder device is characterized in that the moving distance for changing the alarm determination value is one pitch or more of the upper digit scale.

Images