JPH0122884B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an improvement in a data reading device for a storage effect sensor, typically a charge coupled device (CCD).

Generally, in order to remotely measure the travel distance or rotation angle of a moving object from a reference point, attach a scale to the object, reset the scale's graduations to zero at the reference position, and then measure the movement from there. The distance and rotation angle may be counted optically or electromagnetically and displayed on a display, for example.

However, since the method described above is a so-called incremental method in which the amount of relative movement between the scale and the sensor that reads it is measured using an electrical signal generated during movement, a reference point must be detected.

Therefore, in order to eliminate such problems,
An absolute method that encodes the scale can be considered. In this method, the length from the reference point of the scale is encoded and marked on the scale, and by reading this, the length from the reference point is obtained without detecting the reference point. be. In carrying out this method, it is conceivable to use a cumulative effect sensor to read the encoded scale and obtain subdivided values. A storage effect sensor is composed of a photo element section, a gate section, and a shift register section, as typified by a charge coupled device (CCD). The photo element section is constructed by arranging a large number of elements in, for example, a straight line, and each element has a function of accumulating a charge corresponding to the intensity of light incident thereon. The shift register section has the same number of bits as the number of elements in the photo element section, and the charge of the corresponding element is transferred to each bit in parallel via the gate section, and when a scanning pulse is applied, each bit is transferred in parallel. The configuration is such that the charges are output in time series. In such a storage effect type sensor, the width of each element constituting the photo element section can usually be set on the order of 10-odd microns, making it possible to read the scale graduations and interpolate between the graduations. That is, as shown in FIG. 1, a bright area A (light is transmitted) and a dark area B (light is not transmitted).
A storage effect sensor, for example CCDY, consisting of a photo element part F, a gate part G, and a shift register part R, is provided facing one side of the scale C, which is formed by arranging the elements alternately in a straight line, and the other side of the scale C. When parallel light P having a width X is irradiated toward the photoelement portion F, charges corresponding to the incident light are accumulated in the element in the portion of the photo element portion F where the light that has passed through the bright portion A is incident. Here, if a gate pulse S is applied to CCDY, a reset pulse Z, and scanning pulses φ ₁ and φ ₂ which are obtained by dividing this pulse Z by 1/2 and having a phase difference of 180° are applied, a differential amplifier is generated. H outputs the output CO shown in FIG. 2d. In this example, the width of the bright area A is set to the width of three elements constituting the photo element section F, and the three elements directly face the bright area A. , three signals are sent out in chronological order exactly corresponding to the above-mentioned correspondence relationship. Therefore, by combining a scale that encodes the length from a reference point with a light-dark pattern and an accumulation effect sensor, it is possible to measure the distance from the reference point of the scale, that is, the length, with an accuracy of less than 10 microns. It becomes possible to do so. Specifically, it is conceivable to provide both as shown in FIG. 3, for example. That is, the scale C is divided into a plurality of blocks of the same length in the longitudinal direction. This figure shows the Npth block and the Np _-1th block. Taking each block, for example, block Np as an example, the width of one element constituting the photo element part F of CCDY is set as the reference width at the left end of this block Np, and the four of the above elements are set as a reference width.
A marker M having a width of
Furthermore, at the right end of block Np _-1 that is in contact with it,
Form a block stop marker D consisting of the dark area of four elements of CCDY, and measure the width of one element in contact with this among the four elements of marker M.
Im. Then, on the right side of the marker M, there is a dark area with a width of two of the above elements and every two widths.
_{K 1 , K 2} , ^K _{3 .} , P ₂ to 2 ¹ , P ₃ to 2 ²
, and the following are similarly made to correspond to binary codes. In addition,
In the figure, address parts P ₁ and P ₃ are formed in the bright part, and the other address parts are formed in the dark part. block
The same setting is made for block Np _-1 , but the value of the binary encoded address field is subtracted by 1 from that for block Np.

Therefore, the scale C configured as above
To accurately measure the distance from the reference point of scale C using and CCDY, do as follows.

That is, each element constituting the photo element section F of CCDY is assigned a bit number as shown in the figure, and a certain bit is designated as an index bit In.
shall be. Then, the width X is approximately equal to the width of 2 blocks.
The scale C at this time is irradiated with parallel light of
Check the projection status to CCDY. The projected position of the marker bit Im is determined from the marker M and the block stop marker D. In Fig. 3, the position of ◎ in the CCDY photo element section corresponds to the marker bit Im. Since the positional relationship between the marker bit Im and the address parts P ₁ , P ₂ , P ₃ ...... is predetermined as described above, the position of the address part projected from the bit ◎ corresponding to the marker bit Im to the photo element part I understand. That is,
First, the state of light reception of bits, bits, bits, bits... is detected. In this example, the bit is so-called ON, so ²⁰ digits are ON. Also the bit is OFF
So, ²¹ digit is OFF and bit is ON as well.
So, the 2nd ^2nd digit is ON, the bit is OFF, and the 2nd ^3rd digit is
It is OFF. Therefore, the absolute address of block Np is 2 ² +2 ⁰ =5, and it can be seen that block Np is the fifth one from the reference point of the scale. Now, if one block is made up of 26 bits as shown in the figure, and the width of one bit is set to 15μ, the width of one block will be 15μ×26=0.39m/m. Therefore,
Block Np is approximately 0.39m/m x 5 from the reference point
= 1.95m/m apart. Next, to find a more accurate distance, count the number of bits from index bit In to the bit opposite marker bit Im. In this example, the number of bits is 5, so the accurate distance is 1.95 m/m - 15μ x 5 = 1.875 m/m.

That is, if the number of bits from index bit In to marker bit Im is Nc, the block address is ACD, the length of one block is Ns which is the number of elements in photo element section F, and the width of one element is mμ. The measured value E is E=m(Ns・ACD−Nc) (unit: μ). Note that the above value is obtained when measuring the address of the block where the marker exists on the right side of the index bit In in Fig. 3, but when measuring the address of the block on the left side, E=m(Ns・ACD+Nc ) is measured as In addition, in order to improve the measurement accuracy to within the width of one bit of the photo element part, a vernier part is added between the marker part and the address part in one block on the scale side, and this vernier part is When the number of bits is N _B , the number of elements in the photo element part within the same length as the vernier part is set in the relationship N _B +1, and the overlapping state of the bits in the vernier part on the scale side and the bits in the photo element part is calculated. This can be done by checking (for example, finding the position of the directly opposite bit).

In this way, by combining a scale with a light-dark pattern and an accumulation effect sensor, it becomes possible to measure distance, that is, length, with high accuracy without detecting a reference point. This can also be directly applied to angle measurements.

However, when trying to actually measure lengths and angles using cumulative effect sensors as mentioned above,
The following problems arose, and there was a drawback that stable measurements could not be carried out over a long period of time.

That is, when the scale shown in FIG. 3 is used to detect the information of block Np, the output C read from the cumulative effect sensor is normally as shown in FIG. 4a. This example is a case where the brightness of the light source and the light transmittance of the bright part on the scale are within the specified range. If the brightness of the light source decreases for some reason, the amplitude of the output signal C will change as shown in Figure 4b, and the amplitude of the output signal will change as shown in Figure 4a.
It is lower than in the case of . Furthermore, if dust adheres to the marker M, for example, the light transmittance of this marker decreases, so that in the output C at that time, only the amplitude of the marker signal decreases, as shown in FIG. 4c. As described above, if the amplitude of the output signal changes due to variations in the brightness of the light source, adhesion of dust, etc., it is inevitable that the system that processes the read signal will malfunction.
In particular, when performing signal processing as described above by combining scale C and CCDY, which have bright and dark areas as shown in Fig. 3, marker bits are Since it is necessary to detect the number of bits up to 100 bits based on the output signal level of each bit, accurate measurement cannot be expected over a long period of time if the output signal level fluctuates as described above.

In reality, the change in the signal read from the cumulative effect sensor changes in response to the reference level V _p which changes with a certain slope as shown in FIG. 4d. At the beginning of reading from the shift register section R, the reference level V _p already drifts by V _a from the ideal O level, for example, and further drifts by V _b from V _a at the end of reading. Such fluctuations in the reference level V _p are caused by the dark current flowing in the photo element section F and the shift register section R, and V _a corresponds to the charge due to the accumulation of dark current in the photo element section F. Further, V _b corresponds to the charge due to accumulation of dark current in the shift register section R. Then, V _a is proportional to the light reception period, and V _b
is proportional to the readout period. Therefore, in order to eliminate fluctuations in the reference level V _p , for example, shorten the light reception period or readout period, or apply optical information in pulses to change the output when there is optical information to the output when there is no optical information. ” can be reduced to a negligible level by reducing the output. However, even if such means are adopted, the aforementioned variations in output level due to variations in brightness of the light source, adhesion of dust, etc. will still remain.

The present invention was made in view of the above circumstances, and its purpose is to always provide information even when the intensity of optical information fluctuates due to fluctuations in the brightness of the light source or due to adhesion of dust, etc. To provide a data reading device for a cumulative effect sensor that can read a data signal of a constant amplitude from a cumulative effect sensor, and can contribute to stable measurement over a long period of time when applied, for example, to length measurement or angle measurement. It is in.

The present invention provides optical information including correction optical information to a storage effect sensor, compares the data signal level of the correction optical information read from the sensor with a predetermined level, and based on the comparison result, By correcting the data signal level of true optical information, the data level of true optical information is made constant, and the above object is achieved.

Hereinafter, details of the present invention will be explained with reference to illustrated embodiments. Note that the figure shows an example in which the present invention is applied to a length measuring device.

In FIG. 5, numeral 1 is a linear scale attached to a moving body (not shown), and this scale 1 has n pieces of the same length in the longitudinal direction, similar to the one shown in FIG. 3, for example. It is divided into blocks. Each block is marked with a marker M in the form of a light and dark pattern similar to that shown in Figure 3.
and address parts P ₁ , P _{2 . .} . are formed.
However, in this example, as shown in FIG. 6, a dark area L with the width of six elements constituting the photo element section F of the storage effect sensor 4, which will be described later, is left on the left side of the marker M, and the above element is A correction bright part _W1 having a width of eight parts is formed. Similarly, a bright correction area _{W 2} with a width of eight elements is formed between the address area P 2 and the address area P ₃ via a dark area with a width of two elements. In addition, Fig. 6 shows one block as an element.
This is an example of 54 pieces.

On one side of the scale 1, a light source 2 formed of, for example, a light emitting diode is arranged, and the light emitted from this light source 2 is converted into parallel light by a lens system 3, and then , scale 1
is irradiated towards. Further, on the other side of the scale 1, a storage effect sensor, for example, a CCD 4, is arranged to detect light transmitted through the scale 1 via a lens system 3'. This CCD 4 is composed of a photo element section F, a gate section G, and a shift register section R, each of which is formed of a plurality of elements as described above. When the reset pulse Z and scanning pulses φ ₁ and φ ₂ whose frequency is divided by 1/2 and whose phases differ by 180° are applied, the contents of the shift register R are transferred to the shift register R. It is configured to output in chronological order. The drive of the CCD 4 is controlled by the microcomputer 6 via the pulse control circuit 5, and the output C of the CCD 4 is controlled by the conversion circuit 7.
After being processed by the microcomputer 6 via the microcomputer 6, it is displayed on the display 8. Note that the light source 2, lens systems 3, 3', and CCD 4 are incorporated into a housing 10 and fixed to a stationary part so as to sandwich the scale 1 in a non-contact manner.

The pulse control circuit 5 is configured as follows. That is, the pulse P ₂ sent from the microcomputer 6 in the relationship described later is
It is given as a reset signal Z to the CCD 4 and also sent to the flip-flop circuit 1 via the inverter 14.
5 (Flip that operates at the rising edge of the clock signal)
It is applied to clock terminal C of . The flip-flop circuit 15 is a T-flip-flop whose output is introduced into the input terminal D, and the output is
The CCD 4 is given a scanning pulse φ ₁ and the Q output is also given as a scanning pulse φ ₂ .

On the other hand, the conversion circuit 7 extracts the output of the CCD 4 using a differential amplifier 21, and also outputs the output C of the amplifier 21 via a gain control circuit 22 to an A/D converter (analog-digital converter) 23.
has been introduced. The gain control circuit 22 includes a seventh
As shown in the figure, an amplifier J and a feedback resistor connected between the output terminal and the inverting input terminal of this amplifier J
Loading resistors R ₁ , R ₂ , R ₃ , R ₄ are inserted in parallel between R _f and the inverting input terminal of amplifier J and ground.
, analog switches FET ₁ , FET ₂ ...... FET ₄ inserted in series with these load resistors, and amplifiers A _n1 , A _n2 , which control these analog switches.
It consists of an operational amplifier consisting of A _n4 . Then, the output C of the differential amplifier 21 is introduced into the non-inverting input terminal of the amplifier J, and each amplifier
A logic level “1” is sent from the microcomputer 6 to the input terminals of A _n1 , A _n2 , …A _n4 as a correction value.
Alternatively, a "0" signal, that is, a binary signal indicating a value of four digits below the decimal point, is given. Note that the load resistances are, for example, R ₁ = 2R _f , R ₂ = 4R _f , R ₃ = 8R _f ,
R ₄ = 16R ₀ , and the analog switch
By selectively turning on FET ₁ to FET ₄ , the gain G (C'/C) can be set to 1/1 in the range of 1 to 1+15/16.
It is set so that it can be changed in increments of 16.

The A/D converter 23 is in a clear state when the A/D control AC of the microcomputer 6 is "1" and performs A/D conversion operation when it is "0". Also, A/
When the D conversion operation is completed, the terminal AE of the microcomputer 6 which was "1" during the operation is changed to "0".

Thus, the apparatus of the present invention configured as described above operates according to a program set in the microcomputer 6. Below, this operation will be explained in the 8th section.
This will be explained using the flow chart shown in the figure.

First, when starting from the routine entrance A, f ₁ : Clears the correction value memory M _CDA to _{M (CDA)+X} .
This is for reading the first data without correction, and clears the correction value memory M _CDA to _{M (CDA)+X} .

f ₂ : Put the first address CDA of the correction value memory M _CDA to _{M (CDA)+X} into the memory _M6 .

_f3 : Put the first address SDA of data memory M _SDA to _{M (SDA)+X} into memory _M5 .

f ₄ : Generate one pulse P _s (“0” → “1”
→“0”). As a result, charges are transferred from the photo element section F to the register section R of the CCD 4.

f ₅ : Put the contents of memory M ₆ into the address register.

f ₆ : Set the contents of the memory indicated by the address register to the correction value CD. At _f5 and _f6 , the necessary correction data from the correction value memory M _CDA to _{M (CDA)+X} is set to the correction value CD, and in this case, the operation starts from the routine entrance A. Therefore, the correction value CD is set to "0000". Therefore, the gain G of the gain control circuit 22 is set to 1.

f ₇ : Put the contents of memory _M5 into the address register.

_f8 : Set A/D control AC to "0". As a result, the A/D converter 23 starts an A/D conversion operation.

_f9 : Wait until A/D conversion end AE becomes "0".

_f10 : Write A/D output data D to the memory indicated by the address register.

_f11 : Set A/D control AC to "1". This puts the A/D converter 23 in a clear state.

f ₁₂ : Generate one pulse P ₂ (“0” → “1”)
→“0”). As a result, the register portion R of the CCD 4 is shifted by one bit.

f ₁₃ : Add 1 to the contents of memory _M6 .

_f14 : Add 1 to the contents of memory _M5 .

f ₁₅ : Whether the contents of memory _M5 are X+ (SDA),
In other words, it is checked whether the necessary X pieces of data have been collected. Therefore, by making X correspond to the number of bits of the shift register, "YES" will be obtained when all the bits of the register R are taken. If NO, move to _f5 . The register section R is scanned by repeating f ₅ to f ₁₅ .

f ₁₆ : Detect bright area data for correction from the data written in the memory, and put this data into _ME in the memory area. Also, set the address of the memory containing the correction value (the correction value used when acquiring the data) corresponding to this data in the memory area.
Put it in M _F. Note that data detection of the correction bright area is performed by checking whether the contents of seven memories sequentially adjacent to each other starting from M _SDA in the memory areas M _SDA to M _(SDA)+X are substantially the same.
That is, if there are seven or more memories with almost the same contents, it is determined that they are bright areas for correction. Then, the last data (seventh data)
M _E(1) , and the memory address containing the corresponding correction value into M _F(1) . In this way, check from M _SDA to M _(SDA)+X . Therefore, the bright area data for correction is M _E(1) , M _E(2) ,
M _E(3) , ...... are entered, and the memory addresses containing the corresponding correction values are respectively
Enter M _F(1) , M _F(2) , M _F(3) , ......

f ₁₇ : Check the maximum value and minimum value of the correction bright area data stored in the memory area M _E to determine whether the difference between the two is within a predetermined tolerance. If it is within the allowable value, it is determined that the influence of uneven light has been corrected to an allowable range (in this case, since the operation is actually started from the routine entrance A, no correction is made);
Move to the routine necessary for the next calculation process. Further, if the value is not within the allowable value, it is determined that the correction is incomplete due to a large influence due to the unevenness of light, and the process moves to _f18 .

f ₁₈ : When operating from routine entrance A as described above, the contents of the memories M _CDA to _{M (CDA) + X} are cleared in f ₁ , so the data of the bright part for correction is belongs to. The processing at _f18 is performed at routine entrance B or will be described later.
This is for when operating through f _21. Calculate the corrected bright area data (uncorrected data) when the gain G = 1 from the previously used correction value and correction bright area data using the following formula. .

Contents of M _E( 〓 ₎ /1+(Correction value memory M _CDA ~ _{M (CDA)+X}
(Contents of those indicated by M _{F (} 〓 ₎ ) However, α is 1, 2, 3, ......
The contents of M _CDA to _{M (CDA)+X} indicate the decimal point. Then, put this result into M _E( 〓 ₎ . Therefore,
The contents of M _E( 〓 ₎ are rewritten from the corrected value of the bright area for correction to the value when no correction is made.

f ₁₉ : Find the maximum value V _SY of the contents of M _E (contents of M _{E (1)} , M _{E (2)} ......).

f ₂₀ : Calculate the correction value for the bright area for correction when reading the next data using the following formula.

V _SY (maximum value obtained with f ₁₉ ) / Contents of M _{E (} 〓 ₎ Put this result into M _{E (} 〓 ₎ . Therefore,
The contents of M _E( 〓 ₎ are rewritten from the value of the uncorrected bright area for correction to the correction value for the corrected bright area when reading the next data.

f ₂₁ : The correction value for the bright area for correction when reading the next data (contents of M _E ) and the corresponding M _F (addresses of correction value memory M _CDA to _{M (CDA) +} (corresponding to the bright area) for all of the correction value memories M _CDA to _{M (CDA)+X} and rewrite the contents. In other words, M _E( 〓 ₎
Find the correction value between and M _{E (} 〓 ₊₁₎ and store it in the correction value memory.
Put it in M _CDA ~ _{M (CDA)+X} . Calculation is based on the formula below.

M _E( 〓 ₎ +M _E( 〓 ₎ −M _E( 〓 ₊₁₎ ／M _F( 〓 ₊₁₎ −M _F( 〓 ₎・β
-1 However, β=1, 2, 3...... and O≦
β<M _F( 〓 ₊₁₎ −M _F( 〓 ₎ . The value obtained using this formula is stored in a memory (one memory within the range of correction value memories M _CDA to M _(CDA)+X ) having an address shown by the following formula.

M _F( 〓 ₎ +β Once the correction value has been changed in this way, the process moves to f ₂ and the data is read again using the newly changed correction value.

If these operations are shown in waveforms, if data is read without correction, differential amplifier 2
Assuming that the outputs V _w1 , V _w2 , V _w3 , ... corresponding to the bright areas for correction are output as shown in Figures 1 to 9 a, there is no correction (G = 1) in this case. The output of the gain control circuit 22 is also exactly the same.
When data is read without correction, if the correction bright area with the maximum value V _SY is W ₃ among the correction bright areas, other correction bright areas
Correction values are determined so that the outputs of W ₁ , W ₂ , and W ₄ are the same as V _SY . Then, the parts other than the bright part for correction are approximated from the correction value of the bright part for correction, that is,
In this example, a correction value is determined by linear approximation from the correction values of adjacent bright areas for correction. Therefore, when the gain control circuit 22 corrects the outputs corresponding to the bright areas for correction, the outputs are at the same level as shown in FIG. 9b, and the data levels are also at almost the same level. As a result, it is possible to read data signals that are not affected by variations in the light source, adhesion of dust, etc., and the data signal processing system can be operated correctly.

Note that the device of the present invention is not limited to the embodiments described above, and can be modified in various ways. In other words, optical information including correction optical information is provided in a pulsed manner, and the above-mentioned correction is performed on the data obtained by subtracting the data without optical information from the data with optical information and removing the accumulated dark current. You may do so. Furthermore, it is also possible to carry out "start scanning" over all bits in the register section at a scanning frequency higher than the scanning frequency at the time of reading before starting data reading, and then perform correction on the read data. good. Moreover, although the above-mentioned embodiment is an example applied to a length measuring device, it can also be applied to an angle measuring device or a pattern recognition device. Furthermore, the configuration of the gain control circuit is not limited to that of the embodiment. Further, as an information source, the correction optical information and the marker may be shared, or one
The number of pieces of correction optical information used within a block can be set freely. Furthermore, the means for correcting the data signal level is not limited to linear approximation; for example, n
The correction value for the data signal level corresponding to the (n+1)th correction optical information may be used to correct the period until the data signal corresponding to the (n+1)th correction optical information is sent out. In determining the correction value, the correction value is not limited to the ratio to the maximum value of the data signal level corresponding to the correction optical information, but may be determined by the ratio to a predetermined level. Each may be determined separately.

As described in detail above, according to the present invention, even if all or part of the optical information changes due to changes in the brightness of the light source or the influence of dust, a data signal of a constant width is always transmitted from the storage effect sensor. can be read,
It is possible to provide a data reading device for a cumulative effect sensor that can prevent malfunctions of the system that processes the data signal.

[Brief explanation of drawings]

Figure 1 shows an example of the arrangement when a cumulative effect sensor and a scale with a bright/dark pattern are combined, and Figure 2 shows various parts when scanning with the cumulative effect sensor in the arrangement shown in Figure 1. Waveform diagram, Figure 3 is a diagram for explaining an example of the relationship between the accumulation effect sensor and the scale with a bright and dark pattern when measuring length by combining them, and Figure 4 is a diagram showing the relationship between the above scale and the above sensor. FIG. 5 is a diagram illustrating the configuration of a device according to an embodiment of the present invention, and FIG. 7 is a configuration diagram of a gain control circuit in the device, FIG. 8 is a flow diagram for explaining the operation of the device, and FIG. 9 is a diagram to explain the configuration.
The figure is a waveform diagram for explaining the operation of the device. 1... Scale, 2... Light source, 3... Lens system, 4... CCD, 5... Pulse control circuit, 6...
...Microcomputer, 7...Conversion circuit, 8...
...Indicator, 22...Gain control circuit, W ₁ , W ₃ ...
Bright area for correction.

Claims (1)

[Scope of Claims] 1. A photo element section configured with a plurality of elements that convert optical information into charges and store them, and a charge accumulated in each element of the photo element section provided corresponding to the photo element section. A storage effect sensor consisting of a register section that can simultaneously store and hold data, and a scale reading coded by a light-transmitting/light-opaque slit that is relatively movable along the storage effect sensor. a scale plate formed by one-dimensionally arranging in the movement direction a correction pattern for providing correction optical information to the storage effect sensor; and data for reading pattern data from the storage effect sensor. a sensor driving means for performing readout scanning; a correction device for correcting and outputting the level of the pattern data signal read from the storage effect type sensor; and the correction pattern data signal read from the storage effect type sensor. 1. A data reading device for a cumulative effect sensor, comprising: a correction value calculation device that calculates a correction value that allows the level of to be a constant value, and controls the correction device based on this correction value. 2. The correction value calculation device calculates the correction value according to the ratio between the maximum level or a predetermined level of the plurality of correction pattern data signals read from the cumulative effect sensor and the level of each correction pattern data signal. data of the cumulative effect type sensor according to claim 1, wherein the correction value for the level of each correction pattern data signal and the correction value for the level of the scale reading pattern data signal are calculated. reader. 3 The correction value calculation device calculates the above-mentioned level for each optical information signal read between the n-th correction pattern data signal and the n+1-th correction pattern data signal read from the storage effect sensor. 3. The storage effect sensor data reading device according to claim 2, wherein a correction value equal to a correction value for the level of the n-th correction pattern data signal is applied. 4. The correction value calculation device calculates the above-mentioned level for each optical information signal read between the n-th correction pattern data signal and the n+1-th correction pattern data signal read from the storage effect sensor. A correction value obtained by linear approximation of a change from a correction value for the level of the n-th correction pattern data signal to a correction value for the level of the n+1-th correction pattern data signal is applied. A data reading device for a cumulative effect sensor according to claim 2. 5. The data reading device according to claim 1, wherein the correction device is constituted by a gain control circuit whose gain changes according to the correction value signal.