JP5307848B2 - Inspection device - Google Patents

Description

  The present invention relates to an inspection apparatus that compares a collected waveform with a predetermined upper and lower limit value and determines whether a product is good or bad.

  Conventionally, in order to prevent a product containing a defect from being shipped to the market, an inspection for determining the quality of a manufactured product has been performed. As one of the inspection methods, a test signal is given to the manufactured product, the behavior of the product at that time is measured through a sensor, etc., and the measured waveform acquired as a time series is compared with a predetermined upper and lower limit value. There is known a method for determining whether or not a product is good (see, for example, Non-Patent Document 1).

Mitsubishi Data Collection Analyzer MELIQC IU2 Series IU2-LOG-LIT1 User's Manual (Manual No .: JZ990D48401 Date of creation: February 2009)

  In this waveform determination method, an upper limit waveform and a lower limit waveform are set in advance, and whether or not the acquired waveform deviates from these upper limit waveform and lower limit waveform (hereinafter may be abbreviated as upper and lower limit waveforms). There is a method. This method is called band comparison (see (3) Band comparison in 3.3.6 Waveform analysis in Non-Patent Document 1).

Hereinafter, a conventional band comparison will be described with reference to FIG.
<Creation of upper and lower limit waveforms>
Assuming that a test signal is given to a product that is known to be a non-defective product, and that the behavior of the product at that time is measured at a sampling interval ΔT, the start of the test is detected as a trigger and the start time T1 is held. At the same time, sampling is started, the end of the inspection is also detected by trigger, and the end time T2 is held and the sampling is stopped. As a result, the time required for one cycle is T (= T2-T1) and the number of waveform points is n points. Suppose that Since the time required for inspection T is constant every time, waveform acquisition is started with a start trigger of one cycle, and after that, when the number of waveform points reaches n points, the same thing can be said. is there. In this way, a normal waveform is obtained. Similarly, a test signal is given to a product that is known to be a defective product, waveform acquisition is performed, and an abnormal waveform is obtained.

  In this way, several normal and abnormal waveforms are acquired, and upper and lower limit waveforms are defined based on these acquired normal and abnormal waveforms. For example, upper and lower limit waveforms are created by shifting a waveform obtained by averaging several acquired normal waveforms in the vertical direction. It is confirmed that any of the acquired abnormal waveforms deviate from the created upper and lower limit waveforms, and finally defined as upper and lower limit waveforms (upper limit waveforms H1 to Hn, lower limit waveforms L1 to Ln). The specified upper and lower limit waveforms are also n points (sampling interval ΔT).

<Determination of measurement waveform>
Similarly, it is assumed that a test signal is also given to a product to be inspected, the behavior at that time is measured at a sampling interval ΔT, and measurement waveforms (V1 to Vn) are obtained as a result. Whether or not the measurement waveform deviates from the upper and lower limit waveforms (hereinafter referred to as characteristic determination) is performed by comparing the value of each point of the measurement waveform with the value of each point of the upper and lower limit waveform. That is, if the value of each point of the measurement waveform is smaller than the value of each point of the upper limit waveform and larger than the value of each point of the lower limit waveform (L1 <V1 <H1, L2 <V2 <H2,..., Ln < (Confirm that Vn <Hn), the measured waveform does not deviate from the upper and lower limit waveforms, and this product is determined to be non-defective.

  As a waveform determination method that is different from the above-described waveform determination method that is determined by the deviation of the upper and lower limit waveforms, a method of analyzing the waveform and extracting the feature points of the waveform and determining the quality based on the result can be considered. The process is complicated, and it takes time and effort to establish a waveform analysis process that can perform right and wrong. Therefore, in the present day of multi-variety variable-volume production, it may not be practical to establish a waveform analysis process that can correctly and correctly perform one product. On the other hand, the method based on the band comparison is effective in such an application because it is only necessary to set the upper and lower limit waveforms.

  As described above, in the conventional method of inspecting a product after manufacturing and determining the quality of the product, it is necessary to provide an inspection process after the manufacturing and processing process, and start manufacturing per product accordingly. The time required from completion to completion also becomes longer. In order to solve this, by measuring the control commands and manufacturing process status during the manufacturing process, and confirming that the resulting waveform does not deviate from the waveform during normal manufacturing process There is a method for judging the quality of a product. This method is sometimes called in-line inspection or on-machine inspection (on-machine verification).

  This waveform determination method may be the same as in FIG. That is, instead of giving a test signal to the manufactured product and measuring the behavior of the product at that time, measure the status of the manufacturing process (one cycle) of one product, and compare the obtained waveforms with the band. That's fine.

  By the way, in the manufacturing process, there is a manufacturing process that is controlled based on a rotating mechanism (that is, in synchronization with the rotating mechanism) such as a press machine. Some of these manufacturing and processing facilities require time from the start of operation (power-on) until the rotation of the rotation mechanism is stabilized, such as requiring warm-up operation. Even if the rotation of the rotation mechanism has such characteristics, there is no problem in the accuracy of the manufacturing process because the control of the manufacturing process itself is synchronized with the rotation mechanism. In other words, even if the rotation of the rotating mechanism has such characteristics, the manufacturing process control itself is synchronized with the rotating mechanism so that there is no problem in the manufacturing process accuracy itself. Yes.

  However, if an on-machine inspection using a conventional waveform determination method is performed on a manufacturing process having such characteristics, the actual product is acceptable, but the waveform determination is rejected. was there.

  For example, it takes time from the start of operation until the rotation of the rotation mechanism is stabilized, and the rotation speed at the start of operation is lower (slower) than at the stable time, and the rotation speed gradually increases. In the case of having the characteristics, as shown in FIG. 8, one cycle of the manufacturing process at the start of operation becomes longer than the time required for one cycle of the manufacturing process at the stable time. Therefore, if the waveform acquisition is performed for a fixed number of points from the processing start trigger so that one cycle at stable time can be acquired, one cycle at the start of operation cannot be acquired, that is, the acquisition time, Even if the number of acquisition points is insufficient and the entire cycle cannot be acquired, it is determined based on whether the upper limit waveform and lower limit waveform are not deviated from the incomplete acquisition waveform. There was a problem that was not made.

  Conversely, it takes time from the start of operation until the rotation of the rotation mechanism stabilizes, and the rotation speed at the start of operation is higher (faster) than when it is stable, and the rotation speed gradually decreases. As shown in FIG. 9, one cycle of manufacturing processing at the start of operation is shorter than the time required for one cycle of manufacturing processing when stable. Therefore, in the setting to acquire a certain number of points from the processing start trigger at a constant sampling period so that one cycle at stable time can be acquired, not only one cycle at the start of operation but also an extra waveform are acquired. In spite of this, there is a problem that correct waveform determination cannot be performed because the determination is made based on whether the acquired waveform deviates from the upper limit waveform and the lower limit waveform.

  In order to solve this problem, if the waveform acquisition is performed from the processing start trigger to the processing end trigger at a constant sampling period, the number of obtained waveforms is different, and the same waveform determination method cannot be applied. Furthermore, as a method for solving this problem, an upper and lower limit waveform for one cycle when manufacturing and processing at a plurality of different rotation speeds according to the measured rotation speed is prepared in advance. A method of measuring the number of rotations (rotation period) and determining the waveform with the upper and lower limit waveforms corresponding to the measured number of rotations is conceivable. It is necessary to prepare a large number of waveforms, and there is a problem that it is not practical because there is a limit to the area in which a large number of upper and lower limit waveforms are stored at the same time.

  The present invention has been made to solve the above problems.

The inspection apparatus according to the present invention includes a signal input unit and a storage unit, and acquires a waveform obtained by the signal input unit for one cycle of a manufacturing process in which the same state change is repeated at a constant sampling interval. A comparison processing unit that includes an acquisition processing unit and compares the measured waveform acquired by the waveform acquisition processing unit with an upper limit waveform and a lower limit waveform for a quality determination criterion stored in the storage unit, and determines whether the product is good or bad in the testing apparatus comprising, based on the time required for one cycle of the measurement waveform, provided with a scaling processing unit for enlarging or reducing process the upper waveform and lower waveforms for the quality criterion with respect to the time axis, the The upper limit waveform and the lower limit waveform for pass / fail judgment criteria are stored in the form of one k-order polynomial or a combination of two or more k-order polynomials, and the reference waveform expansion or The small processing is performed by changing the coefficient of the k-th order polynomial based on the ratio of the time required for one cycle of the measurement waveform and the time required for the upper and lower waveforms for the pass / fail judgment criterion. The upper limit waveform and the lower limit waveform for the pass / fail judgment criterion represented by a polynomial are enlarged or reduced, and the comparison processing is performed by comparing each point of the point sequence of the measurement waveform with the polynomial to determine characteristics. .

  Even if it takes time from the start of manufacturing until the manufacturing and processing apparatus is stabilized, it is possible to correctly determine the quality of the product.

It is a block diagram which shows the hardware constitutions of the inspection apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the relationship between the manufacturing processing apparatus at the time of implementing this invention, and an inspection apparatus. FIG. 3 is a block diagram for explaining functions of the inspection apparatus according to the first embodiment. It is a figure explaining the method of the expansion and waveform determination of the upper limit waveform and lower limit waveform of this invention. It is a flowchart explaining the expansion of the upper limit waveform and the lower limit waveform of the present invention. It is a figure explaining the expansion method of the upper limit waveform and lower limit waveform which concern on Embodiment 2 of this invention. It is explanatory drawing about preparation of the conventional upper limit waveform and a lower limit waveform, and waveform determination. It is a figure which shows the acquisition waveform in the manufacturing process of each 1 cycle at the time of the rotation unstable of a rotary machine, and the stable time. It is a figure which shows the acquisition waveform in the manufacturing process of each 1 cycle at the time of the rotation unstable of a rotary machine, and the stable time.

Embodiment 1 FIG.
FIG. 1 shows a hardware configuration for realizing the present invention. The inspection apparatus 1 includes a microprocessor 2, a data storage memory 3, an operation input unit 4, a data storage unit 5, a signal input unit 6, a signal output unit 7, and an operation display unit 8.

  The inspection apparatus 1 can store a plurality of inspection programs in the data storage memory 3 and the data storage unit 5 that are storage units, and the microprocessor 2 performs signal input unit according to the inspection program specified by the operation input unit 4. The characteristic measurement process is performed on the signal measured in 6, and the determination result is stored in the data storage memory 3 and the data storage unit 5, the signal is output to the signal output unit 7, and the operation display unit 8 performs the determination. Display results. Some types of inspection devices do not include the operation display unit 8. Further, as described in the above-mentioned prior art, according to the inspection program, a test signal is output from the signal output unit 7 and the test signal is given to the manufactured product to be inspected, and the behavior of the product at that time is determined. There is also a method of using the signal input unit 6 to determine pass / fail.

  FIG. 2 shows the relationship between the manufacturing / processing apparatus and the inspection apparatus when the present invention is carried out. The manufacturing / processing apparatus 9 includes a control unit 10. When the control unit 10 outputs a control command 11, the manufacturing / processing state 12 is controlled and manufacturing processing is performed. For example, in the case of a press machine, a motor that is the driving force of the rotation mechanism, a control unit that controls the start / stop / rotation speed of the motor, a shaft, gear, cam, etc. that transmits the rotational motion of the motor to the object to be processed The press part is controlled by these to press the sheet metal to be manufactured and processed. A signal indicating the manufacturing process state 12 is detected by the state detection unit 13 and transmitted to the signal input unit 6 of the inspection apparatus 1.

  FIG. 3 shows a functional block configuration (software block configuration) of processing performed by the microprocessor 2 when the present invention is implemented. Hereinafter, the operation of the inspection apparatus 1 will be described with reference to FIG. The inspection device 1 measures a signal representing the manufacturing / processing state 12 detected by the state detection unit 13 of the manufacturing / processing device 9 through a signal measurement channel (CH) of the signal input unit 6. The waveform acquisition process 16 which is a processing function of the microprocessor 2 samples the signal obtained by the signal input unit 6 at a time interval ΔT and stores the measured waveform 17 in the data storage memory 3 which is a storage unit in time series. Furthermore, it is a process of transferring the data to the data storage unit 5 and storing it. The start trigger and the end trigger for measuring one cycle are detected by connecting the output from the control unit 10 of the manufacturing apparatus 9 to the trigger detection channel (CH) of the signal input unit 6. As described above, when the microprocessor 2 detects the start trigger in the trigger detection channel of the signal input unit 6, the microprocessor 2 samples the signal measurement channel of the signal input unit 6 at the time interval ΔT, and calculates the value in time series as the data storage unit 5. The waveform acquisition process 16 is started, and when the end trigger is detected in the trigger detection channel of the signal input unit 6, the waveform acquisition process 16 ends.

  The upper limit waveform 14 and the lower limit waveform 15 for pass / fail judgment criteria are stored in advance in the data storage memory 3 or the data storage unit 5 before performing the band comparison process 20 which is a characteristic determination process, and are obtained by the waveform acquisition process 16. Enlargement / reduction processing 19 is performed on the measured waveform 17 for the score m18. Thereafter, a band comparison process 20 that is characteristic determination is performed. That is, when the microprocessor 2 completes the waveform acquisition process 16, the microprocessor 2 reads the upper limit waveform 14 and the lower limit waveform 15 stored in the data processing memory 3 or the data storage unit 5, and these and the waveform acquisition measurement waveform 17. The band comparison process 20 is performed. Then, the determination result 21 is stored in the data storage memory 3 or the data storage unit 5 that is a storage unit, output to the signal output unit 7, or supplied to the operation display unit 8. The creation of upper and lower limit waveforms and the determination of the measurement waveform, which are band comparison methods, are as described in the background art.

<Creation of upper and lower limit waveforms>
First, creation of the upper limit waveform 14 and the lower limit waveform 15 for pass / fail judgment criteria will be described with reference to FIG. Assuming that the waveform of one cycle is measured at the sampling interval ΔT in the stable state of the manufacturing process, the start of one cycle is detected as a trigger, the start time T1 is held, sampling is started, and the end of one cycle is also detected as a trigger. Assume that the end time T2 is held and sampling is stopped. As a result, the time required for one cycle is T (= T2−T1) and the number of waveform points is n. Since the required time T is constant every cycle in the stable state of manufacturing and processing, the waveform acquisition is started by the start trigger of one cycle, and then the waveform acquisition is ended when the number of waveform points reaches n points. That is.

  If the product manufactured and processed at that time is separately inspected, and the product is a non-defective product, the waveform measured during the manufacturing process is a normal waveform. If the product is defective, the product is processed during the manufacturing process. The measured waveform is an abnormal waveform. In this way, several normal and abnormal waveforms are acquired, and upper and lower limit waveforms are defined based on these acquired normal and abnormal waveforms. For example, upper and lower limit waveforms are created by shifting a waveform obtained by averaging several acquired normal waveforms in the vertical direction. It is confirmed that any of the acquired abnormal waveforms deviate from the created upper and lower limit waveforms, and finally defined as upper and lower limit waveforms (upper limit waveforms H1 to Hn, lower limit waveforms L1 to Ln). The specified upper and lower limit waveforms are also n points (sampling interval ΔT).

  The above is the same as the conventional method for defining the upper and lower limit waveforms of band comparison in on-board inspection. The following waveform determination method is a part of the present invention which is not present. The waveform determination method according to the present invention will be described below with reference to the drawings.

<Determination of measurement waveform>
In FIG. 4, assuming that the waveform of one cycle during the manufacturing process that has not reached the stable state is measured at the sampling interval ΔT, the start of one cycle is detected as a trigger, the start time Ta is held, sampling is started, and one cycle is started. The end of the signal is also detected as a trigger to hold the end time Tb and stop sampling. As a result, the time required for one cycle is Tc (= Tb-Ta) and the number of waveform points is m points (F1 to Fm). .

  In the present invention, the upper limit waveform 14 and the lower limit waveform 15 at the n point are expanded or reduced to the upper limit waveform and the lower limit waveform at the m point in accordance with the m point 18 (the required time Tc of one cycle) of the acquired measurement waveform 17. Process 19 is performed. That is, the upper limit waveform 14 and the lower limit waveform 15 of the required time T in one cycle are enlarged or reduced to the upper limit waveform and the lower limit waveform of Tc. Thereafter, the upper limit waveform and the lower limit waveform expanded or reduced and the measurement waveform 17 are subjected to band comparison processing 20. In the following, enlargement or reduction may be abbreviated as enlargement / reduction.

  4 and 5 show the enlargement process of the upper / lower waveform 14 for pass / fail judgment criteria when n <m. 4 and 5, a linear interpolation method is described as a representative example of a specific method for enlarging the upper limit waveform 14. The upper limit waveform reduction processing method when n> m is the same as the enlargement procedure, and the lower limit waveform enlargement or reduction processing method is the same as the upper limit waveform enlargement or reduction processing method. Therefore, detailed description is omitted.

  4 and 5, first, the first point H1 (Hi) of the upper limit waveform 14 is arranged as it is as H1 ′ (Hi ′), and H1 ′ is the first point K1 (Kj) of the upper limit waveform after enlargement correction. (Step 502). Next, assuming that i = i + 1 and j = j + 1 (step 503), a point following H1 (Hi-1) of the upper limit waveform, that is, a point H2 (Hi) measured after ΔT from H1, is obtained from H1 ′ to ΔT × TC / Place it as a point H2 ′ (Hi ′) after T (step 504). If the point after ΔT can be calculated from the first point K1 (Kj−1) of the upper limit waveform after the enlargement correction on the straight line connecting H1 ′ (Hi−1 ′) and H2 ′ (Hi ′), the calculation is performed. (Step 505). If not, the process proceeds to step 508, where it is determined whether Hi ′ is the last point of the upper limit waveform (step 508). If it is the last, the process ends. If not, i = i + 1 is set (step 509). Repeat the process. Since only one point can be calculated here, the point is set as a point K2 (Kj) following the first point K1 of the upper limit waveform after enlargement correction (step 506), and j = j + 1 is set for the next (step 507).

  Next, the third point following H1 of the upper limit waveform 14, that is, the point measured after ΔT × 2 from H1, that is, the point H3 following H2, is the point after H1 ′ after ΔT × 2 × TC / T, that is, H2 ′. Is set as a point H3 ′ after ΔT × TC / T (step 507). Then, on the straight line connecting H2 'and H3', if the point after ΔT can be calculated from the already corrected upper limit waveform point K2 after enlargement correction. Since two points can be calculated here, the points are set as a point K3 following the point K2 of the upper limit waveform after enlargement correction, and a point K4 following K3 (step 505).

  By repeating the above method, H1 to Hn of the upper limit waveform 14 at the n point are expanded to the upper limit waveforms K1 to Km at the n point according to the m point (the required time Tc of one cycle) of the measurement waveform 17. (The upper limit waveform of the required time T of one cycle can be expanded to the upper limit waveform of Tc). Similarly, L1 to Ln of the lower limit waveform 15 can be enlarged to lower limit waveforms J1 to Jm (not shown) after the enlargement correction. Therefore, band comparison is performed to determine whether the value of each point of the measurement waveform 17 is smaller than the value of each point of the upper limit waveform after enlargement correction, and whether the value of each point of the lower limit waveform after enlargement correction is larger. By checking with the unit 20, the characteristics of the measurement waveform 17 can be determined, and the determination result 21 can be obtained. That is, if it can be confirmed that J1 <F1 <K1, J2 <F2 <K2,..., Jm <Fm <Km, the measured waveforms F1 to Fm deviate from the upper limit waveform after enlargement correction and the lower limit waveform after enlargement correction. This product is determined to be non-defective.

Embodiment 2. FIG.
4 and 5, the upper and lower limit waveform n points are enlarged / reduced according to the measured waveform 17 Ta, Tb, and the number of points m (required time Tc of one cycle), but the m points are generated. For this purpose, it is necessary to complete the interpolation process for calculating the point sequence for m points based on the point sequence for n points by the start of the next one cycle. Therefore, especially for on-machine inspections with short tact manufacturing and on-machine inspections with a large number of sampling points, the time until the start of the next cycle is short, and there is sufficient computing power to complete these processes. Computer is required.

  If such a problem is a concern, the upper and lower limit waveforms are modeled by a spline curve (approximate by a spline curve) and expressed by a mathematical expression, and the number of points n → m (required time T → Tc for one cycle) is expanded or reduced. Depending on the rate, the mathematical formula is scaled in the time direction, and the upper and lower limit waveforms after scaling are calculated by calculating whether m points of the measured waveform are above or below the scaled mathematical formula. What is necessary is just to use the method of discriminating whether it is above or below. Even if the mathematical modeling process of upper and lower limit waveforms takes some time, it can be done offline, that is, before starting manufacturing and before on-machine inspection. The amount of calculation is much smaller than the linear interpolation.

  Modeling the upper and lower limit waveforms with a single mathematical formula is inappropriate due to the complexity of the mathematical formula, etc. Therefore, a plurality of mathematical formulas can be selected depending on the structure of the upper and lower limit waveforms (the number of local maxima and inflection points). Divided into two models. Examples of mathematical formulas are spline curves, Bezier curves, B-Spline curves expressed by quadratic, cubic, or higher order polynomials, quadratic, cubic, or higher order rational polynomials. There is a NURBS curve represented by

  For example, FIG. 6 shows a case where the upper limit waveform 14 of FIG. 4 is modeled by dividing it by four quadratic spline curves. Here, modeling is performed by dividing the inflection point, that is, the boundary between the convex shape and the concave shape in the upper limit waveform 14 point sequence H1 to Hn. An example is shown. The first concave spline curve Hx (p) approximates the first concave H1 to Hx, and the second convex spx curve Hy to the second convex Hx to Hy. (P) is approximated, and the next concave part can be ignored as a very small part in the vicinity of Hy, and the next convex part Hy to Hz is the third. A quadratic spline curve Hz (p) is approximated, and the last concave shape from Hz to Hn is approximated by a fourth quadratic spline curve Hw (p). In addition, p shows time.

  To explain in detail by taking the first quadratic spline curve Hx (p) as an example, Ax, Bx, and Cx in Formula 1 are determined so that the error from each point of H1 to Hx is minimized. As a representative example of a parameter calculation method that minimizes such an error, a least square method has been conventionally known.

  Hx (p) models a part of the upper limit waveform based on the required time T of one cycle in the stable state of the manufacturing process. This Hx (p) is expanded to Hx (t) in accordance with the required time Tc of one cycle of the measurement waveform acquired during the manufacturing process that has not reached the stable state. Since the time required for one cycle is increased uniformly from T to Tc, the time p for manufacturing and processing in a stable state and the time t for manufacturing and processing in a state where the stable state has not been reached It can be expressed by a relational expression.

  By substituting this into Hx (p), Equation 3 is derived, and this Hx (t) is taken as an enlarged secondary spline curve. That is, a coefficient obtained by multiplying the coefficient Ax of the second-order term of the original mathematical formula before enlargement by (Tc × Tc) / (T × T) is used as the coefficient of the second-order term of the mathematical formula after enlargement. A coefficient obtained by multiplying the coefficient Bx of the first-order term of the original formula by Tc / T may be used as the coefficient of the first-order term of the enlarged formula. Note that the same is true if m / n is used instead of Tc / T.

  Similarly, when modeling with a cubic spline curve, a coefficient obtained by multiplying the coefficient of the cubic term of the original mathematical formula before enlargement by (Tc × Tc × Tc) / (T × T × T) is used. A coefficient obtained by multiplying the coefficient of the second-order term of the original mathematical formula before enlargement by (Tc × Tc) / (T × T) is used as the coefficient of the third-order term of the mathematical formula after enlargement. The coefficient of the second-order term, and the coefficient of the first-order term of the original mathematical formula before expansion multiplied by Tc / T may be the coefficient of the first-order term of the mathematical formula after enlargement.

  After that, by calculating whether m points of the measured waveform are above or below the scaled formula, it is determined whether it is above or below the scaled waveform. do it. For example, since the time (elapsed time from the start of one cycle) of the measured waveform point Fj is Tj, Ly (Tj) <Fj <Hy (Tj) using the corresponding expression Hy (t). ).

  Note that the method of comparing the bands by reducing the m points of the measurement waveform to the same n points as the upper and lower limit waveforms for pass / fail judgment criteria is not preferable from the viewpoint of inspection accuracy. This is because reducing the number of points in a measurement waveform means performing a kind of averaging process on the measurement waveform, so that abnormal points that must be detected are also reduced by this reduction process (averaging process). This is because they may be buried and may not be detected.

1 inspection device, 2 microprocessor,
3 data storage memory, 4 operation input section,
5 Data storage unit, 6 Signal input unit,
7 signal output section, 8 operation display section,
9 Manufacturing processing equipment, 10 Control part,
11 Control command, 12 Manufacturing process state,
13 state detector, 14 upper limit waveform,
15 lower limit waveform, 16 waveform acquisition processing,
17 Measurement waveform, 18 points m,
19 Enlargement / reduction processing, 20 Band comparison processing,
21 Judgment result.

Claims (4)

A waveform input processing unit that includes a signal input unit and a storage unit, and acquires a signal of one cycle of a manufacturing process that is obtained by the signal input unit and repeats the same state change at a constant sampling interval; In the inspection apparatus including the comparison processing unit that determines the quality of the product by comparing the measurement waveform acquired by the acquisition processing unit with the upper limit waveform and the lower limit waveform for the quality determination criterion stored in the storage unit, And an enlargement / reduction processing unit for enlarging or reducing the upper / lower waveform for the pass / fail judgment criterion with respect to the time axis based on a time required for one cycle of the waveform, and an upper limit waveform for the pass / fail judgment reference, The lower limit waveform is stored in the form of one k-th order polynomial or a combination of two or more k-th order polynomials, and the reference waveform enlargement or reduction process is performed on the measured waveform. The pass / fail judgment criterion represented by the k-th order polynomial is changed by changing the coefficient of the k-th order polynomial based on the ratio of the required time of the cycle and the time required for the upper / lower waveform for the pass / fail judgment criterion. An inspection apparatus characterized in that the upper limit waveform and the lower limit waveform are enlarged or reduced, and the comparison process performs characteristic determination by comparing each point of the point sequence of the measurement waveform with the polynomial . A waveform input processing unit that includes a signal input unit and a storage unit, and acquires a signal of one cycle of a manufacturing process that is obtained by the signal input unit and repeats the same state change at a constant sampling interval; In the inspection apparatus including the comparison processing unit that determines the quality of the product by comparing the measurement waveform acquired by the acquisition processing unit with the upper limit waveform and the lower limit waveform for the quality determination criterion stored in the storage unit, And an enlargement / reduction processing unit for enlarging or reducing the upper / lower waveform for the pass / fail judgment criterion with respect to the time axis based on a time required for one cycle of the waveform, and an upper limit waveform for the pass / fail judgment reference, The lower limit waveform is stored in the form of one k-th order polynomial or a combination of two or more k-th order polynomials, and the reference waveform enlargement or reduction process is performed on the measured waveform. By changing the coefficient of the k-th order polynomial based on the ratio of the number m and the point n of the upper and lower waveform for the quality judgment criterion, the quality judgment criterion for the quality judgment criterion represented by the k-th order polynomial is changed. An inspection apparatus characterized in that an upper limit waveform and a lower limit waveform are enlarged or reduced, and the comparison process performs characteristic determination by comparing each point of the sequence of points of the measurement waveform with the polynomial. The one cycle is one cycle of a manufacturing process state in which the same control is repeated, and the manufacturing process is controlled based on a mechanism having a cycle, and the period of the mechanism at the start of the manufacturing process, The inspection apparatus according to claim 1 or 2 , wherein a period of the mechanism after a lapse of time from the start of the manufacturing process is a non-constant manufacturing process. The inspection apparatus according to claim 3 , wherein the mechanism is a rotation mechanism.

Images