JP6910882B2 - Inspection data generator, inspection system and inspection data generation processing program - Google Patents

Description

The present invention provides inspection data that can identify the quality determination condition of the inspection target based on the measurement value data of the good product sample of the same type as the inspection target and the measurement value data of the defective product sample of the same type as the inspection target. It is about the technology to generate.

For example, the following patent documents disclose the invention of a winding diagnostic system (hereinafter, also simply referred to as "diagnostic system") configured to be able to diagnose the quality of winding in a rotating machine such as an electric motor or a generator. There is. When diagnosing the quality of a winding by this diagnostic system, first, an impulse voltage is applied from the impulse voltage generation circuit to the winding to be diagnosed, and the voltage value between both ends of the winding is sampled at a predetermined sampling cycle ( Measure. As a result, a damped vibration waveform corresponding to the winding state is observed. In reality, the rotor is rotated by a predetermined angle and the same process is repeated, but in order to facilitate understanding of the background technology, the description of the rotation of the rotor will be omitted.

Next, the resistance of the equivalent circuit constant in the circuit composed of the winding and the impulse voltage generation circuit is R, the inductance is L, the capacitance is C, and the circuit composed of the winding and the impulse voltage generation circuit is RLC. Considering a series circuit, the multiplication values LC and RC (hereinafter, also referred to as "LC value" and "RC value") of each equivalent circuit constant in the series circuit are calculated using a pseudo inverse matrix. At this time, the reference (LC value) for starting sampling is the time when the measured value (voltage value) obtained by the above measurement process becomes the minimum, or the time when the measured value crosses zero (when the voltage value becomes 0 V). And the starting point of the measured value range for calculating the RC value), and the LC value and the RC value are calculated using a predetermined number (for example, 10 points or more) of the voltage waveform for half a cycle or more.

After that, it is diagnosed whether the winding to be inspected is a non-defective product or a defective product based on the difference between the calculated LC value and the LC value of the non-defective product and the difference between the calculated RC value and the RC value of the non-defective product.

Japanese Patent No. 5721581 (pages 8-13, Fig. 1-18)

However, the configuration of the diagnostic system disclosed in the above patent document has the following problems. Specifically, in the diagnostic system disclosed in the above patent document, an impulse voltage is applied between both ends of the winding to sample (measure) the voltage value between both ends, and the observed attenuated vibration waveform is defined in advance. LC value and RC calculated for a predetermined number (10 points or more of half a cycle or more of the voltage waveform) starting from the measured value at the time of measurement (minimum value or measured value at zero crossing) A configuration is adopted in which the quality of the winding is diagnosed based on the value.

On the other hand, when the applicant tried to inspect the quality of the winding component (winding) according to the same procedure as the procedure for quality diagnosis by the above diagnostic system, the type of winding component and the winding component were found. It has been found that it may be difficult to determine whether the winding is a good product or a defective product depending on the type of defect. In this case, in the non-defective product diagnostic procedure by the diagnostic system disclosed in the above patent document and the good / bad inspection procedure attempted by the applicant, depending on the condition of the winding component (that is, whether the winding component is a non-defective product or a defective product). It utilizes the phenomenon that the damped vibration waveform when an impulse voltage is applied (depending on the presence) is different. Therefore, the applicant observed the damped vibration waveform after applying the impulse voltage to the winding component in which various defects occurred, and found that it depends on the type of the winding component and the type of defect occurring in the winding component. Found that the degree of difference between the damped vibration waveform of a defective product and the damped vibration waveform of a good product may be very low (both waveforms have almost the same shape).

In such a case, the LC value or RC value calculated based on the measured value for the defective product and the LC value or RC value calculated based on the measured value for the non-defective product are about the same value. .. Therefore, in the configuration of the diagnostic system disclosed in the above patent document, when a slight measurement error occurs due to the influence of noise or the like during the measurement process of the winding component to be inspected, the product is defective even though it is a defective product. LC value or RC value similar to LC value or RC value calculated based on measured value for non-defective product is calculated, or LC calculated based on measured value for defective product despite being good product An LC value or RC value that is about the same as the value or RC value may be calculated. Therefore, it is very difficult to correctly determine the quality of the wound component depending on the type of the wound component and the type of defect occurring in the wound component.

The present invention has been made in view of the above problems, and is an inspection data generator and an inspection data generation process capable of providing inspection data capable of accurately determining the quality of an inspection target in which various defects may occur. The main purpose is to provide a program for inspection and an inspection system capable of performing an inspection of an inspection target based on such inspection data.

In order to achieve the above object, the inspection data generator according to claim 1 is performed Na times (Na) for a sample of a good product of the same type as the inspection target in a predetermined sampling cycle for a predetermined time. Is the same type of Na measurement value data A in which a plurality of measurement values a0 for the predetermined time are recorded by the measurement process A of (two or more natural numbers), and the inspection target and the non-defective sample. A plurality of measured values b0 for the predetermined time are recorded by the measurement process B of Nb times (Nb is a natural number of 2 or more) performed under the same measurement conditions as the measurement process A for the defective sample. Nb measurement value data B are acquired, and Nc times (Nc is a natural number of 2 or more) performed on the inspection target under the same measurement conditions as the measurement process A and the measurement process B at the time of inspection of the inspection target. ), Each of the above-mentioned inspection data can specify the quality determination condition of the inspection target based on Nc measurement value data C in which a plurality of measurement values c0 for the predetermined time are recorded. A processing unit that executes a data generation process that is generated based on the measured value data A and each measured value data B is provided, and the processing unit is a plurality of first-order differentiated versions of each measured value a0 in the data generation process. The calculated value a1 and the plurality of calculated values a2 obtained by second-order differentiating each of the measured values a0 are derived for each of the measured value data A, and the plurality of calculated values b1 obtained by first-order differentiating each measured value b0, and The first process of deriving a plurality of calculated values b2 obtained by second-order differentiation of each measured value b0 for each of the measured value data B, each measured value a0 within a predetermined measurement value range, and the measured value. Each of the calculated value a1 in the calculated value range corresponding to the range and the corresponding discriminated value A0 for pass / fail determination are set for each of the measured value data A corresponding based on the calculated value a2 in the calculated value range. The calculation is performed, and the discriminated value A1 for pass / fail determination is calculated for each of the corresponding measured value data A based on the calculated value a1 in the calculated value range, and each of the calculations in the calculated value range. The discriminated value A2 for pass / fail determination is calculated for each of the corresponding measured value data A based on the value a2, the measured value b0 within the measured value range, and the respective calculations within the calculated value range. Based on the value b1 and each of the calculated values b2 in the calculated value range, the discriminated value B0 for pass / fail determination is calculated for each of the corresponding measured value data B, and each of the above calculated values within the calculated value range is calculated. The discriminated value B1 for pass / fail determination is calculated for each of the corresponding measured value data B based on the calculated value b1, and the corresponding measured values are calculated based on the respective calculated values b2 within the calculated value range. The second process of calculating the discriminated value B2 for pass / fail determination is executed for each data B, and the degree of difference between the discriminated value A0 and the discriminated value B0, the discriminated value A1 and the subject The inspection data is generated by defining one of the quality determination conditions based on the degree of difference from the discrimination value B1 and the degree of difference between the discrimination value A2 and the discrimination value B2.

Further, the inspection data generation device according to claim 2 is the inspection data generation device according to claim 1, and the processing unit has the discriminated value A0 and the discriminated value B0 in the data generation process. When the degree of difference is higher than the degree of difference between the discriminated value A1 and the discriminated value B1 and the degree of difference between the discriminated value A2 and the discriminated value B2, the measured value data C Each measured value c0 within the measured value range, a plurality of calculated values c1 obtained by first-order differentiating each measured value c0 within the measured value range, and each measured value c0 within the measured value range being second-order differentiated. One of the good / bad judgment conditions is defined as a condition in which the good / bad judgment value C0 for good / bad judgment is calculated based on the plurality of calculated values c2 and the good / bad of the inspection target is judged based on the good / bad judgment value C0. The degree of difference between the discriminated value A1 and the discriminated value B1 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the discriminated value A2 and the discriminated value B2. When the degree of difference is higher than the degree of difference, the condition for determining the quality of the inspection target based on the determination value C1 by calculating the determination value C1 for quality determination based on each calculation value c1 is the condition for determining the quality. It is defined as one of the conditions, and the degree of difference between the discriminated value A2 and the discriminated value B2 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the discriminated value A1. When the degree of difference from the discriminated value B1 is higher than the degree of difference from the discriminated value B1, the discriminated value C2 for pass / fail determination is calculated based on each calculated value c2, and the quality of the inspection target is discriminated based on the discriminated value C2. Then, the inspection data is generated by defining the condition as one of the quality determination conditions.

Further, the inspection data generation device according to claim 3 is the inspection data generation device according to claim 1 or 2, and the processing unit is the measured value data C at the time of inspection of the inspection target in the data generation process. Specify which of the measured values c0 recorded in the above measured value c0 from which measured value c0 to any other measured value c0 should be the measured value c0 within the measured value range. When the possible target measurement value range is defined as another one of the quality determination conditions, the measurement value a0 and the measurement value b0 corresponding to the start point of the target measurement value range and the end point of the target measurement value range A tentative range of M types (M is a natural number of 2 or more) in which the combination of the corresponding measured value a0 and the measured value b0 is at least different is defined, and each tentative range is defined as the measured value range. And the second process for each of the M types of provisional ranges, the degree of difference between the discriminated value A0 and the discriminated value B0, the discriminated value A1 and the discriminated value B1 The degree of difference between the above and the discriminated value A2 and the degree of difference between the discriminated value B2 are specified for each of the M types of provisional ranges, and the degree of the difference satisfies a predetermined condition. The third process for specifying the provisional range is executed, and the inspection data is generated with the specified provisional range as the target measurement value range.

Further, the inspection data generation device according to claim 4 is the inspection data generation device according to claim 3, and the processing unit has Na in each of the M types of provisional ranges in the third process. The minimum value of the non-defective product, which is the minimum value of the discriminated values A0, and the maximum value of the non-defective product, which is the maximum value of the Na discriminated values A0, are specified, and each of the Nb discriminated values B0 is specified. The difference from the non-defective product minimum value is calculated as the defect degree value for the discriminated value B0 smaller than the non-defective product minimum value, and the discriminated value B0 larger than the non-defective product maximum value is regarded as the non-defective product maximum value. The difference between the two is calculated as the degree of defect value, and the total value of each of the Nb discriminated values B0 is calculated, and the predetermined condition is satisfied based on the calculated total value. The process of specifying the tentative range to be related, the minimum non-defective value which is the minimum value of the Na discriminated values A1 for each of the M types of tentative ranges, and the Na discriminated value A1. The maximum value of the non-defective product, which is the maximum value of the products, is specified, and for each of the Nb discriminated values B1, the difference between the discriminated value B1 smaller than the minimum value of the non-defective product and the minimum value of the non-defective product is set as the defect degree value. For the discriminated value B1 that is larger than the maximum value of the non-defective product, the difference from the maximum value of the non-defective product is calculated as the degree of defect value, and the degree of each defect of the Nb discriminated values B1. A process of calculating the total value of the values and specifying the provisional range that satisfies the predetermined condition based on the calculated total value, and Na objects for each of the M types of provisional ranges. The minimum value of the non-defective product, which is the minimum value of the discrimination values A2, and the maximum value of the non-defective product, which is the maximum value of the Na discriminated values A2, are specified, and the Nb discriminated value B2 is described as described above. The difference from the non-defective product minimum value is calculated as the defect degree value for the discriminated value B2 smaller than the non-defective product minimum value, and the difference from the non-defective product maximum value for the discriminated value B2 larger than the non-defective product maximum value. Is calculated as the defect degree value, and the total value of each defect degree value of the Nb discriminated values B2 is calculated, and the relationship satisfying the predetermined condition based on the calculated total value. The process of specifying the provisional range is executed.

Further, the inspection data generation device according to claim 5 is the inspection data generation device according to claim 4, and the processing unit sets the total value as the non-defective product minimum value and the non-defective product maximum in the third process. Based on the value divided by the difference from the value, the provisional range that satisfies the condition of the predetermined condition is specified.

Further, the inspection data generation device according to claim 6 is the inspection data generation device according to any one of claims 3 to 5, and the processing unit has Na in the third process. The degree of difference between the distribution of the value A0 and the distribution of the Nb discriminated value B0 is quantified for each of the M provisional ranges according to a predetermined standard that becomes a higher value as the degree of difference increases. The degree of difference between the distribution of Na pieces of the discriminated value A1 and the distribution of Nb pieces of the discriminated value B1 is quantified for each of the M provisional ranges according to the predetermined criteria, and Na pieces are obtained. The degree of difference between the distribution of the discriminated value A2 and the distribution of Nb of the discriminated value B2 is quantified for each of the M provisional ranges according to the predetermined standard, and the preliminarily based on the numerical value. Identify the tentative range that satisfies the specified conditions.

Further, the inspection data generation device according to claim 7 is the inspection data generation device according to claim 6, and the processing unit uses one of the M types of provisional ranges in the third process. As a target, the tentative range starting from the measured value a0 whose starting point is the amount of difference from the measured value a0 corresponding to the starting point in the tentative range of the target is equal to or less than the La sampling cycle (La is a natural number). The numerical value of the provisional range whose end point is the measured value a0 whose end point is less than or equal to the Lb sampling cycle (Lb is a natural number) and the provisional range of the target. The amount of difference from the measured value b0 corresponding to the start point in the above is the provisional range starting from the measurement value b0 equal to or less than the La sampling cycle, and the measured value b0 corresponding to the end point in the provisional range of the target. The process of identifying each of the numerical values in the provisional range whose end point is the measured value b0 whose difference amount is equal to or less than the Lb sampling cycle, and associating the minimum value among the specified numerical values with the provisional range of the target. Each of the M temporary ranges is executed as the temporary range of the target, and the larger the associated numerical value is, the greater the degree of the relationship satisfying the predetermined condition is, and the relation satisfying the predetermined condition. Specify the provisional range that becomes.

Further, the inspection data generation device according to claim 8 is the inspection data generation device according to any one of claims 3 to 7, wherein the processing unit has the measured value a0 within the predetermined time. Each of the measured value data A, which periodically changes to a state exceeding the predetermined threshold value and a state below the predetermined threshold value, and the measured value b0 within the predetermined time are described above. Each of the measured value data B, which periodically changes to a state exceeding the predetermined threshold value and a state below the predetermined threshold value, is acquired, and in the data generation process, the predetermined value is specified in advance. The measured value a0 that exceeds the predetermined threshold value at the Kth time (K is a natural number) of the measured values a0 for the time, and the measured value b0 for the predetermined time. The measured value b0 that exceeds the predetermined threshold value at the Kth time, and the predetermined threshold value at the Kth time of the measured values a0 for the predetermined time. Which of the measured value a0 below the predetermined value and the measured value b0 below the predetermined threshold value at the Kth time of the respective measured values b0 for the predetermined time is defined in advance. The tentative range of the M types is defined with reference to one of them.

Further, the inspection data generation device according to claim 9 is the inspection data generation device according to any one of claims 3 to 8, wherein the processing unit is the target measurement value range specified by the third process. Executes a specific result notification process for notifying.

Further, the inspection data generation device according to claim 10 is the inspection data generation device according to claim 9, and the processing unit is at least the measured value data A and the measured value data in the specific result notification process. A waveform based on at least one of B and the measured value data C is displayed on the display device, and the target measured value range is displayed corresponding to the waveform to notify the user.

Further, the inspection data generation device according to claim 11 is the inspection data generation device according to any one of claims 1 to 10, and the processing unit is in charge of each measured value recorded in the measured value data A. a0 is extracted every other J (J1 is a natural number of 2 or more), the extracted measured values a0, the calculated values a1 corresponding to the extracted measured values a0, and the extracted measured values a0. Each of the calculated values a2 corresponding to the above is used to calculate the discriminated value A0, the discriminated value A1 and the discriminated value A2, and each measured value b0 recorded in the measured value data B is calculated. Every other J2 (J2 is a natural number of 2 or more), each measured value b0 extracted, each calculated value b1 corresponding to each extracted measured value b0, and each extracted measured value b0 correspond. Each of the calculated values b2 is used to calculate the discriminated value B0, the discriminated value B1, and the discriminated value B2, respectively.

Further, the inspection system according to claim 12 executes the inspection data generation device according to any one of claims 1 to 11 and the measurement process A to generate the measurement value data A, and the measurement process. A measuring device configured to execute B to generate the measured value data B and execute the measurement process C to generate the measured value data C, and an inspection generated by the inspection data generating device. The data and the inspection device for inspecting the quality of the inspection target based on the measurement value data C generated by the measurement device are provided so that the quality of the inspection target can be inspected.

Further, the inspection data generation processing program according to claim 13 is performed Na times (Na is 2) for a sample of a good product of the same type as the inspection target in a predetermined sampling cycle for a predetermined time. The Na measurement value data A in which a plurality of measurement values a0 for the predetermined time are recorded by the measurement process A of the above natural number), and the defective product of the same type as the inspection target and the good sample. For the sample, Nb pieces in which a plurality of measured values b0 for the predetermined time are recorded by the Nb times (Nb is a natural number of 2 or more) of the measurement process B performed under the same measurement conditions as the measurement process A. In addition to acquiring the measurement value data B of the above, Nc times (Nc is a natural number of 2 or more) performed on the inspection target under the same measurement conditions as the measurement process A and the measurement process B at the time of inspection of the inspection target. The inspection data that can identify the quality determination condition of the inspection target based on the Nc measurement value data C in which the plurality of measurement values c0 for the predetermined time are recorded by the process C is the measurement value data. A plurality of calculated values obtained by first-order differentiating each measured value a0 in the data generation process while causing the processing unit of the inspection data generation device to execute the data generation process generated based on A and each measured value data B. A1 and a plurality of calculated values a2 obtained by second-order differentiating each of the measured values a0 are derived for each of the measured value data A, and a plurality of calculated values b1 obtained by first-order differentiating each of the measured values b0, and each of them. The first process of deriving a plurality of calculated values b2 obtained by second-order differentiation of the measured value b0 for each of the measured value data B, and the respective measured values a0 within the predetermined measured value range and the measured value range. Based on each of the calculated values a1 in the corresponding calculated value range and each of the calculated values a2 in the calculated value range, the discriminated value A0 for pass / fail determination is calculated for each of the corresponding measured value data A. , The discriminated value A1 for pass / fail determination is calculated for each of the corresponding measured value data A based on the respective calculated values a1 in the calculated value range, and the respective calculated values a2 within the calculated value range. For each of the corresponding measured value data A, the discriminated value A2 for pass / fail determination is calculated, the measured value b0 within the measured value range, and the calculated value b1 within the calculated value range. , And each of the corresponding measured value data B based on each of the calculated values b2 in the calculated value range, the discriminated value B0 for pass / fail determination is calculated, and the calculated value range is calculated. The discriminated value B1 for pass / fail determination is calculated for each measured value data B corresponding based on the calculated value b1 in the above, and the corresponding calculated value b2 in the calculated value range is used for the corresponding calculation value B1. The second process of calculating the discriminated value B2 for pass / fail determination is executed for each of the measured value data B, and the degree of difference between the discriminated value A0 and the discriminated value B0 is the discriminated value. The inspection data is generated by defining one of the pass / fail judgment conditions based on the degree of difference between A1 and the discriminated value B1 and the degree of difference between the discriminated value A2 and the discriminated value B2. The processing unit is made to execute the processing to be performed.

In the inspection data generation device according to claim 1, in the data generation processing in which the processing unit generates inspection data capable of specifying the quality determination condition of the inspection target, each measured value a0 of the non-defective sample is first-order differentiated. A plurality of calculated values a1 and a plurality of calculated values a2 obtained by second-order differentiating each measured value a0 are derived for each measured value data A, and each measured value b0 for a defective sample is first-order differentiated. The first process of deriving a plurality of calculated values b2 obtained by second-order differentiation of the calculated value b1 and each measured value b0 for each measured value data B, each measured value a0, each calculated value a1, and each calculated value a2. The judgment value A0 for good / bad judgment is calculated for each of the corresponding measurement value data A based on Each calculation is performed, and the discriminated value A2 for pass / fail determination is calculated for each measurement value data A corresponding to each calculation value a2, and each measurement value b0, each calculation value b1 and each calculation value b2 are calculated. Based on each of the corresponding measurement value data B, the judgment value B0 for good / bad judgment is calculated, and the judgment value B1 for good / bad judgment is calculated for each measurement value data B corresponding to each calculation value b1. The second process of calculating and calculating the discriminated value B2 for pass / fail determination for each measured value data B corresponding based on each calculated value b2 is executed, and the discriminated value A0 and the discriminated value B0 are executed. One of the pass / fail judgment conditions is specified and inspected based on the degree of difference between the discriminated value A1 and the discriminated value B1, and the degree of difference between the discriminated value A2 and the discriminated value B2. Generate data for.

Further, in the inspection data generation device according to claim 2, the processing unit determines that the degree of difference between the discriminated value A0 and the discriminated value B0 is the degree of difference between the discriminated value A1 and the discriminated value B1. When the degree of difference between the discriminated value A2 and the discriminated value B2 is higher than the degree of difference, each measured value c0 within the measured value range and each measured value c0 within the measured value range in each measured value data C are first-order differentiated. Based on the calculated value c1 of the above and the plurality of calculated values c2 obtained by second-order differentiation of each measured value c0 within the measured value range, the discriminated value C0 for pass / fail judgment is calculated, and the good / bad of the inspection target is calculated based on the discriminated value C0. Is defined as one of the good / bad judgment conditions, and the degree of difference between the discriminated value A1 and the discriminated value B1 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the discriminated value. When the degree of difference between the value A2 and the discriminated value B2 is higher than the degree of difference, the discriminant value C1 for pass / fail determination is calculated based on each calculated value c1 and the pass / fail of the inspection target is discriminated based on the discriminated value C1. Is defined as one of the good / bad judgment conditions, and the degree of difference between the discriminated value A2 and the discriminated value B2 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the discriminated value A1. When the degree of difference from the discriminated value B1 is higher than the degree of difference, the condition that the discriminated value C2 for pass / fail judgment is calculated based on each calculated value c2 and the quality of the inspection target is discriminated based on the discriminated value C2. Inspection data is generated by specifying it as one of the pass / fail judgment conditions.

Further, in the inspection system according to claim 12, the inspection data generator that executes the data generation process and the measurement process for the non-defective sample, the defective sample, and the inspection target are executed to measure the measured value data A and the measured value. It is configured to be equipped with a measuring device that generates data B and measured value data C and an inspection device that inspects the quality of the inspection target based on the inspection data and the measured value data C so that the quality of the inspection target can be inspected. .. Further, in the inspection data generation processing program according to claim 13, the processing unit of the inspection data generation device is made to execute the above data generation processing.

Therefore, according to the inspection data generation device according to claims 1 and 2, the inspection system according to claim 12, and the inspection data generation processing program according to claim 13, the discriminated value A0 and the discriminated value B0 For an inspection target in which a type of defect that may cause a sufficiently high degree of difference is calculated, the discriminated value C0 is calculated based on each measured value c0 in the measured value data C for the inspection target to determine the quality. For inspection targets in which inspection data is generated that defines the pass / fail judgment conditions, and the degree of difference between the discriminated value A0 and the discriminated value B0 is not so high, a type of defect may occur. Based on the discriminated value C1 calculated based on the calculated value c1 obtained by first-order differentiating each measured value c0 in the data C, and the calculated value c2 obtained by second-order differentiating each measured value c0 in the measured value data C for the inspection target. Since the inspection data in which the pass / fail judgment condition for determining the pass / fail is determined based on the calculated discriminated value C2 is generated, by executing the inspection based on this inspection data, various inspection targets can be obtained. Good products can be accurately identified.

In the inspection data generation device according to claim 3, in the data generation processing, the processing unit has the other measurement value c0 from each measurement value c0 recorded in the measurement value data C at the time of inspection of the inspection target. Target measurement value range when defining the target measurement value range that can specify which measurement value c0 up to the measurement value c0 should be the measurement value c0 within the measurement value range as another one of the quality determination conditions Specify M types of tentative ranges in which the combination of the measured values a0 and b0 corresponding to the start point of the above and the measured values a0 and b0 corresponding to the end point of the target measured value range are at least different, and each tentative range is defined. As the measured value range, the first process and the second process are executed for each of the M types of temporary ranges, and the degree of difference between the discriminated value A0 and the discriminated value B0, the discriminated value A1 and the discriminated value B1 The degree of difference between the two and the discriminated value A2 and the discriminated value B2 are specified for each of the M types of provisional ranges, and the degree of difference satisfies the predetermined condition. The third process for specifying the above is executed, and the inspection data is generated with the specified provisional range as the target measurement value range.

Therefore, according to the inspection data generation device according to claim 3, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, the type of inspection target. Or, depending on the difference in the measurement environment, from the measured values c0 recorded in the measured value data C for the inspection target, the discriminated value A0, the discriminated value A1 or the discriminated value A2 for the non-defective sample. And, the target measurement value range in which the discriminated value B0, the discriminated value B1 or the discriminated value B2 of the defective sample has a significantly different relationship is automatically specified and the inspection data is generated. By inspecting the inspection target based on the inspection data, it is possible to accurately determine the quality of the inspection target where various defects may occur.

In the inspection data generation device according to claim 4, in the third process, the processing unit has a non-defective product minimum value, which is the minimum value among the Na discriminated values A0, and Na for each of the M types of provisional ranges. The maximum non-defective product value, which is the maximum value among the discriminated values A0, is specified, and for each Nb discriminated value B0, the difference from the non-defective product minimum value is set for the discriminated value B0 smaller than the minimum non-defective product value. For the discriminated value B0 that is calculated as the defect degree value and is larger than the non-defective product maximum value, the difference from the non-defective product maximum value is calculated as the defect degree ratio value, and the sum of each defect degree value of Nb discriminated value B0. A process of calculating a value and specifying a provisional range that satisfies a predetermined condition based on the calculated total value, and a minimum value among Na discriminated values A1 for each of the M types of provisional ranges. The minimum value of the non-defective product and the maximum value of the non-defective product, which is the maximum value among the discriminated values A1 of Na, are specified, and for each Nb discriminated value B1, the discriminated value B1 smaller than the minimum value of the non-defective product is obtained. Calculates the difference from the minimum non-defective product value as the degree of defect value, and for the discriminated value B1 larger than the maximum value of non-defective product, the difference from the maximum value of non-defective product is calculated as the degree of defect value, and Nb discriminated values. A process of calculating the total value of each defect degree value of B1 and specifying a tentative range that satisfies a predetermined condition based on the calculated total value, and Na coverings for each of the M types of tentative ranges. The minimum value of the non-defective product, which is the minimum value of the discriminant value A2, and the maximum value of the non-defective product, which is the maximum value of the discriminated values A2 of Na, are specified, and the minimum value of the non-defective product is specified for each Nb discriminated value B2. For the discriminated value B2 smaller than, the difference from the minimum non-defective product value is calculated as the defect degree value, and for the discriminated value B2 larger than the maximum non-defective product value, the difference from the maximum non-defective product value is calculated as the defect degree value. At the same time, the total value of each defect degree value of Nb discriminated values B2 is calculated, and a process of specifying a provisional range that satisfies a predetermined condition based on the calculated total value is executed.

Therefore, according to the inspection data generation device according to claim 4, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, in a non-defective sample. Since it is possible to eliminate the influence of the variation of the discriminated value that may occur, it is possible to preferably specify the provisional range suitable for the target measurement value range and generate the inspection data.

According to the inspection data generation device according to claim 5, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, the processing unit is the third. By specifying a provisional range that satisfies a predetermined condition based on the value obtained by dividing the total value by the difference between the non-defective product minimum value and the non-defective product maximum value, the non-defective product minimum value and the non-defective product maximum value can be obtained. Regarding the tentative range in which the difference between the two is the same as the discriminated value B0, the discriminated value B1 and the discriminated value B2, the tentative range in which the difference between the non-defective product minimum value and the non-defective product maximum value is large (the discriminated value A0 for the non-defective product sample) , The provisional range in which the difference between the non-defective product minimum value and the non-defective product maximum value is smaller than the provisional range in which the discriminant value A1 and the discriminated value A2 have a large variation (discrimination value A0, discriminated value A1 and discriminated value A1 for the non-defective product sample). Since a larger value is calculated in the provisional range in which the variation of the discriminated value A2 is small), the difference between the discriminated value A0, the discriminated value A1 and the discriminated value A2 for the non-defective sample is large. The inspection data can be generated by preferably specifying the provisional range in which the discrimination value B0, the discrimination value B1 and the discrimination value B2 can be calculated, that is, the provisional range in which the quality of the inspection target can be more reliably determined. ..

In the inspection data generation device according to claim 6, the processing unit determines the degree of difference between the distribution of Na discriminated values A0 and the distribution of N b discriminated values B0 in the third process. The higher the degree, the higher the value. According to a predetermined standard, each of the M provisional ranges is quantified, and the degree of difference between the distribution of Na discriminated values A1 and the distribution of N b discriminated values B1 is determined. Each of the M provisional ranges is quantified according to a predetermined standard, and the degree of difference between the distribution of Na discriminated value A2 and the distribution of Nb discriminated value B2 is M according to the predetermined standard. Each of the tentative ranges is quantified, and the tentative ranges that satisfy the predetermined conditions are specified based on the numerical values.

Therefore, according to the inspection data generation device according to claim 6, an inspection system provided with such an inspection data generation device, and an inspection data generation processing program for executing such processing, a non-defective sample is obtained. Accurately identify a provisional range in which the distribution of the discriminated value A0, the discriminated value A1 and the discriminated value A2, and the distribution of the discriminated value B0, the discriminated value B1 and the discriminated value B2 for the defective sample are large. can do.

In the inspection data generation device according to claim 7, in the third process, the processing unit targets one of the M types of temporary ranges and sets the measured value a0 corresponding to the start point in the target temporary range. A tentative range in which the amount of difference is a tentative range starting from a measured value a0 that is less than or equal to the La sampling cycle, and an end point is a tentative range in which the amount of difference from the measured value a0 corresponding to the end point in the tentative range of interest is less than or equal to the Lb sampling cycle. The amount of difference between the numerical value of the tentative range to be measured and the measured value b0 corresponding to the start point in the tentative range of the target is the tentative range starting from the measured value b0 equal to or less than the La sampling cycle, and the end point in the tentative range of the target. A process of identifying each of the values in the tentative range whose end point is the measured value b0 whose amount of difference from the corresponding measured value b0 is less than or equal to the Lb sampling cycle, and associating the minimum value of each specified value with the target tentative range. Each of the M tentative ranges is executed as the target tentative range, and the tentative range that satisfies the predetermined condition is satisfied as the degree of the relationship that satisfies the predetermined condition is larger as the associated numerical value is larger. To identify.

Therefore, according to the inspection data generation device according to claim 7, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, a defective sample. A provisional range in which the discriminated value B0, the discriminated value B1 and the discriminated value B2 become large due to the sudden deviation of the measured values during the measurement process B is specified as the target measured value range. It occurs in a tentative range in which the discriminated value B0, the discriminated value B1 and the discriminated value B2 are sufficiently large values even in the tentative range where the start point and the end point are the same without causing a situation, that is, in a defective sample. It is possible to generate inspection data by specifying a provisional range including a measured value that is different from the measured value for a non-defective sample due to the influence of a constant defect as a target measured value range.

In the inspection data generation device according to claim 8, the processing unit periodically enters a state in which the measured value a0 exceeds a predetermined threshold value and a state in which the measured value a0 falls below the predetermined threshold value within a predetermined time. Each measured value data A that changes cyclically, and each measured value b0 that periodically changes to a state in which the measured value b0 exceeds a predetermined threshold value and a state in which the measured value b0 falls below a predetermined threshold value within a predetermined time. In addition to acquiring the measured value data B, in the data generation process, the measured value a0 exceeding the predetermined threshold value at the Kth time of each measured value a0 for the predetermined time, and the predetermined measured value a0. The measured value b0 that exceeds the predetermined threshold value at the Kth time of each measured value b0 for the time and the Kth time of the measured value a0 for the predetermined time are predetermined. Either the measured value a0 below the threshold value or the measured value b0 below the predetermined threshold value at the Kth time of each measured value b0 for the predetermined time, whichever is predetermined. As a standard, M types of provisional ranges are specified.

Therefore, according to the inspection data generation device according to claim 8, an inspection system provided with such an inspection data generation device, and an inspection data generation processing program for executing such processing, a non-defective sample is obtained. Even if there is a slight deviation in the starting points of the measured value data A and the measured value data B during the measurement process A and the measurement process B for the defective sample, the same standard is used for all the measured value data. Temporary range based on the measured value that satisfies (the measured value a0, b0 that exceeds the predetermined threshold value at the Kth time, or the measured value a0, b0 that falls below the predetermined threshold value at the Kth time). Since the start point of the measurement value can be specified, the effect of a slight deviation of the start point in the measurement value data A and the measurement value data B can be eliminated, and a provisional range suitable for the target measurement value range can be specified.

According to the inspection data generation device according to claim 9, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, the processing unit is the third. By executing the specific result notification processing that notifies the target measurement value range specified by the processing of, it is certain to the user what range of measurement values is to be inspected for the quality of the inspection target. And it can be easily recognized.

According to the inspection data generation device according to claim 10, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, the processing unit determines the specific result. In the notification process, a waveform based on at least one of the measured value data A, the measured value data B, and the measured value data C is displayed on the display device, and the target measured value range is displayed corresponding to the waveform. By notifying with, it is possible to make the user more easily recognize the quality of the inspection target to be inspected for the measured value included in what range.

According to the inspection data generation device according to claim 11, the inspection system provided with such an inspection data generation device, and the inspection data generation processing program for executing such processing, the processing unit determines the measured value. Each measured value a0 recorded in the data A is extracted every other J, and corresponds to each extracted measured value a0, each calculated value a1 corresponding to each extracted measured value a0, and each extracted measured value a0. Each calculated value a2 is used to calculate the discriminated value A0, the discriminated value A1 and the discriminated value A2, and each measured value b0 recorded in the measured value data B is extracted and extracted every J2. The discriminated value B0, the discriminated value B1 and the discriminated value using each measured value b0, each calculated value b1 corresponding to each extracted measured value b0, and each calculated value b2 corresponding to each extracted measured value b0. By calculating each value B2, the discriminated value A0, the discriminated value A1 and the discriminated value A2, the discriminated value B0, the discriminated value B1 and the discriminated value B2 are targeted for all the measured values included in the provisional range. Since the number of measured values to be calculated is smaller than that in the case of calculating, the burden on the processing unit and the capacity of the memory for temporarily storing the calculation result can be sufficiently reduced.

It is a block diagram which shows the structure of inspection system 1. It is explanatory drawing for demonstrating an example of the damped vibration waveform measured when the impulse voltage is applied to the inspection object. It is explanatory drawing for demonstrating an example of the waveform W0 of the measured value D0. It is explanatory drawing for demonstrating an example of the waveform W1 of the calculated value D1. It is explanatory drawing for demonstrating an example of the waveform W2 of the calculated value D2. It is explanatory drawing for demonstrating an example of the waveform W0f of the measured value D0f. It is explanatory drawing for demonstrating an example of the waveform W1f of the calculated value D1f. It is explanatory drawing for demonstrating an example of the waveform W2f of the calculated value D2f. It is a display screen view which shows an example of the specific result display 31 displayed on the display part 22 of the data processing apparatus 3 in inspection system 1. Explanatory drawing for explaining an example of LC value (RC value) calculated based on the measured value for a non-defective sample Xa, and LC value (RC value) calculated based on the measured value for a defective sample Xb. Is. It is explanatory drawing for demonstrating an example of the relationship between the distribution of LC value (RC value) of a good sample Xa, and LC value (RC value) of defective sample Xb. It is explanatory drawing for demonstrating another example of the relationship between the distribution of LC value (RC value) of a good sample Xa, and LC value (RC value) of defective sample Xb. It is explanatory drawing for quantifying the relationship between the type of the inspection target value range, the LC value (RC value) of a non-defective sample Xa, and the LC value (RC value) of a defective sample Xb. It is another explanatory diagram for quantifying the relationship between the type of the inspection target value range, the LC value (RC value) of a non-defective sample Xa, and the LC value (RC value) of a defective sample Xb. FIG. 5 is a display screen diagram showing an example of a measurement result display screen 40 displayed on the display unit 22 of the data processing device 3 in the inspection system 1.

Hereinafter, embodiments of an inspection data generator, an inspection system, and an inspection data generation processing program will be described with reference to the accompanying drawings.

First, the configuration of the inspection system 1 will be described. The inspection system 1 shown in FIG. 1 is an "impulse test system" which is an example of an "inspection system" in which a "measurement device", a "data generation device for inspection", and an "inspection device" are integrally provided, and measures the measurement. A device 2 and a data processing device 3 are provided so that the quality of the inspection target X (an example of the “inspection target”) can be inspected. In this case, the inspection target X is an example of the “inspection target”, and in this example, an example of inspecting the quality of the wound component (coil) as the inspection target X will be described. Regarding this inspection target X, a non-defective sample Xa (an example of a “non-defective sample”) which is a winding component of the same type (of the same model product) as the inspection target X, and a non-defective sample Xa (an example of a “non-defective sample”) of the same type as the inspection target X (same model product). It is assumed that defective sample Xb (an example of "defective sample"), which is a winding component, is prepared respectively.

The measuring device 2 corresponds to a "measuring device", and as an example, is configured to be capable of executing various measurement processes targeting the inspection target X, the non-defective sample Xa, and the defective sample Xb under the control of the data processing device 3. ing. Specifically, the measuring device 2 includes a measurement signal generation unit 11, an A / D conversion unit 12, a processing unit 13, a storage unit 14, and the like. The measurement signal generation unit 11 applies an impulse voltage as a measurement signal between both ends of the measurement target (inspection target X, non-defective sample Xa, defective sample Xb, etc.) under the control of the processing unit 13. As an example, the A / D conversion unit 12 performs A / D conversion (sampling: measurement) of the voltage value between both ends of the measurement target in a designated cycle (sampling cycle: measurement cycle) according to the control of the processing unit 13. The measured value D0 (sampling value) is sequentially output to the processing unit 13.

The processing unit 13 comprehensively controls the measuring device 2. Specifically, the processing unit 13 controls the measurement signal generation unit 11 to apply an impulse voltage to the measurement target, and also controls the A / D conversion unit 12 to perform A / D conversion of the voltage value at an arbitrary cycle. The process (sampling process) is executed. Further, the processing unit 13 stores the measured value D0 output from the A / D conversion unit 12 in the storage unit 14, generates the measured value data Dm based on the measured value D0, and stores the measured value data Dm in the storage unit 14. , The generated measured value data Dm is output to the data processing device 3. The storage unit 14 stores the data of the operation program of the processing unit 13, the above-mentioned measured value D0, the measured value data Dm, and the like. The actual measuring device 2 is configured to include various operation switches for instructing the operating conditions of the measuring device 2, and a display unit for displaying a measurement condition setting screen, a measurement value display screen, and the like. However, illustration and description of these will be omitted.

On the other hand, the data processing device 3 corresponds to an "inspection data generator", and inspects the inspection target X based on the measured value data Dm of the non-defective sample Xa and the defective sample Xb output from the measuring device 2. Generate inspection data Di to perform. Further, the data processing device 3 corresponds to an "inspection device", and inspects the quality of the inspection target X based on the generated inspection data Di and the measured value data Dm for the inspection target X output from the measurement device 2. do. In this case, in the inspection system 1 of this example, as an example, the program data Dp corresponding to the "inspection data generation processing program" is installed in the existing personal computer to configure the data processing device 3.

Specifically, the data processing device 3 includes an operation unit 21, a display unit 22, a processing unit 23, and a storage unit 24. The operation unit 21 includes a keyboard and pointing devices such as a mouse and a touch panel (not shown), and outputs an operation signal corresponding to the operation for these to the processing unit 23. The display unit 22 is an example of a “display device” and displays various display screens such as the specific result display 31 shown in FIG. 9 and the measurement result display screen 40 shown in FIG. 15 under the control of the processing unit 23. ..

The processing unit 23 is an example of the “processing unit” and controls the data processing device 3 in a comprehensive manner. Specifically, as will be described later, the processing unit 23 controls the measuring device 2 according to the program data Dp to perform a measurement process targeting the non-defective sample Xa (an example of “measurement process A”), and the defective sample Xb. (An example of "Measurement process B"), and a process of generating inspection data Di based on the measured value data Dm output from the measuring device 2 (an example of "Data generation process"). ) Is executed. The process of generating the inspection data Di will be described in detail later.

Further, the processing unit 23 controls the measuring device 2 to execute the measurement process (an example of “measurement process C”) targeting the inspection target X according to the program data Dp, and also generates the inspection data Di and the generated inspection data Di. An inspection process for inspecting the quality of the inspection target X is executed based on the measured value data Dm output from the inspection system 1. This inspection process will also be described in detail later. The storage unit 24 includes program data Dp (data of the operation program of the processing unit 23), measured value data Dm output from the inspection system 1, calculated value data Dc1 and Dc2 generated by the processing unit 23, and inspection data Di. And so on.

Next, the inspection method of the inspection target X by the inspection system 1 will be described with reference to the attached drawings. It is assumed that the work of installing the program data Dp in the data processing device 3 and the work of connecting the measuring device 2 and the data processing device 3 have already been completed.

When generating the inspection data Di for inspecting the inspection target X, first, the measurement value data Dm is obtained by the measurement process targeting the non-defective sample Xa, and the measured value is obtained by the measurement process targeting the defective sample Xb. Obtain the data Dm. At this time, a plurality of measurement processes for the non-defective sample Xa are executed to obtain a plurality of measurement value data Dm, and a plurality of measurement processes for the defective sample Xb are executed for a plurality of times. By obtaining the measured value data Dm, when the inspection data Di is later generated based on these measured value data Dm, the variation of the measured value D0 that may occur in the measurement process performed at the time of the inspection of the inspection target X It is possible to generate inspection data Di that reflects the degree.

Specifically, first, the operation unit 21 of the data processing device 3 is operated to instruct the start of the inspection data Di generation process. At this time, the processing unit 23 causes the display unit 22 to display a message "Please set a non-defective sample" as an example according to the program data Dp. In response to this, the user connects the measuring device 2 (measurement signal generation unit 11 and A / D conversion unit 12) to the non-defective sample Xa. Next, the processing unit 23 of the data processing device 3 controls the measuring device 2 according to the program data Dp, so that, as an example, the non-defective sample Xa is measured 30 times (Na = 30 times) (“Measuring process A”. "Example)) is executed.

At this time, in the measuring device 2, the processing unit 13 first controls the A / D conversion unit 12 to obtain a voltage at a sampling cycle (an example of a “predetermined sampling cycle”) instructed by the processing unit 23. Start sampling (measurement) of the value. As a result, the measured value D0 for the non-defective sample Xa (voltage value between both ends of the non-defective sample Xa: an example of “measured value a0”) is sequentially output from the A / D conversion unit 12 and stored in the storage unit 14. Further, the processing unit 13 controls the measurement signal generation unit 11 to apply an impulse voltage to the non-defective sample Xa. At this time, the measured value D0 (voltage value) between both ends of the non-defective sample Xa changes as shown in the damped vibration waveform shown in FIG.

Next, as an example, the processing unit 13 A / within this time TA at the time point te when the time TA instructed by the processing unit 23 has elapsed from the time point ts immediately before starting the application of the impulse voltage to the non-defective sample Xa. A plurality of measured values D0, D0 ... Output from the D conversion unit 12 are recorded, and the measured value data Dm (“Na pieces in which a plurality of measured values a0 for a time predetermined by the measurement process A are recorded respectively). An example of one of the “measured value data A”) is generated, and the generated measured value data Dm is stored in the storage unit 14. Further, the processing unit 13 outputs the generated measured value data Dm to the data processing device 3. As a result, one of the Na = 30 measurement processes is completed.

In this case, the above time TA is a time during which the damped vibration generated between both ends of the non-defective sample Xa is sufficiently damped after the application of the impulse voltage is completed, and is the time to be inspected X (non-defective sample Xa and defective sample Xb). It is predetermined according to the type (model) of. Further, in the inspection system 1 of this example, as an example, the time TA is instructed by the data processing device 3 to the measuring device 2 by presenting "the number of measured values D0 constituting the measured value data Dm". Has been adopted.

On the other hand, in the data processing device 3, the processing unit 23 stores the measured value data Dm output from the measuring device 2 in the storage unit 24 as one of the data regarding the non-defective sample Xa. After that, in the measuring device 2, the processing unit 13 sequentially executes the second and subsequent measurement processes out of the Na = 30 measurement processes. As a result, the measured value data Dm of Na = 30 pieces of the non-defective sample Xa output from the measuring device 2 is stored in the storage unit 24 of the data processing device 3 (“predetermined time (this example). Then, Na pieces that periodically change into a state in which the measured value a0 exceeds a predetermined threshold value (for example, 0V) within the time TA) and a state in which the measured value a0 falls below the predetermined threshold value (0V). An example of a state in which the process of "acquiring the measured value data A of" has been completed).

Further, when the storage of the measured value data Dm of Na = 30 pieces for the non-defective sample Xa is completed, the processing unit 23 sets the defective sample in place of the non-defective sample as an example according to the program data Dp. Please display the message "Please." On the display unit 22. In response to this, the user connects the measuring device 2 (measurement signal generation unit 11 and A / D conversion unit 12) to the defective sample Xb. Next, the processing unit 23 of the data processing device 3 controls the measuring device 2 according to the program data Dp, so that, as an example, the defective sample Xb is measured 30 times (Nb = 30 times) (“Measuring process”. An example of "B") is executed.

Although detailed description is omitted, the measurement process of Nb = 30 times for the defective sample Xb is performed under the same measurement conditions as the measurement process of Na = 30 times for the non-defective sample Xa. Is executed by. As a result, Nb = 30 measured value data in which the measured values D0 (voltage value between both ends of the defective sample Xb: an example of “measured value b0”) for Nb = 30 defective sample Xb are recorded. Dm (an example of “measured value data B”) is stored in the storage unit 24 of the data processing device 3 (“measured value b0 is predetermined within a predetermined time (in this example, time TA)). Acquire Nb measured value data B that periodically change to a state exceeding the set threshold value (0V in this example) and a state falling below the predetermined threshold value (0V). " An example of the completed state).

Further, the processing unit 23 inspects when the storage (acquisition) of Na = 30 measured value data Dm for the non-defective sample Xa and Nb = 30 measured value data Dm for the defective sample Xb is completed. The generation process of the inspection data Di for inspecting the quality of the target X (an example of the “data generation process”) is started.

In this case, in the inspection system 1 (data processing device 3) of this example, the processing unit 23 of the data processing device 3 acquires the above-mentioned Na = 30 measured value data Dm and Nb = 30 according to the program data Dp. Based on the measured value data Dm of the above, the same measurement conditions as the above-mentioned measurement processes (measurement process A) for the non-defective sample Xa and the above-mentioned measurement processes (measurement process B) for the defective sample Xb at the time of inspection of the inspection target X. A plurality of measured values D0 (“measured value c0”) for the above time TA (predetermined time) by the measurement process (measurement process C) of 5 times (Nc = 5 times example) for the inspection target X performed in An example) of Nc = 5 measured value data Dm (an example of “measured value data C”) is used to generate inspection data Di that can specify the condition for determining the quality of the inspection target X. Has been adopted.

Specifically, in the inspection system 1 (data processing device 3) of this example, the processing unit 23 performs Na = 30 measured value data Dm for the non-defective sample Xa and Nb for the defective sample Xb according to the program data Dp. = By analyzing 30 measured value data Dm, when inspecting the inspection target X, each measured value D0 and the measured value D0 recorded in the measured value data Dm for the inspection target X are first-order differentiated and calculated value D1. , And a method of inspecting the inspection target X based on the LC value and RC value calculated using the calculated value D2 obtained by second-order differentiation of the measured value D0, and the LC value and RC calculated using the calculated value D1. A method of inspecting the inspection target X based on a value (a multiplication value calculated based on the first-order differential component), and an LC value and an RC value (calculated based on the second-order differential component) calculated using the calculated value D2. Which of the methods to inspect the inspection target X based on the multiplication value to be performed) and the filtering process for each measured value D0 prior to the calculation of the LC value and the RC value when inspecting the inspection target X. Whether or not to perform the above is specified as one of the "good / bad judgment conditions". The calculation of the calculated values D1 and D2, the calculation of the LC value and the RC value, the content of the filtering process, and the like will be described in detail later.

Further, in the inspection system 1 (data processing device 3) of this example, the processing unit 23 has Na = 30 measured value data Dm for the non-defective sample Xa and Nb = 30 for the defective sample Xb according to the program data Dp. By analyzing the measured value data Dm of the above, when the inspection target X is inspected, any of the measured values D0, D0, ... The "target measurement value range" that can specify whether each measurement value D0, D0 ... Up to any measurement value D0 should be the measurement value D0 within the "predetermined measurement value range" is the "good / bad judgment condition". It is specified as the other one. Each process for specifying this "target measurement value range" will also be described in detail later.

In the process of generating the inspection data Di, the processing unit 23 first sets each measurement value D0 (an example of "measurement value a0") on the first floor for each measurement value data Dm of Na = 30 pieces for the non-defective sample Xa. A plurality of differentiated operation values D1 (an example of "calculation value a1") and a plurality of operation values D2 obtained by second-order differentiation of each measurement value D0 (an example of "operation value a2") are obtained for each measurement value data A. A plurality of calculated values D1 (“calculated value b1”) obtained by first-order differentiating each measured value D0 (an example of “measured value b0”) for each Nb = 30 measured value data Dm for the defective sample Xb. (Example), and a plurality of calculated values D2 (an example of the "calculated value b2") obtained by second-order differentiation of each measured value D0 are derived (an example of the "first process").

In this case, as an example, when the waveform of each measured value D0 in any of the measured value data Dm for the non-defective sample Xa is the waveform W0 shown in FIG. 3 (hereinafter, also referred to as “voltage waveform W0”), each calculated value. The waveform of D1 is like the waveform W1 shown in FIG. 4 (hereinafter, also referred to as “first-order differential waveform W1”), and the waveform of each calculated value D2 is the waveform W2 (hereinafter, “second-order differential waveform”) shown in FIG. It is also called "W2").

Further, the processing unit 23 executes a filtering process for each measured value D0 for each measured value data Dm of Na = 30 pieces of the non-defective sample Xa, and each measured value D0 after the filtering process (“measured value”). A plurality of calculated values D1 (another example of "calculated value a1") obtained by first-order differentiating (another example of a0), and a plurality of calculated values D2 ("calculated") obtained by second-order differentiating each measured value D0 after filtering. Another example of "value a2") is derived for each of the measured value data A, and filtering processing for each measured value D0 is performed for each Nb = 30 measured value data Dm for the defective sample Xb. Is executed, and each measured value D0 after the filtering process (another example of the “measured value b0”) is first-order differentiated into a plurality of calculated values D1 (another example of the “calculated value b1”), and after the filtering process. A plurality of calculated values D2 (another example of the "calculated value b2") obtained by second-order differentiation of each measured value D0 are derived (another example of the "first process").

Specifically, the processing unit 23 first has 60 measured value data of Na = 30 measured value data Dm for the non-defective sample Xa and Nb = 30 measured value data Dm for the defective sample Xb. For Dm, a filtering process (filtering using a low-pass filter) for converting a measured value D0 below a predetermined threshold value to the same value as the threshold value is executed. In the following description, the measured value data Dm after the filtering process is also referred to as the measured value data Dmf in order to distinguish it from the measured value data Dm that has not been filtered. Further, in order to distinguish from each measured value D0 of the measured value data Dm, each measured value D0 of the measured value data Dmf is also referred to as a measured value D0f.

Next, the processing unit 23 derives a plurality of calculated values D1 obtained by first-order differentiating each measured value D0f for each of the measured value data Dmf of Na = 30 pieces for the non-defective sample Xa, and generates the calculated value data Dc1. , A plurality of calculated values D2 obtained by second-order differentiation of each measured value D0f are derived to generate calculated value data Dc2. Similarly, the processing unit 23 derives a plurality of calculated values D1 obtained by first-order differentiating each measured value D0f for each Nb = 30 measured value data Dmf for the defective sample Xb, and obtains the calculated value data Dc1. At the same time as generating, a plurality of calculated values D2 obtained by second-order differentiation of each measured value D0f are derived to generate calculated value data Dc2.

In the following description, in order to distinguish from the calculated value data Dc1 and Dc2 generated based on the measured value data Dm, the calculated value data Dc1 and Dc2 generated based on the measured value data Dmf are referred to as the calculated value data. Also referred to as Dc1f and Dc2f. Further, in order to distinguish from the calculated values D1 and D2 calculated based on each measured value D0, the calculated values D1 and D2 calculated based on each measured value D0f are also referred to as calculated values D1f and D2f.

In this case, as an example, when the waveform of each measured value D0f in any of the measured value data Dmf for the non-defective sample Xa is the waveform W0f shown in FIG. 6 (hereinafter, also referred to as “voltage waveform W0f”), each calculated value. The waveform of D1f becomes like the waveform W1f shown in FIG. 7 (hereinafter, also referred to as “first-order differential waveform W1f”), and the waveform of each calculated value D2f is the waveform W2f (hereinafter, “second-order differential waveform” shown in FIG. 8). It is also called "W2f").

Next, the processing unit 23 defines a plurality of types of "temporary ranges (an example of" M types of temporary ranges ")" that are candidates for the above-mentioned "target measurement value range". In this case, in the data processing device 3 (program data Dp) of this example, as will be described later, each measured value data Dm (Dmf) in which the measured values D0 (D0f) for the time TA are recorded is "temporarily ranged". The LC value and RC value are calculated with different values, and the LC value and RC value for the non-defective sample Xa and the LC value and RC value for the defective sample Xb are significantly different from each other. To identify.

Further, in the data processing device 3 (program data Dp) of this example, each calculated value data Dc1 (Dc1f) in which the calculated value D1 (D1f) for the time TA is recorded has a different "temporary range" and LC. The values and RC values are calculated respectively, and the "target measurement value range" in which the LC value or RC value for the non-defective sample Xa and the LC value or RC value for the defective sample Xb are significantly different is specified.

Further, in the data processing device 3 (program data Dp) of this example, the “temporary range” is different for each calculated value data Dc2 (Dc2f) in which the calculated values D2 (D2f) for the time TA are recorded, and LC is performed. The values and RC values are calculated respectively, and the "target measurement value range" in which the LC value or RC value for the non-defective sample Xa and the LC value or RC value for the defective sample Xb are significantly different is specified.

Therefore, according to the program data Dp, the processing unit 23 sets the measured value D0 (D0f), the calculated value D1 (D1f) or the calculated value D2 (D2f) corresponding to the start point of the above "target measured value range", and the "target measurement". A plurality of types of "temporary ranges" with different combinations of the measured values D0 (D0f), the calculated values D1 (D1f), and the calculated values D2 (D2f) corresponding to the end points of the "value range" are defined.

In this case, in the data processing device 3 (program data Dp) of this example, as an example, by defining M types of "temporary range" for the measured value data Dm (measured value D0), the M types of "temporary range" are defined. Corresponding to the "range", M types of "temporary range" for the measured value data Dmf (measured value D0f), M types of "temporary range" for the calculated value data Dc1 (calculated value D1), and calculated value data Dc1f ( M types of "temporary range" for the calculated value D1f), M types of "temporary range" for the calculated value data Dc2 (calculated value D2), and M types of "temporary range" for the calculated value data Dc2f (calculated value D2f). The configuration (method) in which the "range" is defined is adopted.

Further, in the data processing device 3 (program data Dp) of this example, as an example, the measurement exceeding 0 V at the beginning of the measured values D0, D0 ... For the time TA hours recorded in the measured value data Dm. The value D0 is set as the reference measurement value D0 (“K = the first measurement value exceeding 0V as a predetermined threshold value” as a reference), and “the reference measurement value D0 to ○ The measured value D0 that satisfies the condition of "before (or after XX)" is set as the start point of the "temporary range", and the measured value that satisfies the condition of "after XX from the measured value D0 as the starting point". A configuration is adopted in which the "temporary range" is defined with D0 as the end point of the "temporary range". For the above "threshold value", any value other than 0V can be set.

Specifically, as one of the M types of "temporary range", as an example, the measured value D0 at the reference point P0 (time point t0) shown in FIG. 2 is used as a reference, and the measured value D0 at the reference point P0 is counted. The measured value D0 at the time point t1 corresponding to the number of times T1 is set as the start point P1 of the "provisional range", and the measured value D0 at the time point t2 after the number corresponding to the time T2 counted from the measured value D0 at the start point P1. Is defined as the end point P2 of the "temporary range". Further, as another one of the M types of "temporary range", the measured value D0 at the reference point P0 is used as a reference, and the time point t1a after the number corresponding to the time T1a counted from the measured value D0 at the reference point P0 The measured value D0 is set as the start point P1a of the "temporary range", and the measured value D0 at the time point t2a after the number corresponding to the time T2a counted from the measured value D0 of the start point P1a is set as the end point P2a of the "temporary range". Is stipulated in.

In this case, in the data processing device 3 (program data Dp) of this example, for the purpose of reducing the burden required for generating the inspection data Di, for example, the LC value or RC based on the measured value data Dm, as will be described later. When calculating the values, etc., the measured values D0, D0 ... Recorded in the measured value data Dm are extracted every 5 pieces (example of "J1 = J2 = 5 pieces"), and the extracted measured values D0, D0 ... -A configuration is adopted in which the LC value and RC value are calculated using only.

Therefore, by defining M types of "temporary ranges" according to the program data Dp, as an example, as shown in FIGS. 13 and 14, "measured value D0 200 pieces before the reference measured value D0" and "reference measured value D0" "Measured value D0 195 before the measured value D0", ... "Measured value D0 after 5995 reference measured values D0" and "Measured value D0 after 6,000 reference measured values D0" 1,240 types of start points, "measured value D0 50 after the start point", "measured value D0 55 after the start point", ... "measured value D0 3,995 after the start point" and "start point" M = 980,840 kinds of "temporary range" consisting of a combination of 791 kinds of end points with "measured value D0 after 4,000 pieces" are LC value and RC based on measured value data Dm (measured value D0). It is defined as the range in which the value is calculated.

In both figures, the end point of each "temporary range" is represented by "the number of measured values D0 from the starting point (the length of the target measured value range)". Further, although detailed description is omitted, since the "temporary range" for the measured value data Dm (measured value D0) is defined, as described above, the "temporary range" for the measured value data Dmf (measured value D0f) is defined. "Range", "Temporary range" for calculated value data Dc1 (calculated value D1), "Temporary range" for calculated value data Dc1f (calculated value D1f), "Temporary range" for calculated value data Dc2 (calculated value D2) , And the "temporary range" for the calculated value data Dc2f (calculated value D2f) are also defined in the same manner.

Next, according to the program data Dp, the processing unit 23 applies Na = 30 measured value data for each non-defective sample Xa, M = 980 for each Dm, and each measured value D0 within the “temporary range” of 840 types, and the non-defective sample Xa. Na = 30 calculated value data Dc1 M = 980, each calculated value D1 in 840 kinds of "temporary range", and Na = 30 calculated value data Dc2 for non-defective sample Xa M = 980 , The LC value and RC value, which are examples of the "discriminated value A0", the "discriminated value A1", and the "discriminated value A2", are calculated based on each calculated value D2 in the "temporary range" of 840 types. ..

Specifically, the LC value and RC value for the non-defective sample Xa can be calculated as in [Equation 1] using the pseudo inverse matrix. In addition, "v" in [Equation 1] and other mathematical formulas referred to later represents the measured value D0 (voltage value between both ends of the non-defective sample Xa), and "v" in which "・" is written in parallel above is The calculated value D1 (first-order differential value of the measured value D0) is represented, and "v" with "..." written above represents the calculated value D2 (second-order differential value of the measured value D0). In addition, "(1)", "(2)" ... "(n)" written side by side on the right side of "v" are "the first value in the temporary range" and "the second value in the temporary range". "Value" ... Indicates that it is the "nth (last) value in the provisional range". Further, "T" represents a transposed matrix.

In this case, the "temporary range" in which the measured value D0 for the specified time does not exist in the measured value data Dm, the "temporary range" in which the calculated value D1 for the specified time does not exist in the calculated value data Dc1, and When there is a "temporary range" in which the calculated value D2 for the specified time does not exist in the calculated value data Dc2, as an example, the last measured value D0 in the measured value data Dm is used as the insufficient measured value D0. , The last calculated value D1 in the calculated data Dc1 is used as the insufficient calculated value D1, and the last calculated value D2 in the calculated data Dc2 is used as the insufficient calculated value D2 to use the LC value or RC value. It is calculated or those "temporary ranges" are excluded from the candidates of the "target measurement value range" described later.

Such arithmetic processing is executed by changing the "temporary range" for each corresponding measured value data Dm. As a result, 30 × 980,840 = 29,425,200 LC values (an example of “determined value A0 for good / bad determination”) and 29,425,200 RC values (““ Good / bad judgment value A0 ”) for the non-defective sample Xa). Another example of the quality determination discriminated value A0) is calculated, and the calculation result is associated with the type of the “temporary range” and stored in the storage unit 24, respectively.

Although detailed description is omitted, each measured value data Dm, each calculated value data Dc1 and each calculated value data Dc2 are also used for each measured value data Dmf, each calculated value data Dc1f and each calculated value data Dc2f. The same arithmetic processing as the above arithmetic processing is executed respectively. As a result, 30 × 980,840 = 29,425,200 LC values (another example of “determined value A0 for quality determination”) and 29,425,200 RC values for the non-defective sample Xa. A value (another example of the discriminated value A0 for determining the quality) is calculated, and the calculation result is associated with the type of the "temporary range" and stored in the storage unit 24, respectively. Hereinafter, in order to distinguish from the LC value and RC value calculated based on the measured value data Dm, the calculated value data Dc1 and the calculated value data Dc2, based on the measured value data Dmf, the calculated value data Dc1f and the calculated value data Dc2f. The calculated LC value and RC value are also referred to as LCf value and RCf value.

^{In this case, "φ T} " in the above [Equation 1] is as shown in [Equation 2].

Therefore, "φ ^T φ" is [Equation 3].

Here, if "(φ ^T φ) ⁻ " is set to [Equation 4],

^{"(Φ T} φ) ^-1 φ ^T " in the above [Equation 1] can be expressed as [Equation 5].

Therefore, the above [Equation 1] can be expressed as [Equation 6].

By calculating this [Equation 6], [Equation 7] indicating the LC value and the RC value is derived.

This [Equation 7] includes the component of the above [Equation 6], the component corresponding to the measured value D0, the component corresponding to the calculated value D1 (the first derivative of the measured value D0), and the calculated value D2 (measurement). It is composed of components corresponding to the second derivative value of the value D0). Therefore, [Equation 7] can be expressed as [Equation 8] by transforming and rearranging each component.

In this case, as in [Equation 9], the solution (multiplication value) of the multiplication formula of the component of the first-order differential value in [Equation 8] is the LC1 value and the RC1 value, and the solution of the multiplication formula of the component of the second-order differential value (Equation 9]. It can be understood that the LC value is the sum of the LC1 value and the LC2 value, and the RC value is the sum of the RC1 value and the RC2 value, where the multiplication value) is the LC2 value and the RC2 value.

Therefore, using [Equation 9] which is a modification of [Equation 1] described above for deriving the LC value and RC value, it is recorded in each measurement value D0 and calculation value data Dc1 recorded in the measurement value data Dm. LC1 value and RC1 value (an example of "discriminated value A1"), LC2 value and RC2 value ("subject") based on each calculated value D1 and each calculated value D2 recorded in the calculated value data Dc2. An example of the discrimination value A2 ”) is calculated respectively. Such arithmetic processing is executed by changing the "temporary range" for each corresponding measured value data Dm. As a result, 30 × 980,840 = 29,425,200 LC1 values and RC1 values and 29,425,200 LC2 values and RC2 values for the non-defective sample Xa are calculated, and the calculation result is “temporary”. Each is stored in the storage unit 24 in association with the type of "range".

Although detailed description is omitted, each measured value data Dm, each calculated value data Dc1 and each calculated value data Dc2 are also used for each measured value data Dmf, each calculated value data Dc1f and each calculated value data Dc2f. The same arithmetic processing as the above arithmetic processing is executed respectively. As a result, 30 × 980, 840 = 29,425,200 LC1 values and RC1 values (“discrimination value A1 for quality determination” and other examples) and 29,425,200 pieces of the non-defective sample Xa. The LC2 value and the RC2 value (another example of the "determined value A2 for pass / fail determination") are calculated, and the calculation result is stored in the storage unit 24 in association with the type of the "temporary range".

Hereinafter, in order to distinguish the LC1 value and RC1 value, the LC2 value and the RC2 value calculated based on the measured value data Dm, the calculated value data Dc1 and the calculated value data Dc2, the measured value data Dmf, the calculated value data Dc1f and the calculation The LC1 value, RC1 value, LC2 value, and RC2 value calculated based on the value data Dc2f are also referred to as LC1f value, RC1f value, LC2f value, and RC2f value.

Next, according to the program data Dp, the processing unit 23 sets Nb = 30 measured value data for the defective sample Xb, M = 980 for each Dm, and each measured value D0 within the “temporary range” of 840 types, and the defective sample. Nb = 30 calculated value data for Xb M = 980 for each Dc1, each calculated value D1 in 840 kinds of "temporary range", and Nb = 30 calculated value data Dc2 for defective sample Xb Based on each calculated value D2 in the "temporary range" of M = 980 and 840 types, the LC value, RC value, LCf value and RCf value, which are examples of the "discriminated value B0", and the "discriminated value B1" The LC1 value, RC1 value, LC1f value and RC1f value, which are examples, and the LC2 value, RC2 value, LC2f value, and RC2f value, which are examples of the "discriminated value B2", are calculated, respectively. Since the calculation of each "discriminated value" for the defective sample Xb is the same as the calculation for each "discriminated value" for the non-defective sample Xa, detailed description thereof will be omitted. With the above, the "second process" is completed.

Subsequently, the processing unit 23 sets the "temporary range" and "temporary range" in which the degree of difference between the "determined value A0" and the "determined value B0" satisfies the "predetermined condition" according to the program data Dp. The "provisional range" in which the degree of difference between the "determined value A1" and the "determined value B1" satisfies the "predetermined condition", and the "determined value A2" and the "determined value B2" Each of the "temporary ranges" in which the degree of difference between the two is satisfied with the "predetermined condition" is specified (execution of the "third process").

Specifically, the processing unit 23 describes a "temporary range" in which the LC value for the non-defective sample Xa and the LC value for the defective sample Xb are significantly different, and the RC value for the non-defective sample Xa and the defective sample Xb. "Temporary range" where the RC value is significantly different, "Temporary range" where the LCf value for the non-defective sample Xa and the LCf value for the defective sample Xb are significantly different, and the RCf value and the defective sample for the non-defective sample Xa Each "temporary range" that is significantly different from the RCf value for Xb is specified.

Further, the processing unit 23 has a "temporary range" in which the LC1 value for the non-defective sample Xa and the LC1 value for the defective sample Xb are significantly different, and the RC1 value for the non-defective sample Xa and the RC1 value for the defective sample Xb. "Temporary range" where the LC1f value for the non-defective sample Xa and the LC1f value for the defective sample Xb are significantly different, and the RC1f value and the defective sample Xb for the non-defective sample Xa. Each "temporary range" that is significantly different from the RC1f value is specified.

Further, the processing unit 23 has a "temporary range" in which the LC2 value for the non-defective sample Xa and the LC2 value for the defective sample Xb are significantly different, and the RC2 value for the non-defective sample Xa and the RC2 value for the defective sample Xb. "Temporary range" where the LC2f value for the non-defective sample Xa and the LC2f value for the defective sample Xb are significantly different, and the RC2f value and the defective sample Xb for the non-defective sample Xa. Each "temporary range" that is significantly different from the RC2f value is specified.

More specifically, for example, when specifying a "temporary range" in which the LC value for the non-defective sample Xa and the LC value for the defective sample Xb are significantly different, for each of M = 980,840 types of "temporary range". In addition, "minimum non-defective product" which is the minimum value of Na = 30 LC values for the inspection target X, and "maximum non-defective product" which is the maximum value of Na = 30 LC values for the inspection target X. Specify each "value". In this case, as an example, when the LC value for the non-defective sample Xa and the LC value for the defective sample Xb in any of the "temporary ranges" are calculated as shown in FIG. Part 23 specifies LCmin shown in the figure as the “minimum value of non-defective product” for the non-defective sample Xa, and LCmax as the “maximum value of non-defective product” for the non-defective sample Xa. In the figure, the LC value (RC value) of the non-defective sample Xa is represented by “◯”. Further, in the same figure and FIGS. 11 and 12 referred to later, the LC value (RC value) of the defective sample Xb is represented by “●”.

Next, the processing unit 23 sets the difference between each LC value for Nb = 30 defective sample Xb and the LC value smaller than the above-mentioned "minimum non-defective product" as the "minimum non-defective product". For LC values that are calculated as "values" and are larger than the "maximum value of non-defective products", the difference from the "maximum value of non-defective products" is calculated as the "degree of defect value", and from the "minimum value of non-defective products" to the "maximum value of non-defective products". For the LC values within the range up to, the arithmetic processing for setting the “defective degree value” to “0” is executed for each “temporary range”.

Specifically, as an example, as shown in FIG. 10, the processing unit 23 has an LCmin for LCd-a, which is an LC value for a defective sample Xb smaller than LCmin, which is an example of the "minimum value for non-defective products". The difference between LCdk-a and LCdk-a is calculated as the "defectiveness degree value". Further, for LCdb, which is the LC value for the defective sample Xb larger than LCmax, which is an example of the “maximum value of non-defective product”, LCdk-b, which is the difference from LCmax, is calculated as the “defective degree value”. Further, for LCdc, which is the LC value in the LC Legend, which is in the range from LCmin to LCmax, the “defectiveness degree value” is set to “0”.

Subsequently, the processing unit 23 calculates the total value of the “defectiveness degree value” for Nb = 30 defective sample Xb according to the program data Dp for each “temporary range”. In this case, as shown in the left figure of FIG. 11, the difference from LClength, which is the range of variation in the LC value for the non-defective sample Xa, is small during the measurement process of Nb = 30 times for the defective sample Xb. Regarding the "temporary range" that includes the measured value D0 from which the LC value is calculated, even if some of the LC values with Na = 30 have a sufficiently large difference from LCmax or from LCmin. Even if it does, a small value is calculated as the total value of the "defectiveness degree value". Further, as shown in the right figure of the figure, a "temporary range" including a measured value D0 in which an LC value having a large difference from LClength is calculated during Nb = 30 measurement processes for a defective sample Xb. With respect to, even if some of the LC values of Nb = 30 are values within LClength, a large value is calculated as the total value of the “defectiveness degree value”.

Next, the processing unit 23 divides the total value of the above "defect degree value" by the difference between LCmin and LCmax, LClength (the magnitude of the variation in the LC value with respect to the non-defective sample Xa) (good sample Xa). The ratio of the total value of the "defect degree value" to the magnitude of the variation of the LC value is calculated for each "temporary range", and the calculation result is stored in the storage unit 24 in association with each "temporary range". .. At this time, as in the examples shown in the left and right figures of FIG. 12, the total value of each “defective degree value” of Nb = 30 LC values for the defective sample Xb was tentatively the same. Occasionally, a "temporary range" with a wide LClength (a "temporary range" with a large variation in the measured value D0 for a non-defective sample Xa) as shown in the example on the left has a narrower LClength as shown in the example on the right. In the "range" (the "provisional range" in which the variation of the measured value D0 for the non-defective sample Xa is small), the value obtained by dividing the total value by LClength is a larger value.

Subsequently, the processing unit 23 determines how much each "temporary range" is the "target measurement value range" based on the calculated value for each "temporary range" (value obtained by dividing the total value by LClength) according to the program data Dp. "Numerical value" (a numerical value indicating how much the LC value for the non-defective sample Xa and the LC value for the defective sample Xb have a "relationship satisfying the predetermined conditions") Identify.

Specifically, the processing unit 23 first determines the degree of difference between the distribution of Na = 30 LC values for the non-defective sample Xa and the distribution of Nb = 30 LC values for the defective sample Xb. The larger the degree of difference, the higher the numerical value. According to the "predetermined standard", each "provisional range" is quantified (scored). As a result, as shown in FIG. 13, a "numerical value" (value such as "0.0" or "0.1" in the same figure) for each "temporary range" is specified.

Next, the processing unit 23 targets one of the "temporary ranges" of M = 980,840 types, and the amount of difference from the measured value D0 corresponding to the start point in the target "temporary range" is 5 sampling cycles or less. (Example of "La = 5") It is a "temporary range" starting from the measured value D0, and the amount of difference from the measured value D0 corresponding to the end point in the target "temporary range" is 5 sampling cycles or less ( A "temporary range" whose end point is the measured value D0 of "Lb = 5") (in FIG. 13, when any "temporary range" is targeted, it is illustrated around the "temporary range". The process of specifying the "numerical value" of each of the eight "temporary ranges") and associating the minimum value of each specified "numerical value" with the target "temporary range" is performed, and each "temporary range" is "temporarily ranged". Execute as "temporary range of target".

Specifically, as an example, in the example shown in FIG. 13, the "start point" is the "measured value D0 2,000 after the measured value D0 of the reference point P0", and the "length (end point)" is the "start point". The "temporary range" of "from the measured value D0 to the 1,995th measured value D0 (measured value D0 after 1,995)" is the measured value D0 corresponding to the starting point in the "temporary range". The starting point is the measured value D0 from "after 1,995 pieces of the measured value D0 of the reference point P0" to "after 2,005 pieces of the measured value D0 of the reference point P0" in which the amount of difference is 5 sampling cycles or less. 8 of the "temporary range" that is a "temporary range" and whose end point is the measured value D0 from the "starting point measured value D0 to 1,990th" to the "starting point measured value D0 to 2,000th". The minimum of the eight "numerical values" specified for the "temporary range" is specified as the "numerical value" associated with that "temporary range".

That is, the "start point" is the "measured value D0 2,000 after the measured value D0 of the reference point P0", and the "length (end point)" is the "1,995th measured value from the measured value D0 of the start point". Regarding the "provisional range" of "up to D0 (measured value D0 after 1,995 pieces)", as shown in FIG. 14, the "starting point" is the measured value after 1,995 pieces from the measured value D0 of the reference point P0. In "D0", the "length (end point)" was specified for the "provisional range" of "from the measured value D0 of the start point to the 1,990th measured value D0 (measured value D0 after 1,990)". The "numerical value" with "1.6" is specified and associated as the minimum value.

Similarly, in the example shown in FIG. 13, the "start point" is the "measured value D0 after 2,005 measured values of the reference point P0", and the "length (end point)" is the "measured value D0 of the start point". For the "temporary range" of "from to the 2,000th measured value D0 (measured value D0 after 2,000)", the amount of difference from the measured value D0 corresponding to the starting point in the "temporary range" is 5. In the "provisional range" starting from the measured value D0 from "2,000 after the measured value D0 of the reference point P0" to "2,010 after the measured value D0 of the reference point P0", which is less than the sampling cycle. There are eight "temporary ranges" with the measured values D0 from the "starting point measured values D0 to 1,995th" to the "starting point measured values D0 to 2,005th" as the end point. The minimum of the eight "numerical values" specified for the "range" is specified as the "numerical value" associated with that "temporary range".

That is, the "start point" is the "measured value D0 after 2,005 measured values of the reference point P0", and the "length (end point)" is the "2,000th measured value from the measured value D0 of the start point". Regarding the "provisional range" of "up to D0 (measured value D0 after 2,000)", as shown in FIG. 14, the "starting point" is the measured value after 2,010 from the measured value D0 of the reference point P0. In "D0", the "length (end point)" was specified for the "provisional range" of "from the measured value D0 of the start point to the 1,995th measured value D0 (measured value D0 after 1,995)". A "numerical value" with "2.0" is specified and associated as the minimum value.

In this case, the distribution of Na = 30 LC values for the non-defective sample Xa and the distribution of Nb = 30 LC values for the defective sample Xb due to the defects actually occurring in the defective sample Xb. There is a "temporary range" in which the degree of difference from is large. In other "temporary ranges" in which the "temporary range" and the "start point" and "length (end point)" are similar, the distribution of LC values for the non-defective sample Xa and the distribution of the LC values for the defective sample Xb. The degree of difference tends to be large. On the other hand, it is not caused by defects actually occurring in the defective sample Xb, but is affected by measurement errors and noise during the measurement process for the non-defective sample Xa and during the measurement process for the defective sample Xb, and the nature of the calculation. Due to such factors, the degree of difference between the distribution of Na = 30 LC values for the non-defective sample Xa and the distribution of Nb = 30 LC values for the defective sample Xb is large. There is also. In this case, in such a "temporary range" and another "temporary range" in which the "start point" and "length (end point)" are similar, the distribution of the LC value for the non-defective sample Xa and the LC value for the defective sample Xb The degree of difference from the distribution may not be so large.

Therefore, as described above, the minimum value of the "numerical values" for other "temporary ranges" having the same "start point" and "length (end point)" is associated with the target "temporary range" (target). By replacing it with the "numerical value" of the "temporary range"), the degree of difference between the LC value distribution for the non-defective sample Xa and the LC value distribution for the defective sample Xb has increased due to the influence of measurement errors and the like. It is avoided that "a peculiar numerical value (a large numerical value that is likely not a numerical value calculated due to an actual defect)" remains associated with "".

By executing such a process for each "temporary range" of M = 980 and 840 types, as shown in FIG. 14, the distribution and error of the LC value for the "non-defective sample Xa" for each "temporary range". A “numerical value” corresponding to the “degree of difference from the distribution of the LC value” of the non-defective sample Xb is associated and stored in the storage unit 24.

Further, the processing unit 23 causes the display unit 22 to display the specific result display 31 shown in FIG. 9 based on the “numerical value” specified as described above. In the specific result display 31, each section (cell) corresponds to each section (cell) of each "temporary range" in the explanatory diagram of FIG. 14, and a large "numerical value" is specified in FIG. 14 as a "temporary range". The section corresponding to "" is darker (color closer to black), and the section corresponding to the "temporary range" in which the small "numerical value" is specified in FIG. 14 is lighter (color closer to white). Is displayed. By displaying such a specific result display 31, the "degree of difference between the LC value distribution for the non-defective sample Xa and the LC value distribution for the defective sample Xb" is greatly different in what "temporary range". It is possible to intuitively recognize whether or not to do so.

Subsequently, in the processing unit 23, the larger the associated "numerical value" is, the greater the difference between the distribution of the LC value for the non-defective sample Xa and the distribution of the LC value for the defective sample Xb (in advance). Assuming that the degree of (relationship that satisfies the specified conditions) is large, the "provisional range" that is a candidate for the "target measurement value range" is specified. Specifically, the processing unit 23 specifies, as an example, the "provisional range" having the largest associated "numerical value" as a candidate for the "target measured value range". When there are a plurality of "temporary ranges" having the largest associated "numerical value", as an example, a "temporary range" having a short corresponding LClength (that is, an LC value or the like due to a small number of constituent measured values D0, etc. Priority is given to the "temporary range"), which takes less time to calculate.

As described above, the identification of the "temporary range" in which the LC values differ greatly between the non-defective sample Xa and the defective sample Xb is completed. After that, for each "discriminated value" other than the LC value, follow the same procedure as the above procedure for the LC value, and set a "temporary range" in which the values differ greatly between the non-defective sample Xa and the defective sample Xb. Identify. As a result, the "third process" is completed.

Next, the processing unit 23 determines each "discriminated value" (LC value, RC value, LC1 value, RC1 value, LC2 value) whose values differ greatly between the non-defective sample Xa and the defective sample Xb according to the program data Dp. From each "temporary range" of RC2 value, LCf value, RCf value, LC1f value, RC1f value, LC2f value and RC2f value), it is suitable to be the "target measurement value range" when inspecting the inspection target X. Specify the "temporary range".

Specifically, as an example, the degree of difference between the LC value for the non-defective sample Xa and the LC value for the defective sample Xb is the difference between the LC1 value for the non-defective sample Xa and the LC1 value for the defective sample Xb. When the degree of difference between the LC2 value of the non-defective sample Xa and the LC2 value of the defective sample Xb is higher than the degree of the difference, the LC value as the "discriminated value C0" is used, and the above-mentioned "third" is used. The condition that the "temporary range" of the LC value specified by "processing" is set as the "target measurement value range" is specified as a candidate for the "good / bad judgment condition".

Further, the degree of difference between the LC1 value of the non-defective sample Xa and the LC1 value of the defective sample Xb is the degree of difference between the LC value of the non-defective sample Xa and the LC value of the defective sample Xb, and the non-defective sample. When the difference between the LC2 value for Xa and the LC2 value for the defective sample Xb is higher than the degree of difference, the LC1 value as the "discriminated value C1" is used and specified by the above "third process". The condition that the "temporary range" of the LC1 value is set to the "target measurement value range" is specified as a candidate for the "good / bad judgment condition".

Further, the degree of difference between the LC2 value of the non-defective sample Xa and the LC2 value of the defective sample Xb is the degree of difference between the LC value of the non-defective sample Xa and the LC value of the defective sample Xb, and the non-defective sample. When the difference between the LC1 value for Xa and the LC1 value for the defective sample Xb is higher than the degree of difference, the LC2 value as the "discriminated value C2" is used and specified by the above "third process". The condition that the "temporary range" of the LC2 value is set to the "target measurement value range" is specified as a candidate for the "good / bad judgment condition".

Similarly, the degree of difference in RC value, RC1 value and RC2 value between the non-defective sample Xa and the defective sample Xb, and the degree of difference in the LCf value, LC1f value and LC2f value between the non-defective sample Xa and the defective sample Xb. , And by comparing the degree of difference between the RCf value, the RC1f value and the RC2f value between the non-defective sample Xa and the defective sample Xb, candidates for the “good / bad determination condition” are specified.

Subsequently, the processing unit 23 identifies the candidate having the highest degree of difference in the "discriminated value" from the four specified "good / bad determination condition" candidates, and the "temporary range" of the identified candidate. Is determined as the "target measurement value range". When there are a plurality of candidates having the highest degree of difference in the "discriminated value", the processing unit 23 causes the selected candidates to be selected as the "target measurement value range" from the specified candidates, and the selected candidates are selected. The "temporary range" of is determined as the "target measurement value range". Further, the processing unit 23 records the information that can identify the determined "target measurement value range" to generate the inspection data Di, and associates the generated inspection data Di with the type (model) of the inspection target X. It is stored in the storage unit 24. As described above, the "data generation process" is completed, and the preparation for inspecting the quality of the inspection target X is completed.

In this case, as an example, when the measured value D0 in the measured value data Dm for the non-defective sample Xa is a value like the voltage waveform W0 shown in FIG. 3, the calculated value in the calculated value data Dc1 obtained by first-order differentiating the measured value D0. D1 has a value similar to that of the first-order differential waveform W1 shown in FIG. As can be understood by comparing the voltage waveform W0 and the first-order differential waveform W1, each calculated value D1 sufficiently emphasizes a slight change in each measured value D0. Therefore, the degree of difference between the good sample Xa of the LC value or RC value calculated based on the measured value D0 and the defective sample Xb is low, and it is difficult to determine the quality of the inspection target X based on the LC value or RC value. Even in such a case, there is a possibility that the inspection target X can be accurately inspected by determining the quality based on the LC1 value calculated based on the calculated value D1 and the RC1.

Further, when the measured value D0 and the calculated value data Dc of the non-defective sample Xa are values such as the above voltage waveform W0 and the first-order differential waveform W1, the calculated value D2 in the calculated value data Dc2 obtained by second-order differentializing the measured value D0. Is a value like the second-order differential waveform W2 shown in FIG. As can be understood by comparing the first-order differential waveform W1 and the second-order differential waveform W2, in each calculated value D2, a slight change in each measured value D0 is emphasized more than in each calculated value D1. Therefore, the degree of difference between the good sample Xa of the LC1 value or the RC1 value calculated based on the calculated value D1 and the defective sample Xb is low, and it is difficult to determine the quality of the inspection target X based on the LC1 value or the RC1 value. Even in such a case, there is a possibility that the inspection target X can be accurately inspected by determining the quality based on the LC2 value or RC2 value calculated based on the calculated value D2.

Further, as can be understood by comparing the voltage waveform W0 and the first-order differential waveform W1 with the second-order differential waveform W2, the voltage waveform W0 and the first-order differential waveform W1 are attenuated vibration waveforms according to the state of the non-defective sample Xa. However, the second-order differential waveform W2 has a noise-like waveform as a whole, and the components of attenuation and vibration according to the state of the non-defective sample Xa are hardly confirmed. Therefore, in each measured value D0 (voltage waveform W0) and each calculated value D1 (first-order differential waveform W1), the LC value, LC1 value, RC value, and RC1 value become large by changing the "target measurement value range". On the other hand, in each calculated value D2 (second-order differential waveform W2), the LC2 value and RC2 value may hardly change no matter how the "target measurement value range" is defined. .. Therefore, depending on the type of the inspection target X, the quality is determined based on the LC value or RC value calculated based on the measured value D0, or the LC1 value or RC1 value calculated based on the calculated value D1. There is a possibility that it can be inspected accurately.

On the other hand, in each measured value D0f, a part of noise components (components that change depending on the measurement environment) that are not directly related to the state of the inspection target X is excluded by the filtering process for each measured value D0. NS. Therefore, depending on the measurement environment, the LC value or RC value calculated based on the measured value D0 is calculated based on the measured value D0f even when it is difficult to accurately inspect the inspection target X due to the influence of noise or the like. There is a possibility that the inspection target X can be accurately inspected by determining the quality based on the LCf value and RCf.

Further, the calculated value D1f in the calculated value data Dc1f obtained by first-order differentiating the measured value D0f is a value like the first-order differential waveform W1f shown in FIG. In this case, as can be understood by comparing the first-order differential waveform W1 and the first-order differential waveform W1f described above, each calculated value D1f is calculated based on the measured value D0f from which noise components and the like have been removed by filtering processing. As a result, the slight change that can be confirmed in each calculated value D1 is sufficiently small. Therefore, in a measurement environment in which the measured value D0 changes slightly due to the influence of noise or the like, the inspection target X can be accurately determined by determining the quality based on the LC1f value or RC1f calculated based on the calculated value D1f. May be inspected.

Further, the calculated value D2f in the calculated value data Dc2f obtained by second-order differentializing the measured value D0f is a value similar to the second-order differential waveform W2f shown in FIG. In this case, as can be understood by comparing the second-order differential waveform W2 and the second-order differential waveform W2f described above, each calculated value D2f is calculated based on the measured value D0f from which noise components and the like are removed by the filtering process. As a result, the slight change due to the influence of the noise component or the like is reduced, and the damping / vibration component that cannot be confirmed by the calculated value D2 (second-order differential waveform W2) can be confirmed. As a result, by defining an appropriate "target measurement value range", the degree of difference in the "discriminated value" between the non-defective sample Xa and the defective sample Xb may be sufficiently large. Therefore, in a measurement environment in which the measured value D0 changes slightly due to the influence of noise or the like, the inspection target X is accurately determined by determining the quality based on the LC2f value or RC2f value calculated based on the calculated value D2f. May be inspected.

On the other hand, when inspecting the inspection target X by the inspection system 1, the processing unit 23 of the data processing device 3 controls the measuring device 2 according to the program data Dp, and measures the above-mentioned non-defective sample Xa and defective sample Xb. The measurement process (“measurement process C”) is executed under the same measurement conditions as at the time. At the time of the measurement process for the inspection target X, as described above, the measurement process of Nc = 5 times is executed as an example. As a result, Nc = 5 measured value data Dm (measured value data C) for the inspection target X are output from the measuring device 2.

Further, the processing unit 23 stores the measured value data Dm in the storage unit 24 each time the measured value data Dm is output from the measuring device 2, and as shown in FIG. 15, the waveform W based on the measured value data Dm is stored. Is displayed on the measurement result display screen 40 of the display unit 22. Further, the processing unit 23 indicates a target measurement value range display 41 indicating the “target measurement value range” recorded in the inspection data Di, a start point display 42a indicating the start point of the “target measurement value range”, and an end point. An example of "specific result notification processing" in which the end point display 42b and the end point display 42b are displayed overlaid on the waveform W ("notify by displaying the target measurement value range corresponding to the waveform of the measurement value data C"": shown in the figure. Display of broken lines and numerical values).

As a result, the user who sees the target measurement value range display 41, the start point display 42a, and the end point display 42b can use the "discriminated value" (LC value, RC value, LC1 value, etc.) based on any of the measured values D0 in the waveform W. The RC1 value, LC2 value, RC2 value, LCf value, RCf value, LC1f value, RC1f value, LC2f value, RC2f value, etc.) are calculated to recognize whether the quality of the inspection target X is inspected.

Next, the processing unit 23 includes the measured values D0, D0 ... Recorded in the measured value data Dm in the "target measured value range" recorded in the inspection data Di according to the program data Dp. The measured values D0, D0 ... Are specified, and based on the specified measured values D0, D0 ..., the LC value, RC value, LCf value, RCf value (discriminated value C0), LC1 value, RC1 value, and LC1f The process of calculating the "discriminated value" specified by the inspection data Di among the value and RC1f value (discriminated value C1), and the LC2 value and RC2 value, LC2f value and RC2f value (discriminated value C2). , Execute for each measured value data Dm.

Subsequently, the processing unit 23 determines that the calculated “discriminated value” is ± 5 with respect to LClength for the non-defective product sample Xa corresponding to the “provisional range” defined as the predetermined non-defective product range (for example, the “target measured value range””. It is determined whether or not it is included in the% range). At this time, when all of the five "discriminated values" calculated based on Nc = 5 measured value data Dm are within the non-defective product range, the processing unit 23 determines the inspection target X as a non-defective product. .. Further, when one or more of the five "discriminated values" is a value outside the non-defective product range, the processing unit 23 determines the inspection target X as a defective product. With the above, the quality inspection of the inspection target X is completed.

As described above, in the data processing apparatus 3, in the "data generation processing" in which the processing unit 23 generates the inspection data Di capable of specifying the "good / bad determination condition" of the inspection target X, each measurement of the non-defective sample Xa is performed. A plurality of calculated values D1 obtained by first-order differentiating the value D0 and a plurality of calculated values D2 obtained by second-order differentiating each measured value D0 are derived for each measured value data Dm, and each measured value D0 for the defective sample Xb. The "first process" for deriving a plurality of calculated values D1 obtained by first-order differentiation and a plurality of calculated values D2 obtained by second-order differentializing each measured value D0 for each measured value data Dm, and each of the non-defective sample Xa. The LC value and RC value for quality determination are calculated for each measured value data Dm corresponding based on the measured value D0 and the respective calculated values D1 and D2, and the corresponding calculated value D1 is calculated for the non-defective sample Xa. The LC1 value and RC1 value for quality determination are calculated for each measurement value data Dm, and the LC2 value for quality determination and the LC2 value for quality determination are calculated for each measurement value data Dm corresponding based on each calculation value D2 for the non-defective sample Xa. The RC2 value is calculated, and the LC value and RC value for quality determination are calculated for each measurement value data Dm corresponding to each measurement value D0 and each calculation value D1 and D2 for the defective sample Xb. , LC1 value and RC1 value for pass / fail determination are calculated for each corresponding measured value data Dm based on the calculated value D1 for the defective product sample Xb, and based on each calculated value D2 for the defective product sample Xb. The "second process" for calculating the LC2 value and RC2 value for quality determination for each corresponding measured value data Dm is executed, and the LC value and RC value for the non-defective sample Xa and the defective sample Xb are obtained. The degree of difference between the LC value and RC value, the degree of difference between the LC1 value and RC value 1 for the non-defective sample Xa and the LC1 value and RC value 1 for the defective sample Xb, and the LC2 value for the non-defective sample Xa. The inspection data Di is generated by defining one of the "good / bad judgment conditions" based on the degree of difference between the RC value 2 and the LC2 value and the RC value 2 for the defective sample Xb. ..

Specifically, in this data processing device 3, the processing unit 23 determines the degree of difference between the LC value and RC value of the non-defective sample Xa and the LC value and RC value of the defective sample Xb with respect to the non-defective sample Xa. The degree of difference between the LC1 value or RC value 1 and the LC1 value or RC value 1 for the defective sample Xb, and the LC2 value or RC value 2 for the non-defective sample Xa and the LC2 value or RC value for the defective sample Xb. When the degree of difference from 2 is higher, each measured value D0 in the "measured value range" and each measured value D0 in the "measured value range" in each measured value data Dm for the inspection target X are first-order differentiated. Based on the plurality of calculated values D1 and the plurality of calculated values D2 obtained by second-order differentiation of each measured value D0 in the "measured value range", the LC value or RC value for pass / fail determination is calculated and based on the LC value or RC value. The condition for determining the quality of the inspection target X is defined as one of the "quality determination conditions", and the LC1 value and RC value 1 for the non-defective sample Xa and the LC1 value and RC value 1 for the defective sample Xb are defined. The degree of difference is the degree of difference between the LC value or RC value for the non-defective sample Xa and the LC value or RC value for the defective sample Xb, and the LC2 value or RC value 2 for the non-defective sample Xa and the defective sample Xb. When the degree of difference from the LC2 value or RC value 2 is higher than the degree of difference between The condition for judging the quality of the inspection target X is defined as one of the "good / bad judgment conditions", and the LC2 value and RC value 2 for the good sample Xa and the LC2 value and RC value 2 for the defective sample Xb are defined. The degree of difference is the degree of difference between the LC value or RC value for the non-defective sample Xa and the LC value or RC value for the defective sample Xb, and the LC1 value or RC value 1 for the non-defective sample Xa and the defective sample Xb. When the degree of difference from the LC1 value or RC value 1 is higher than the degree of difference from the LC1 value or RC value 1, the LC2 value or RC2 value for pass / fail determination is calculated based on each calculated value D2 for the inspection target X, and based on the LC2 value or RC2 value. The condition for determining the quality of the inspection target X is defined as one of the “quality determination conditions”, and the inspection data Di is generated.

Further, in this inspection system 1, "measurement process A", "measurement process B" and "measurement process C" are executed for the non-defective product sample Xa, the defective product sample Xb and the inspection target X, and the measured value data A and the measured value are executed. The measuring device 2 that generates the data B and the measured value data C and the "data generation process" are executed to generate the inspection data Di, and the generated inspection data Di and the measured value data Dm for the inspection target X are generated. A data processing device 3 for inspecting the quality of the inspection target X based on the above is provided so that the quality of the inspection target X can be inspected. Further, in this program data Dp, the processing unit 23 of the data processing apparatus 3 is made to execute the above-mentioned "data generation processing".

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, a kind of defect may occur in which the degree of difference between the "discriminated value A0" and the "discriminated value B0" is sufficiently high. For the inspection target X, the inspection data in which the "good / bad judgment condition" for determining the quality by calculating the "discrimination value C0" based on each measurement value D0 in the measurement value data Dm for the inspection target X is defined. Di is generated, and the degree of difference between the "discriminated value A0" and the "discriminated value B0" is not so high. For the inspection target X, which may cause a kind of defect, each of the measurement value data Dm for the inspection target X Calculated based on the "discriminated value C1" calculated based on the calculated value D1 obtained by first-order differentiating the measured value D0, and the calculated value D2 obtained by second-order differentiating each measured value D0 in the measured value data Dm for the inspection target X. Since the inspection data Di that defines the "good / bad judgment condition" for determining the quality based on the "determined value C2" is generated, the inspection is executed based on this inspection data Di. It is possible to accurately determine the non-defective products of various inspection targets X.

Further, in the data processing device 3, the processing unit 23 performs the "data generation processing" from any of the measured values D0 recorded in the measured value data Dm at the time of inspection of the inspection target X. The "target measurement value range" that can specify which of the measurement values D0 up to the measurement value D0 should be the measurement value D0 within the "measurement value range" is defined as one of the "good / bad judgment conditions". Sometimes, M types (in this example, measured value data) in which the combination of the measured value D0 corresponding to the start point of the "target measured value range" and the measured value D0 corresponding to the end point of the "target measured value range" are at least different. "Temporary range" of 980,840 types for each of Dm and calculated value data Dc1 and Dc2) is defined, and each "temporary range" is set as "measured value range" and "first process" and "second process" are performed. In addition to executing each of the M types of "provisional range", the degree of difference between the LC value and RC value for the non-defective sample Xa and the LC value and RC value for the defective sample Xb, the LC1 value for the non-defective sample Xa, and so on. The degree of difference between the RC1 value and the LC1 value or RC value 1 for the defective sample Xb, and the degree of difference between the LC2 value or RC2 value for the non-defective sample Xa and the LC2 value or RC value 2 for the defective sample Xb. Is specified for each of the M types of "temporary range", and the "third process" for specifying the "temporary range" in which the degree of difference satisfies the predetermined condition is executed, and the specified "temporary range" is executed. The inspection data Di is generated with the "range" as the "target measurement value range".

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, each measurement recorded in the measurement value data Dm for the inspection target X according to the type of the inspection target X and the difference in the measurement environment. Among the values D0, D0 ..., the LC value, LC1 value and LC2 value and RC value, RC1 value and RC2 value for the non-defective sample Xa, and the LC value, LC1 value and LC2 value and RC for the defective sample Xb. Since the "target measurement value range" that has a significantly different relationship between the value, RC1 value, and RC2 value is automatically specified and the inspection data Di is generated, the inspection target X is determined based on this inspection data Di. By inspecting, it is possible to accurately determine the quality of the inspection target X, which may cause various defects.

Further, in the data processing device 3, in the "third processing", the processing unit 23 has out of Na "discriminated values A0 (LC value and RC value)" for each "temporary range" of M types. The "minimum value of non-defective product", which is the minimum value of For each "RC value)", for the "discriminated value B0" that is smaller than the "minimum value of non-defective product", the difference from the "minimum value of non-defective product" is calculated as the "degree of defect value", and is larger than the "maximum value of non-defective product". For the "discriminated value B0", the difference from the "maximum non-defective product value" is calculated as the "defectiveness degree value", and the total value of each "defectiveness degree value" of Nb "discriminated value B0" is calculated. , The process of specifying the "temporary range" that satisfies the condition that satisfies the predetermined condition based on the calculated total value, and the Na "determined value A1 (LC1 value or LC1 value") for each of the M types of "temporary range". The "minimum value of non-defective product" which is the minimum value of "RC1 value)" and the "maximum value of non-defective product" which is the maximum value of Na "identified value A1" are specified, and Nb "identified product" is identified. For each value B1 (LC1 value or RC1 value), for the "discriminated value B1" that is smaller than the "minimum value of non-defective product", the difference from the "minimum value of non-defective product" is calculated as the "degree of defect value", and "good product". For the "discriminated value B1" that is larger than the "maximum value", the difference from the "non-defective product maximum value" is calculated as the "defectiveness degree value", and each "defectiveness degree value" of Nb "discriminated value B1" is calculated. A process of calculating the total value of and specifying a "temporary range" that satisfies a predetermined condition based on the calculated total value, and Na "discriminated" for each of the M types of "temporary range". The "minimum value of non-defective product" which is the minimum value of the "value A2 (LC2 value or RC2 value)" and the "maximum value of non-defective product" which is the maximum value of the "identified value A2" of Na are specified. For each Nb "discriminated value B2 (LC2 value or RC2 value)", for the "discriminated value B2" smaller than the "non-defective product minimum value", the difference from the "non-defective product minimum value" is used as the "defectiveness degree value". For the "discriminated value B2" that is calculated and is larger than the "maximum value of non-defective product", the difference from the "maximum value of non-defective product" is calculated as the "degree of defect value", and Nb "discriminated value B2" The total value of each "defect degree value" is calculated, and a process of specifying a "temporary range" that satisfies a predetermined condition based on the calculated total value is executed.

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, it is possible to eliminate the influence of the variation of the "discriminated value" that may occur even in the non-defective sample Xa. The inspection data Di can be generated by preferably specifying a suitable "temporary range".

Further, according to the data processing device 3, the inspection system 1, and the program data Dp, the processing unit 23 sets the total value by the difference between the "minimum value of non-defective product" and the "maximum value of non-defective product" in the "third processing". By specifying the "provisional range" that satisfies the conditions specified in advance based on the divided values, the "discriminated value B0" has the same difference from the "minimum value of non-defective product" and "maximum value of non-defective product". , Regarding the "provisional range" that becomes the "discriminated value B1" and the "discriminated value B2", the "temporary range" in which the difference between the "minimum value of non-defective product" and the "maximum value of non-defective product" is large ("subject" for the non-defective product sample Xa The "provisional range" in which the difference between the "minimum value of non-defective product" and the "maximum value of non-defective product" is smaller than that of the "provisional range" in which the "discrimination value A0", "discriminated value A1" and "discriminated value A2" vary widely. (A "temporary range" in which the variation of "discriminated value A0", "discriminated value A1", and "discriminated value A2" for the non-defective sample Xa is small) is calculated, and therefore the non-defective sample Xa is calculated. "Discriminated value A0", "discriminated value A1" and "discriminated value A2" have a large difference with respect to the range of variation of "discriminated value B0", "discriminated value B1" and "discriminated value B2". The inspection data Di can be generated by preferably specifying the "temporary range" that can be calculated, that is, the "temporary range" that can more reliably determine the quality of the inspection target X.

Further, in the data processing device 3, the processing unit 23 determines the degree of difference between the distribution of Na "discriminated values A0" and the distribution of Nb "discriminated values B0" in the "third processing". Is quantified for each of M "provisional ranges" according to a predetermined standard, which becomes a higher value as the degree of difference increases, and the distribution of Na "discriminated values A1" and Nb "discriminated values" are quantified. The degree of difference from the distribution of "B1" is quantified for each of M "provisional ranges" according to a predetermined standard, and the distribution of Na "discriminated values A2" and Nb "discriminated values B2" are quantified. The degree of difference from the distribution of "" is quantified for each of M "temporary ranges" according to a predetermined standard, and the "temporary range" that satisfies the predetermined conditions based on those "numerical values" To identify.

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, the distribution of the "discriminated value A0", the "discriminated value A1", and the "discriminated value A2" for the non-defective sample Xa and the defective product It is possible to accurately identify a "provisional range" in which the distribution of the "discriminated value B0", the "discriminated value B1", and the "discriminated value B2" for the sample Xb is large.

Further, in the data processing device 3, the processing unit 23 measures one of the M types of "temporary ranges" in the "third process" and corresponds to the start point in the target "temporary range". The amount of difference from the value D0 is the "temporary range" starting from the measured value D0 that is less than or equal to the La sampling cycle, and the amount of difference from the measured value D0 corresponding to the end point in the target "temporary range" is the Lb sampling cycle. The measured value D0 in which the difference between the numerical value of the non-defective sample Xa in the "temporary range" having the following measured value D0 as the end point and the measured value D0 corresponding to the starting point in the target "temporary range" is less than or equal to the La sampling cycle. Defective product in the "temporary range" that is the "temporary range" as the start point and whose end point is the measured value D0 whose end point is the measured value D0 whose amount of difference from the measured value D0 corresponding to the end point in the target "temporary range" is less than or equal to the Lb sampling cycle. The process of identifying each of the numerical values for the sample Xb and associating the minimum value among the specified numerical values with the target "temporary range" is executed with M "temporary ranges" as the target "temporary ranges". , Assuming that the larger the associated numerical value is, the greater the degree of the relationship satisfying the predetermined condition is, the "temporary range" that satisfies the predetermined condition is specified.

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, the “determined value B0” is caused by the sudden deviation of the measured value D0 that occurs during the “measurement process B” of the defective sample Xb. , "Temporary range" such as "Discriminated value B1" and "Discriminated value B2" are not specified as "Target measured value range", and their start and end points are the same. Even in the "temporary range", the "provisional range" in which the "discriminated value B0", the "discriminated value B1" and the "discriminated value B2" are sufficiently large values, that is, the constant occurrence in the defective sample Xb The inspection data Di can be generated by specifying the "provisional range" including the measured value D0, which is different from the measured value D0 for the non-defective sample Xa due to the influence of defects, as the "target measured value range". ..

Further, in the data processing device 3, the processing unit 23 periodically changes the measured value D0 to a state in which the measured value D0 exceeds the predetermined threshold value and a state in which the measured value D0 falls below the predetermined threshold value within a predetermined time. The measured value data Dm for the changing non-defective sample Xa and the measured value data Dm for the defective sample Xb are acquired, and in the "data generation process", each measured value for a predetermined time for the non-defective sample Xa is obtained. The Kth measurement value D0 exceeding the predetermined threshold value in the Kth time of D0, and the Kth measurement value D0 of each measurement value D0 for the predetermined time for the defective sample Xb are predetermined. The measured value D0 that exceeds the threshold value, the measured value D0 that is below the predetermined threshold value at the Kth time of each measured value D0 for the predetermined time for the non-defective sample Xa, and the defective sample. Of the respective measured values D0 for the predetermined time for Xb, the M type of "provisional" is based on one of the measured values D0 below the predetermined threshold value at the Kth time, whichever is predetermined. Specify the "range".

Therefore, according to the data processing device 3, the inspection system 1, and the program data Dp, the starting point of the measured value data Dm at the time of "measurement process A" for the non-defective sample Xa and at the time of "measurement process B" for the defective sample Xb. Even if there is a slight deviation in the measurement value D0 that satisfies the same criteria in all the measurement value data Dm (the measurement value D0 that exceeds the predetermined threshold value at the Kth time, or the Kth time). Since the starting point of the "provisional range" can be specified with reference to the measured value D0) below the predetermined threshold value, the starting point in the measured value data Dm for the good sample Xa and the defective sample Xb is small. It is possible to eliminate the influence of the deviation and specify the "temporary range" suitable for the "target measurement value range".

Further, according to the data processing device 3, the inspection system 1, and the program data Dp, the processing unit 23 executes the "specific result notification process" for notifying the "target measurement value range" specified by the "third process". By doing so, it is possible to make the user surely and easily recognize whether or not the inspection target X is inspected for the measured value D0 included in the range.

Further, according to the data processing device 3, the inspection system 1 provided with such a data processing device 3, and the program data Dp for executing such processing, the processing unit 23 performs the "specific result notification processing" in the "specific result notification processing". At least, a waveform based on at least one of the measured value data Dm for the non-defective sample Xa, the measured value data Dm for the defective sample Xb, and the measured value data Dm for the inspection target X is displayed on the display unit 22. By notifying the user by displaying the "target measurement value range" corresponding to the waveform, the user is informed of what range the measurement value D0 included in is to be inspected for the quality of the inspection target X. It can be recognized more easily.

Further, according to the data processing device 3, the inspection system 1, and the program data Dp, the processing unit 23 extracts each measured value D0 recorded in the measured value data Dm for the non-defective sample Xa every other J1. Using each extracted measured value D0, each calculated value D1 corresponding to each extracted measured value D0, and each calculated value D2 corresponding to each extracted measured value D0, "discriminated value A0" and "discriminated value" "A1" and "discriminated value A2" are calculated respectively, and each measured value D0 recorded in the measured value data Dm for the defective sample Xb is extracted every two J2, and each extracted measured value D0 is extracted. Using each calculated value D1 corresponding to each measured value D0 and each calculated value D2 corresponding to each extracted measured value D0, "discriminated value B0", "discriminated value B1", and "discriminated value B2" are used. By calculating each of the above, the "discriminated value A0", the "discriminated value A1", the "discriminated value A2", and the "discriminated value B0" are targeted for all the measured values D0 included in the "temporary range". , Compared with the case of calculating the "determined value B1" and the "determined value B2", the load on the processing unit 23 and the calculation result are temporarily set because the number of the measured values D0 to be calculated is smaller. The capacity of the memory to be stored can be sufficiently reduced.

The configurations of the "inspection data generation device" and the "inspection system" and the processing procedure described in the "inspection data generation processing program" are the configurations of the above inspection system 1 (data processing device 3). It is not limited to an example or an example of description of program data Dp.

For example, the degree of difference between the LC value or RC value of the non-defective sample Xa and the LC value or RC value of the defective sample Xb (the degree of difference between the “discriminated value A0” and the “discriminated value B0”). The degree of difference between the LC1 value and RC1 value of the non-defective sample Xa and the LC1 value and RC1 value of the defective sample Xb (the degree of difference between the "discriminated value A1" and the "discriminated value B1"), and the non-defective product. The most difference between the LC2 value and RC2 value of sample Xa and the LC2 value and RC2 value of defective sample Xb (the degree of difference between "discriminated value A2" and "discriminated value B2"). LC value or RC value (discriminated value C0), LC1 value or RC1 value (discriminated value C1), and LC2 value or RC2 value at the time of inspection of the inspection target X corresponding to a high degree of "discriminated value". An example of generating inspection data Di by specifying which of (discriminated value C2) is used has been described, but the inspection target X is selected by using the “discriminated value” having the second highest degree of difference. It is also possible to adopt a configuration / method for generating inspection data Di for inspection.

Further, in the "data generation process", the number of executions of "measurement process A" targeting the non-defective sample Xa = Na and the number of executions of "measurement process B" targeting the defective sample Xb = Nb are the same number ( In this example, an example in which Na = Nb = 30) has been described, but the number of times each measurement process is executed can be any multiple times other than 30 times, and the execution of "measurement process A" can be performed. The number of times and the number of times the "measurement process B" is executed can be different from each other.

In this case, the Na "measurement value data A" generated by the "measurement process A" of Na times is not limited to the data generated by the measurement process of Na times for one non-defective sample Xa. , Data generated by a total of Na measurements for a plurality of non-defective samples Xa can be used. Similarly, the Nb "measured value data B" generated by the Nb "measurement process B" is limited to the data generated by the Nb measurement process for one defective sample Xb. However, the data generated by the measurement process of a total of Nb times for a plurality of defective sample Xb can be used. When using the data generated by the total Na measurement processing for a plurality of non-defective sample Xa or the total Nb measurement process for a plurality of defective samples Xb, each non-defective sample It is preferable to perform the measurement process for each Xa a plurality of times and the measurement process for each defective sample Xb a plurality of times.

Further, the number of executions of "measurement process A" in the "data generation process" = Na, the number of executions of "measurement process B" = Nb, and the number of executions of "measurement process C" at the time of inspection of the inspection target X = Nc. In this example, an example in which the numbers are different from each other (Na = Nb = 30: Nc = 5) has been described. The number of times can be the same as either (or both) of "measurement process B".

Further, the number of types of the "temporary range" specified in the "data generation process" is not limited to M = 980 and 840 types, and any plurality of types of M = 2 or more are specified and the "target measurement value range" is specified. Can be a candidate for. Further, a configuration / method for calculating each "discriminated value" with the first measured value D0 to the last measured value D0 recorded in the measured value data Dm as the "measured value range", that is, the "temporary range" or It is also possible to adopt a configuration / method for calculating the "discriminated value" without specifying the "target measured value range".

Further, when calculating the "discriminated value A0", the "discriminated value A1" and the "discriminated value A2", and the "discriminated value B0", the "discriminated value B1" and the "discriminated value B2", the measured values An example of extracting and using the measured value D0 every 5 J1 = J2 = 5 from the measured values D0, D0 ... Recorded in the data Dm has been described, but the measured value used for the calculation of the "discriminated value" The extraction interval of D0 (the value of J1 and the value of J1) can be any natural number other than "5", and the value of J1 and the value of J2 can be different numbers from each other. .. Further, in the calculation of the "discriminated value", all the measured values D0, D0 ... In the "temporary range" or all the measured values D0, D0 ... Recorded in the measured value data Dm are used. You can also do it.

In addition, when specifying the minimum value from the "numerical values" of the "temporary range" with the same start point and end point for any "temporary range", the start point or end point for the target "temporary range" An example of targeting a "temporary range" in which the end points and the like differ by La = Lb = 1 sampling period has been described, but the La value and the Lb value at the time of this processing are arbitrary numbers other than "1". It can also be set to La ≠ Lb.

Further, the measured value D0 that exceeds the "threshold value (0V in this example)" at the beginning (K = 1st time) of the measured values D0, D0 ... Recorded in the measured value data Dm is used as a reference. As an example, an example of specifying the start point of the "temporary range" has been described, but the Kth time is not limited to the first time (first time), and can be any number of times of the second time or more. It is also possible to specify the start point of the "temporary range" with reference to the measured value D0 that is lower than the "Kth threshold value" among the measured values D0, D0 ... Recorded in the measured value data Dm. .. In this case, the Kth time is not limited to the first time (first time), and can be any number of times of the second time or more. Further, instead of the "measured value below the threshold value" or the "measured value exceeding the threshold value", the minimum value of the measured values D0, D0 ... Recorded in the measured value data Dm or The maximum value can also be the starting point of the "temporary range".

Further, for the "discriminated value B0 (LC value, RC value)" smaller than the "minimum non-defective product (LCmin, RCmin)", the difference from the "minimum non-defective product value" is set as the "defective degree value" and the "maximum non-defective product". Regarding the "discriminated value B0" which is larger than the "value (LCmax, RCmax)", an example in which the difference from the "maximum non-defective product value" is set as the "defective degree value" has been described. The difference between the average value or the center of gravity value of "" and the "determined value B0" for the defective sample Xb can also be used as the "defective degree value". Similarly, the difference between the average value or the center of gravity value of the "discriminated value A1" for the non-defective sample Xa and the "discriminated value B1" for the defective sample Xb can be regarded as the "defective degree value", or the non-defective sample Xa can be used. The difference between the average value or the center of gravity value of the "determined value A2" and the "determined value B2" for the defective sample Xb can also be used as the "defectiveness degree value". In addition, following an example of specifying a "provisional range that satisfies a predetermined condition" based on a value obtained by dividing the "total defect degree value" by the "difference between the minimum non-defective product value and the maximum non-defective product value". As described above, it is also possible to specify a "temporary range" that satisfies the conditions based on the "total value of the degree of defects" itself.

Further, "discriminated value A0", "discriminated value A1" and "discriminated value A2", and "discriminated value B0", "discriminated value B1" and "discriminated value B2" are LC values and LC1. Not limited to the value and LC2 value, RC value, RC1 value and RC2 value, LC value, LC1 value, LC2 value, RC value, RC1 value, RC2 value multiplied by an arbitrary coefficient, LC value as RC value Any value can be used, such as the divided value, the LC1 value divided by the RC1 value, and the LC2 value divided by the RC2 value.

Further, the measured values D0, D0 ... Of the measured value data Dm acquired from the measuring device 2 are used as they are to calculate the "discriminated value", and the acquired measured value data Dm is subjected to arbitrary filtering processing (arbitrary value). (Processing to exclude measured values below or above an arbitrary value) is executed, and each "discriminated value" is calculated using the measured value D0f after filtering processing. As described above, a configuration / method in which the "discriminated value" is calculated by using the measured values D0, D0 ... Of the measured value data Dm acquired from the measuring device 2 as they are, and the filtering process is not performed is adopted. You can also.

In addition, the "provisional range" in which the associated "numerical value" is the largest (the "provisional range" in which the "discriminated value" for the non-defective sample Xa and the "discriminated value" for the defective sample Xb are the most different). Was specified as a candidate for the "target measurement value range", and an example of specifying the "target measurement value range" from the specified "provisional range" was explained. Not limited to such an example, the "provisional range" in which the degree of difference between the "discriminated value" for the non-defective sample Xa and the "discriminated value" for the defective sample Xb is the second or later is the "target measured value range". It can also be specified as a candidate for.

Further, at the time of inspection of the inspection target X, the target measurement value range display 41, the start point display 42a, and the end point display 42b are displayed in correspondence with the waveform W (the waveform based on the measurement value data Dm for the inspection target X) to "target". Although the "specific result notification process" for notifying the "measured value range" has been described as an example, the measured value data Dm for the non-defective sample Xa and the measured value data for the defective sample Xb prior to the inspection for the inspection target X have been described. Execution of processing to display an arbitrary display (target measurement value range display 41, start point display 42a, end point display 42b, etc.) that specifies the "target measurement value range" corresponding to the waveform based on Dm is executed as "specific result notification processing". You can also do it.

Further, at the time of inspection of the inspection target X, the waveform based on the measurement value data Dm for the inspection target X is displayed, and the waveform based on the measurement value data Dm for the non-defective sample Xa and the measurement value data Dm for the defective sample Xb are displayed. It is also possible to execute a process of displaying at least one of the based waveforms and displaying an arbitrary display for specifying the "target measurement value range" corresponding to those waveforms as the "specific result notification process". Further, the "target measurement value range" is not limited to the process of displaying the target measurement value range display 41, the start point display 42a, the end point display 42b, and the like, and information that can specify the "target measurement value range" (the position of the start point and the position of the start point). It is also possible to perform a process of outputting the end point position) as a voice message.

Further, an example of displaying information that can specify the "target measured value range" on the display unit 22 which is a component of the data processing device 3 in the inspection system 1 as the "inspection system" has been described, but the display as an external device has been described. It is also possible to display information that can identify the "target measurement value range" on the device. Further, an example in which the data processing device 3 functions as an "inspection data generation device" and an "inspection device" has been described, but from the "inspection data generation device" that generates the inspection data Di and the "measuring device". An "inspection system" can also be configured by separately providing an "inspection device" that inspects the "inspection target" based on the "measured value data C" and the "inspection data". Further, the "inspection data generator", the "measuring device" and the "inspection device" are configured by one device (the element corresponding to the "inspection data generator" in the "inspection device", and the "measuring device". It is also possible to integrally provide an element corresponding to).

In addition, an example of inspecting a wound component as an "inspection target" has been described, but the "inspection target" for inspecting quality by an "inspection system" is not limited to this, and various electronic devices such as capacitors and resistors are used. It is possible to inspect between a component and any inspection point on the circuit board as an "inspection target".

1 Inspection system 2 Measuring device 3 Data processing device 11 Measurement signal generator 12 A / D conversion unit 13, 23 Processing unit 14, 24 Storage unit 21 Operation unit 22 Display unit 31 Specific result display 40 Measurement result display screen 41 Target measurement value Range display 42a Start point display 42b End point display D0 Measured value D1, D2 Calculated value Dc1, Dc2 Calculated value data Di Inspection data Dm Measured value data Dp Program data P0 Reference point P1, P1a Start point a
P2, P2a End point t0, t1, t1a, t2, t2a, te, ts Time point T1, T1a, T2, T2a, TA time W, W0 to W2, W0f to W2f Waveform X Inspection target Xa Good sample Xb Defective sample

Claims (13)

Na times (Na is a natural number of 2 or more) measured over a predetermined time in a predetermined sampling cycle for a non-defective sample of the same type as the inspection target for the predetermined time. Nb times performed under the same measurement conditions as the measurement process A for the Na measurement value data A in which the plurality of measurement values a0 of the above are recorded and the defective sample of the same type as the inspection target and the good sample. By the measurement process B of (Nb is a natural number of 2 or more), Nb of measurement value data B in which a plurality of measurement values b0 for the predetermined time are recorded are acquired, and the inspection of the inspection target is performed. A plurality of measured values c0 for the predetermined time by the Nc times (Nc is a natural number of 2 or more) of the measurement process C, which is sometimes performed on the inspection target under the same measurement conditions as the measurement process A and the measurement process B. Data generation process for generating inspection data that can identify the pass / fail judgment condition of the inspection target based on Nc measurement value data C recorded in each of the measurement value data A and the measurement value data B. Equipped with a processing unit that executes
In the data generation process, the processing unit
A plurality of calculated values a1 obtained by first-order differentiating each measured value a0 and a plurality of calculated values a2 obtained by second-order differentiating each measured value a0 are derived for each of the measured value data A, and each measured value b0. The first process of deriving a plurality of calculated values b1 obtained by first-order differentiation of the above, and a plurality of calculated values b2 obtained by second-order differentiation of each measured value b0 for each measured value data B,
Corresponds based on the respective measured values a0 within the predetermined measured value range, the respective calculated values a1 within the calculated value range corresponding to the measured value range, and the respective calculated values a2 within the calculated value range. Each of the measured value data A is calculated with a pass / fail determination value A0, and each of the corresponding measurement value data A is subject to pass / fail determination based on the calculated value a1 within the calculated value range. The discrimination value A1 is calculated respectively, and the judgment value A2 for pass / fail judgment is calculated for each measurement value data A corresponding to each calculation value a2 within the calculation value range, and the measurement value is calculated. For each measurement value data B corresponding to each measurement value b0 in the range, each calculation value b1 in the calculation value range, and each calculation value b2 in the calculation value range for pass / fail determination. The discriminated value B0 is calculated respectively, and the discriminated value B1 for pass / fail judgment is calculated for each of the corresponding measured value data B based on the calculated value b1 within the calculated value range, and the calculated value is calculated. A second process of calculating the discriminated value B2 for pass / fail determination for each measured value data B corresponding to each of the calculated value b2 within the range is executed.
The degree of difference between the discriminated value A0 and the discriminated value B0, the degree of difference between the discriminated value A1 and the discriminated value B1, and the difference between the discriminated value A2 and the discriminated value B2. An inspection data generator that specifies one of the pass / fail determination conditions based on the degree and generates the inspection data. In the data generation process, the processing unit determines that the degree of difference between the discriminated value A0 and the discriminated value B0 is the degree of difference between the discriminated value A1 and the discriminated value B1 and the discriminated value. When the degree of difference between the value A2 and the discriminated value B2 is higher than the degree of difference, the measured value c0 within the measured value range and the measured value c0 within the measured value range in each measured value data C are set as one. The discriminated value C0 for pass / fail judgment is calculated based on the plurality of calculated values c1 obtained by the degree differentiation and the plurality of calculated values c2 obtained by second-order differentiating the measured values c0 within the measured value range. The condition for determining the quality of the inspection target based on the above is defined as one of the quality determination conditions.
The degree of difference between the discriminated value A1 and the discriminated value B1 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the difference between the discriminated value A2 and the discriminated value B2. When the degree is higher than the above-mentioned degree, the condition that the quality of the inspection target is determined based on the discriminated value C1 by calculating the discriminated value C1 for pass / fail determination based on each calculated value c1 is the condition for determining the quality. Defined as one of
The degree of difference between the discriminated value A2 and the discriminated value B2 is the degree of difference between the discriminated value A0 and the discriminated value B0, and the difference between the discriminated value A1 and the discriminated value B1. When the degree is higher than the above-mentioned degree, the condition that the quality of the inspection target is determined based on the judgment value C2 by calculating the judgment value C2 for judgment based on each calculation value c2 is the condition for determining the quality. The inspection data generation device according to claim 1, wherein the inspection data is generated by defining it as one of the above. In the data generation process, the processing unit performs any of the measured values c0 recorded in the measured value data C at the time of inspection of the inspection target from any of the measured values c0 to any other measured value c0. When defining the target measurement value range in which it is possible to specify whether each measurement value c0 up to the above should be the measurement value c0 within the measurement value range as another one of the quality determination conditions,
M types in which the combination of the measured value a0 and the measured value b0 corresponding to the start point of the target measured value range and the measured value a0 and the measured value b0 corresponding to the end point of the target measured value range are at least different ( M defines a tentative range of 2 or more natural numbers), and the first process and the second process are executed for each tentative range of the M type, respectively, with each tentative range as the measured value range. The degree of difference between the discriminated value A0 and the discriminated value B0, the degree of difference between the discriminated value A1 and the discriminated value B1, and the difference between the discriminated value A2 and the discriminated value B2. The degree is specified for each of the M types of provisional ranges, and a third process for specifying the provisional range in which the degree of difference satisfies the predetermined condition is executed.
The inspection data generation device according to claim 1 or 2, wherein the specified provisional range is used as the target measurement value range to generate the inspection data. In the third process, the processing unit
For each of the M types of provisional ranges, the minimum value of the non-defective product, which is the minimum value of the Na discriminated values A0, and the maximum value of the non-defective product, which is the maximum value of the Na discriminated values A0, are specified. Then, for each of the Nb discriminated values B0, the difference from the non-defective product minimum value is calculated as the defect degree value for the discriminated value B0 smaller than the non-defective product minimum value, and is larger than the non-defective product maximum value. For the discriminated value B0, the difference from the maximum value of the non-defective product is calculated as the defect degree value, and the total value of each defect degree value of the Nb discriminated value B0 is calculated and calculated. The process of specifying the provisional range that satisfies the predetermined condition based on the value, and
For each of the M types of provisional ranges, the minimum non-defective value, which is the minimum value of the Na discriminated values A1, and the maximum non-defective product, which is the maximum value of the Na discriminated values A1, are specified. Then, for each of the Nb discriminated values B1, the difference from the non-defective product minimum value is calculated as the defect degree value for the discriminated value B1 smaller than the non-defective product minimum value, and is larger than the non-defective product maximum value. For the discriminated value B1, the difference from the maximum value of the non-defective product is calculated as the defect degree value, and the total value of each defect degree value of the Nb discriminated value B1 is calculated and calculated. The process of specifying the provisional range that satisfies the predetermined condition based on the value, and
For each of the M types of provisional ranges, the minimum value of the non-defective product, which is the minimum value of the Na discriminated values A2, and the maximum value of the non-defective product, which is the maximum value of the Na discriminated values A2, are specified. Then, for each of the Nb discriminated values B2, the difference from the non-defective product minimum value is calculated as the defect degree value for the discriminated value B2 smaller than the non-defective product minimum value, and is larger than the non-defective product maximum value. For the discriminated value B2, the difference from the maximum value of the non-defective product is calculated as the defect degree value, and the total value of each defect degree value of the Nb discriminated value B2 is calculated and calculated. The inspection data generation device according to claim 3, wherein the process of specifying the provisional range that satisfies the condition of satisfying the predetermined condition based on the value is executed. In the third process, the processing unit sets a provisional range that satisfies the predetermined condition based on a value obtained by dividing the total value by the difference between the minimum value of the non-defective product and the maximum value of the non-defective product. The inspection data generation device according to claim 4, which is specified. In the third process, the processing unit increases the degree of difference between the distribution of Na of the discriminated values A0 and the distribution of Nb of the discriminated values B0 as the degree of difference increases. Each of the M provisional ranges was quantified according to a predetermined standard, and the degree of difference between the distribution of Na of the discriminated value A1 and the distribution of Nb of the discriminated value B1 was defined in advance. Each of the M provisional ranges is quantified according to the standard, and the degree of difference between the distribution of Na of the discriminated value A2 and the distribution of Nb of the discriminated value B2 is determined according to the predetermined standard. The inspection data generation device according to any one of claims 3 to 5, wherein each of the M temporary ranges is quantified, and based on the numerical values, the temporary ranges that satisfy the predetermined conditions are specified. In the third process, the processing unit targets one of the M types of temporary ranges, and the amount of difference from the measured value a0 corresponding to the start point in the target temporary range is equal to or less than the La sampling cycle. (La is a natural number) is the tentative range starting from the measured value a0, and the amount of difference from the measured value a0 corresponding to the end point in the tentative range of the target is equal to or less than the Lb sampling cycle (Lb is a natural number). The starting point is the measured value b0 in which the amount of difference between the numerical value in the temporary range having the measured value a0 of the natural number) as the end point and the measured value b0 corresponding to the starting point in the temporary range of the target is less than or equal to the La sampling cycle. The numerical value of the tentative range whose end point is the measured value b0 which is the tentative range and whose end point is the measured value b0 whose end point is equal to or less than the Lb sampling cycle and whose amount of difference from the measured value b0 corresponding to the end point in the tentative range of the target is, respectively. The process of identifying and associating the minimum value of each of the specified numerical values with the provisional range of the target is executed for each of the M provisional ranges as the provisional range of the target, and the provisional range having a large associated numerical value is executed. The inspection data generation device according to claim 6, wherein a provisional range for specifying a provisional range having a relationship that satisfies the predetermined condition is determined as the degree of the relationship that satisfies the predetermined condition is larger. The processing unit periodically changes the measured value a0 into a state in which the measured value a0 exceeds the predetermined threshold value and a state in which the measured value a0 falls below the predetermined threshold value within the predetermined time. Data A, and each of the measured values that periodically changes into a state in which the measured value b0 exceeds the predetermined threshold value and a state in which the measured value b0 falls below the predetermined threshold value within the predetermined time. The data B is acquired, and in the data generation process, the Kth (K is a natural number) of the measured values a0 for the predetermined time exceeds the predetermined threshold value. The measured value a0, the measured value b0 exceeding the predetermined threshold value at the Kth time of the respective measured values b0 for the predetermined time, and the said for the predetermined time. The measured value a0 below the predetermined threshold value at the Kth time of each measured value a0, and the predetermined K time of the measured values b0 for the predetermined time. The inspection data generation device according to any one of claims 3 to 7, which defines the provisional range of the M types with reference to one of the measured values b0 below the threshold value, which is defined in advance. The inspection data generation device according to any one of claims 3 to 8, wherein the processing unit executes a specific result notification process for notifying the target measurement value range specified by the third process. In the specific result notification process, the processing unit causes the display device to display a waveform based on at least one of the measured value data A, the measured value data B, and the measured value data C, and displays the target. The inspection data generation device according to claim 9, wherein the measured value range is displayed in correspondence with the waveform to notify the user. The processing unit extracts each measured value a0 recorded in the measured value data A every other J1 (J1 is a natural number of 2 or more), and the extracted measured value a0 and each extracted measured value. The discriminated value A0, the discriminated value A1 and the discriminated value A2 are calculated using the calculated value a1 corresponding to a0 and the calculated value a2 corresponding to the extracted measured value a0, respectively. At the same time, each measured value b0 recorded in the measured value data B is extracted every other J2 (J2 is a natural number of 2 or more), and the extracted measured value b0 and the extracted measured value b0 are used. A request for calculating the discriminated value B0, the discriminated value B1 and the discriminated value B2 using the corresponding calculated value b1 and the respective calculated value b2 corresponding to the extracted measured value b0, respectively. Item 4. The inspection data generation device according to any one of Items 1 to 10. The inspection data generator according to any one of claims 1 to 11.
The measurement process A is executed to generate the measurement value data A, and the measurement process B is executed to generate the measurement value data B, and the measurement process C is executed to generate the measurement value data C. With a measuring device configured to be generateable,
The inspection data generated by the inspection data generator and the inspection device for inspecting the quality of the inspection target based on the measurement value data C generated by the measurement device are provided to check the quality of the inspection target. An inspection system that is configured to be inspectable. Na times (Na is a natural number of 2 or more) measured over a predetermined time in a predetermined sampling cycle for a non-defective sample of the same type as the inspection target for the predetermined time. Nb times performed under the same measurement conditions as the measurement process A for the Na measurement value data A in which the plurality of measurement values a0 of the above are recorded and the defective sample of the same type as the inspection target and the good sample. By the measurement process B of (Nb is a natural number of 2 or more), Nb of measurement value data B in which a plurality of measurement values b0 for the predetermined time are recorded are acquired, and the inspection of the inspection target is performed. A plurality of measured values c0 for the predetermined time by the Nc times (Nc is a natural number of 2 or more) of the measurement process C, which is sometimes performed on the inspection target under the same measurement conditions as the measurement process A and the measurement process B. Data generation process for generating inspection data that can identify the pass / fail judgment condition of the inspection target based on Nc measurement value data C recorded in each of the measurement value data A and the measurement value data B. Is executed by the processing unit of the inspection data generator, and
In the data generation process
A plurality of calculated values a1 obtained by first-order differentiating each measured value a0 and a plurality of calculated values a2 obtained by second-order differentiating each measured value a0 are derived for each of the measured value data A, and each measured value b0. The first process of deriving a plurality of calculated values b1 obtained by first-order differentiation of the above, and a plurality of calculated values b2 obtained by second-order differentiation of each measured value b0 for each measured value data B,
Corresponds based on the respective measured values a0 within the predetermined measured value range, the respective calculated values a1 within the calculated value range corresponding to the measured value range, and the respective calculated values a2 within the calculated value range. Each of the measured value data A is calculated with a pass / fail determination value A0, and each of the corresponding measurement value data A is subject to pass / fail determination based on the calculated value a1 within the calculated value range. The discrimination value A1 is calculated respectively, and the judgment value A2 for pass / fail judgment is calculated for each measurement value data A corresponding to each calculation value a2 within the calculation value range, and the measurement value is calculated. For each measurement value data B corresponding to each measurement value b0 in the range, each calculation value b1 in the calculation value range, and each calculation value b2 in the calculation value range for pass / fail determination. The discriminated value B0 is calculated respectively, and the discriminated value B1 for pass / fail judgment is calculated for each of the corresponding measured value data B based on the calculated value b1 within the calculated value range, and the calculated value is calculated. A second process of calculating the discriminated value B2 for pass / fail determination for each measured value data B corresponding to each of the calculated value b2 within the range is executed.
The degree of difference between the discriminated value A0 and the discriminated value B0, the degree of difference between the discriminated value A1 and the discriminated value B1, and the difference between the discriminated value A2 and the discriminated value B2. An inspection data generation processing program that specifies one of the pass / fail determination conditions based on the degree and causes the processing unit to execute a process of generating the inspection data.

Images