Inspection device, learned model generation device, inspection method, learned model generation method, inspection program, and learned model program

Abstract

This invention improves the reliability of inspection of electronic components. The inspection device adjusts the value obtained by inputting measurement data (50) into the learned model (35) represented by the sum of the first model f (Rdc4, Rdc2) and the second model g (Ls) based on the adjustment value (Rtng2), and calculates it as the estimated value (Rse) of AC resistance. The first model f (Rdc4, Rdc2) takes the first measured value (Rdc4) of DC resistance measured by the four-terminal method and the second measured value (Rdc2) of DC resistance measured by the two-terminal method as explanatory variables, and takes the first resistance value (Rc) as the value of the resistance component generated by the measurement system using the two-terminal method as the target variable. The second model g (Ls) takes the fourth measured value (Ls) of inductance as explanatory variable, and takes the second resistance value (Rs0) as the value of the resistance component generated by the object being measured as the target variable. The adjustment value (Rtng2) is calculated based on the difference between the value of the AC resistance used as a reference and the estimated value g (Ls_av2) of the second resistance value calculated by inputting at least one measured data (50) into the second model g (Ls).

Description

Technical Field

[0001] This invention relates to inspection apparatus, learning-completed model generation apparatus, inspection method, learning-completed model generation method, inspection program, and learning-completed model program, for example, to an inspection apparatus for inspecting inductor components. Background Technology

[0002] Previously, inspection devices were known to measure the electrical characteristics of electronic components such as chips and inductors, and to determine whether the electronic components were in good condition based on the measurement results. For example, Patent Document 1 discloses an inspection device that measures the AC resistance and inductance of an inductor element to be inspected, calculates a Q value using these measured values, and determines whether the inductor element was in good condition based on the calculated Q value.

[0003] In Patent Document 1, a method for measuring the AC resistance of an inductor element is described, in which the AC resistance is calculated by subtracting the estimated value of the contact resistance of the measuring probe used in the two-terminal method from the measured value of the AC resistance measured by the two-terminal method. Furthermore, Patent Document 1 describes calculating the estimated value of the contact resistance by subtracting the measured value of the DC resistance measured by the two-terminal method from the measured value of the DC resistance measured by the four-terminal method, and correcting the measured value of the AC resistance using a value obtained by multiplying the estimated value of the contact resistance by a coefficient of 0 or more and less than 1.

[0004] The inspection apparatus disclosed in Patent Document 1 calibrates the measured value of AC resistance based on the premise that the relationship between the series resistance and AC resistance in the inductor element is linear. However, the relationship between the series resistance and AC resistance of actual inductor elements is unclear; for example, if the relationship is non-linear, the measured value of AC resistance may not be properly calibrated. Furthermore, in the inspection apparatus disclosed in Patent Document 1, to avoid overcalibration of AC resistance, the measured value of AC resistance is calibrated using a value obtained by multiplying the estimated value of the contact resistance by a coefficient. However, if this coefficient is inappropriate, the AC resistance may not be properly calibrated.

[0005] Existing technical documents Patent documents Patent Document 1: Japanese Patent No. 6949675 Summary of the Invention The problem the invention aims to solve Prior to this application, the inventors conducted the following research: By performing machine learning on the measured values ​​of the series resistance and AC resistance of the inductor element, a model was generated to calculate the estimated value of the AC resistance based on the measured value of the DC resistance of the inductor element of the inspection object. This model was then used to determine the conformity of the inductor element. The research revealed the following problems.

[0006] Generally, when inspecting a DUT using an inspection device (automatic measuring instrument), a calibration process is performed before the inspection begins. In the calibration process, firstly, a reference device (gold standard device) is selected from multiple DUTs. Next, in a low-noise environment, the electrical characteristics of three reference devices (e.g., open-circuit reference, short-circuit reference, and load reference) are manually measured using a retainer, and these measured values ​​are used as the defined values ​​of the reference devices. It should be noted that when using pre-defined open-circuit and short-circuit references, only the reference device of the DUT is measured. Next, the three reference devices with defined values ​​are measured using the inspection device, and the measured values ​​under each reference are obtained. Then, a correction formula is generated to correct the measured values ​​so that the error between the measured values ​​under each reference and the pre-obtained defined values ​​is reduced, and this formula is stored in the inspection device. Thus, a series of calibration processes are completed. After the inspection device begins operation, it corrects the values ​​obtained from measuring the DUT based on the above correction formula and outputs the corrected values ​​as the measured values ​​of the DUT. This reduces measurement errors during inspection.

[0007] As mentioned above, the defined values ​​of the reference devices are measured in a low-noise environment, therefore the measurement error included in the defined values ​​of the reference devices is small. However, if the defined values ​​of the reference devices include errors, it will affect all measurement results of the DUT performed subsequently by the inspection device.

[0008] For example, if there is a difference between the measurement data used in machine learning to generate the learned model and the defined values ​​of the reference device used in the calibration process, the accuracy of the inspection using the learned model may be reduced.

[0009] The present invention was made in view of the above-mentioned problems, and its purpose is to improve the reliability of inspection of electronic components.

[0010] Solution for solving the problem A representative embodiment of the present invention provides an inspection apparatus characterized by comprising: a data acquisition unit for acquiring measurement data, the measurement data including a first measured value of the DC resistance of the test object measured by a four-terminal method, a second measured value of the DC resistance of the test object measured by a two-terminal method, a third measured value of the AC resistance of the test object measured by a two-terminal method, and a fourth measured value of the inductance of the test object measured by a two-terminal method; a storage unit for storing a learned model, the learned model being configured to enable a computer to calculate the third measured value based on the input first measured value, second measured value, and fourth measured value; and an estimation unit for calculating the third measured value corresponding to the first measured value, second measured value, and fourth measured value acquired by the data acquisition unit based on the learned model stored in the storage unit. The estimated value; and the adjustment value calculation unit, which calculates an adjustment value for adjusting the value calculated by the learned model, the learned model being represented by the sum of a first model and a second model, the first model using the first measured value and the second measured value as explanatory variables, and using a first resistance value as the value of the resistance component generated by the measurement system using the two-terminal method as the target variable, the second model using the fourth measured value as an explanatory variable, and using a second resistance value as the value of the resistance component generated by the measured object as the target variable, the adjustment value calculation unit calculating the adjustment value based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value calculated by inputting the fourth measured value included in at least one of the measurement data into the second model, the estimation unit adjusting the value calculated by the learned model based on the adjustment value and calculating it as the estimated value of the third measured value.

[0011] Invention Effects The inspection device according to the present invention can improve the reliability of inspection of electronic components. Attached Figure Description

[0012] Figure 1 This is a diagram showing the configuration of the inspection system in Implementation Method 1.

[0013] Figure 2 This is a diagram illustrating an example of the configuration of the learning-completed model generation device in the inspection system of Embodiment 1.

[0014] Figure 3 This is a graph used to illustrate the deviation between the estimated value of the second resistance value Rs0 of the AC resistance calculated by model g(Ls) and the defined value (AC resistance Rs) of the DUT in the calibration process.

[0015] Figure 4This is a flowchart illustrating the process of generating a learned model by the learned model generation device in Implementation 1.

[0016] Figure 5 This is a flowchart illustrating the process of machine learning (step S3) in implementation method 1.

[0017] Figure 6 This is a diagram illustrating an example of the configuration of the data processing control device in the inspection apparatus of Embodiment 1.

[0018] Figure 7 This is a flowchart illustrating the inspection process performed by the inspection device in Implementation 1.

[0019] Figure 8 This is a flowchart illustrating the process of calculating the adjustment value (step S12) in Implementation Method 1.

[0020] Figure 9 This is a flowchart illustrating the process of inspection (step S13) performed by the inspection device in Implementation 1.

[0021] Figure 10 This is a diagram illustrating an example of the configuration of the learning-completed model generation device in the inspection system of Embodiment 2.

[0022] Figure 11 This is a flowchart illustrating the process of machine learning (step S3) in implementation method 2.

[0023] Figure 12 This is a diagram illustrating an example of the configuration of the data processing control device in the inspection apparatus of Embodiment 2.

[0024] Figure 13 This is a flowchart illustrating the process of adjusting value calculation (step S12) in the inspection process of Implementation Method 2. Detailed Implementation

[0025] 1. Overview of the implementation method First, a summary of representative embodiments of the invention disclosed in this application will be given. It should be noted that, in the following description, as an example, the reference numerals in the accompanying drawings corresponding to the constituent elements of the invention are indicated by parentheses.

[0026] [1] A representative embodiment of the present invention provides an inspection device (2), characterized in that it comprises: a data acquisition unit (21) for acquiring measurement data (50), the measurement data (50) including a first measured value (Rdc4) of the DC resistance of the object to be measured by a four-terminal method, a second measured value (Rdc2) of the DC resistance of the object to be measured by a two-terminal method, a third measured value (Rs) of the AC resistance of the object to be measured by a two-terminal method, and a fourth measured value (Ls) of the inductance of the object to be measured by a two-terminal method; a storage unit (22) for storing a learning completion model (35), the learning completion model (35) enabling a computer to calculate the third measured value based on the input first measured value, the second measured value, and the fourth measured value; and an estimation unit (23) for calculating an estimated value of the third measured value corresponding to the first measured value, the second measured value, and the fourth measured value acquired by the data acquisition unit based on the learning completion model stored in the storage unit; The adjustment value calculation unit (26) calculates an adjustment value (Rtng2) for adjusting the value calculated by the learning completed model. The learning completed model is represented by the sum of a first model (f(Rdc4,Rdc2)) and a second model (g(Ls)). The first model (f(Rdc4,Rdc2)) uses the first measured value and the second measured value as explanatory variables and uses the first resistance value (Rc), which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model (g(Ls)) uses the fourth measured value as an explanatory variable and uses the second resistance value (Rs0), which is the value of the resistance component generated by the measured object, as the target variable. The adjustment value calculation unit calculates the adjustment value based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value calculated by inputting the fourth measured value included in at least one of the measured data into the second model. The estimation unit adjusts the value calculated by the learning completed model based on the adjustment value and calculates it as the estimated value of the third measured value.

[0027] 〔2〕In the inspection device described in 〔1〕 above, it may also be that the value of the AC resistance as a reference is the average value (Rs_av2) of the third measurement values included in the plurality of measurement data of the measurement object for which the difference between the first measurement value and the second measurement value is less than a threshold value (|Rdc2 - Rdc4| < Rdth). The adjustment value calculation unit calculates the estimated value (g(Ls_av2)) of the second resistance value by inputting the average value (Ls_av2) of the fourth measurement values included in the plurality of measurement data of the measurement object for which the difference between the first measurement value and the second measurement value is less than the threshold value into the second model, and calculates the difference between the average value of the third measurement values and the estimated value of the second resistance value as the adjustment value (Rtng2 = Rs_av2 - g(Ls_av2)).

[0028] 〔3〕In the inspection device described in 〔1〕 above, it may also be that the value of the AC resistance as a reference is the defined value (Rs_d2) of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value (g(Ls_av2)) of the second resistance value by inputting the average value (Ls_av2) of the fourth measurement values included in the measurement data of the plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold value (|Rdc2 - Rdc4| < Rdth) into the second model, and calculates the difference between the defined value and the estimated value of the second resistance value as the adjustment value (Rtng2 = Rs_d2 - g(Ls_av2)).

[0029] [4] Another representative embodiment of the present invention provides an inspection device (2A) characterized by comprising: a data acquisition unit (21) for acquiring measurement data (50), the measurement data (50) including a first measured value (Rdc4) of the DC resistance of the object to be measured by a four-terminal method, a second measured value (Rdc2) of the DC resistance of the object to be measured by a two-terminal method, and a third measured value (Rs) of the AC resistance of the object to be measured by a two-terminal method; a storage unit (22) for storing a learning completion model (35A) for enabling a computer to function in such a way as to calculate the third measured value based on the input first measured value and the second measured value; an estimation unit (23A) for calculating an estimated value of the third measured value corresponding to the first measured value and the second measured value acquired by the data acquisition unit based on the learning completion model stored in the storage unit; and an adjustment value calculation unit (26A) for calculating an adjustment value using... The adjustment value is calculated by adjusting the value calculated through the learned model, which is represented by the sum of a first model (f(Rdc4,Rdc2)) and a second model (g(Rdc4)). The first model (f(Rdc4,Rdc2)) uses the first and second measured values ​​as explanatory variables and a first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model (g(Rdc4)) uses the first measured value as an explanatory variable and a second resistance value (Rs0), which is the value of the resistance component generated by the measured object, as the target variable. The adjustment value calculation unit calculates the adjustment value based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value calculated by inputting the first measured value included in at least one of the measured data into the second model. The estimation unit adjusts the value calculated through the learned model based on the adjustment value and calculates it as the estimated value of the third measured value.

[0030] 〔5〕In the inspection device described in the above-mentioned 〔4〕, it may also be that the value of the AC resistance as the reference is the average value (Rs_av2) of the third measurement value included in the plurality of measurement data of the measurement object whose difference between the first measurement value and the second measurement value is less than the threshold value (|Rdc2 - Rdc4| < Rdth). The adjustment value calculation unit inputs the average value (Rdc4_av2) of the first measurement value included in the plurality of measurement data of the measurement object whose difference between the first measurement value and the second measurement value is less than the threshold value into the second model to calculate the estimated value (g(Rdc4_av2)) of the second resistance value, and calculates the difference between the average value of the third measurement value and the estimated value of the second resistance value (Rs_av2 - g(Rdc4_av2)) as the adjustment value.

[0031] 〔6〕In the inspection device described in the above-mentioned 〔4〕, it may also be that the value of the AC resistance as the reference is the defined value (Rs_d2) of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit inputs the average value (Rdc4_av2) of the first measurement value included in the plurality of measurement data of the measurement object whose difference between the first measurement value and the second measurement value is less than the threshold value (|Rdc2 - Rdc4| < Rdth) into the second model to calculate the estimated value (g(Rdc4_av2)) of the second resistance value, and calculates the difference between the defined value and the estimated value of the second resistance value (Rs_d2 - g(Rdc4_av2)) as the adjustment value.

[0032] 〔7〕In the inspection device according to any one of the above-mentioned 〔1〕 to 〔6〕, it may also be that the inspection device further includes a correction unit (24), and the correction unit (24) performs the following correction process: based on the estimated value of the third measurement value, and based on the first measurement value and the second measurement value acquired by the data acquisition unit, corrects the third measurement value, and outputs the corrected third measurement value as the value of the AC resistance of the measurement object.

[0033] 〔8〕In the inspection device described in the above-mentioned 〔7〕, it may also be that the correction unit calculates the resistance component (Rc) generated by the measurement system using the two-terminal method according to the first model based on the first measurement value and the second measurement value acquired by the data acquisition unit, and corrects the third measurement value based on the resistance component generated by the measurement system using the two-terminal method as the correction process.

[0034] [9] In the inspection device described in [8] above, the calibration unit may calculate the error between the estimated value of the third measured value calculated by the estimation unit and the third measured value obtained by the data acquisition unit, and perform the calibration process if the error is less than a predetermined threshold.

[0035]

[10] A method for inspection according to a representative embodiment of the present invention is characterized by comprising: a first step (S121, S131-S133), acquiring measurement data (50), the measurement data (50) including a first measured value (Rdc4) of the DC resistance of the object to be measured by a four-terminal method, a second measured value (Rdc2) of the DC resistance of the object to be measured by a two-terminal method, a third measured value (Rs) of the AC resistance of the object to be measured by a two-terminal method, and a fourth measured value (Ls) of the inductance of the object to be measured by a two-terminal method; a second step (S134), calculating an estimated value of the third measured value corresponding to the first measured value, the second measured value, and the fourth measured value obtained in the first step based on a learned model (35) for enabling a computer to function in such a way as to estimate the third measured value based on the input first measured value, the second measured value, and the fourth measured value; a third step (S122-S126), calculating an adjustment value for adjusting the value calculated by the learned model; a fourth step ( (S134) Adjusting the estimated value of the third measured value calculated in the second step based on the adjustment value; and the fifth step (S135-S138) correcting the third measured value based on the estimated value of the third measured value adjusted in the fourth step, based on the first measured value and the second measured value obtained in the first step, and outputting the corrected third measured value as the AC resistance value of the measured object, wherein the learned model is represented by the sum of the first model and the second model, wherein the first model uses the first measured value and the second measured value as explanatory variables, and uses the first resistance value as the value of the resistance component generated by the measurement system using the two-terminal method as the target variable, wherein the second model uses the fourth measured value as an explanatory variable, and uses the second resistance value as the value of the resistance component generated by the measured object as the target variable, and the third step includes the step of calculating the adjustment value based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value calculated by inputting the fourth measured value included in at least one of the measured data into the second model (S122-S126).

[0036]

[11] In the inspection method described in

[10] above, the value of the AC resistance used as a reference may be the average value (Rs_av2) of the third measurement value included in the multiple measurement data of the test object where the difference between the first measurement value and the second measurement value is less than a threshold. The third step includes: a step (S125) to calculate the estimated value (g(Ls_av2)) of the second resistance value by inputting the average value (Ls_av2) of the fourth measurement value included in the multiple measurement data of the test object where the difference between the first measurement value and the second measurement value is less than the threshold into the second model; and a step (S126) to calculate the difference between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value.

[0037]

[12] In the inspection method described in

[10] above, the value of the AC resistance used as a reference may be the defined value (Rs_d2) of the AC resistance of the reference device in the calibration process. The third step includes: a step (S125) to calculate the estimated value (g(Ls_av2)) of the second resistance value by inputting the average value (Ls_av2) of the fourth measurement value included in the measurement data of a plurality of measurement objects whose difference between the first measurement value and the second measurement value is less than a threshold into the second model; and a step (S126) to calculate the difference (Rs_d2-g(Ls_av2)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0038]

[13] A representative embodiment of the present invention provides an inspection method, characterized in that it includes: a first step (S121, S131-S133), acquiring measurement data (50), wherein the measurement data (50) includes a first measured value (Rdc4) of the DC resistance of the object being measured by a four-terminal method, a second measured value (Rdc2) of the DC resistance of the object being measured by a two-terminal method, and a third measured value (Rs) of the AC resistance of the object being measured by a two-terminal method; a second step (S134), based on the method for enabling a computer to... The learning-completed model (35A) functions by estimating the third measurement value using the first and second measurement values ​​input in the first step, and calculates the estimated value of the third measurement value corresponding to the first and second measurement values ​​obtained in the first step; the third step (S122, S123, S124A~S126A) calculates an adjustment value (Rtng2) for adjusting the value calculated by the learning-completed model; the fourth step (S134) adjusts the estimated value of the third measurement value calculated in the second step based on the adjustment value; And in the fifth step (S135-S138), based on the estimated value of the third measured value adjusted in the fourth step, the third measured value is corrected based on the first measured value and the second measured value obtained in the first step, and the corrected third measured value is output as the AC resistance value of the measured object. The learned model is represented by the sum of the first model (f(Rdc4,Rdc2)) and the second model (g(Rdc4)). The first model (f(Rdc4,Rdc2)) uses the first measured value and the second measured value as explanatory variables. The first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, is used as the target variable. The second model (g(Rdc4)) uses the first measured value as the explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, is used as the target variable. The third step includes the step of calculating the adjustment value based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value calculated by inputting the first measured value included in at least one of the measurement data into the second model (S123, S124A to S126A).

[0039]

[14] In the inspection method described in

[13] above, the value of the AC resistance used as a reference may also be the average value (Rs_av2) of the third measurement value included in the multiple measurement data of the test object for which the difference between the first measurement value and the second measurement value is less than a threshold. The third step includes: a step (S125A) to calculate the estimated value (g(Rdc4_av2)) of the second resistance value by inputting the average value (Rdc4_av2) of the first measurement value included in the multiple measurement data of the test object for which the difference between the first measurement value and the second measurement value is less than the threshold into the second model; and a step (S126A) to calculate the difference (Rs_av2-g(Rdc4_av2)) between the average value of the third measurement value and the estimated value of the second resistance value.

[0040]

[15] In the inspection method described in

[13] above, the value of the AC resistance used as a reference may be the defined value (Rs_d2) of the AC resistance of the reference device in the calibration process. The third step includes: a step (S125A) to calculate the estimated value (g(Rdc4_av2)) of the second resistance value by inputting the average value (Rdc4_av2) of the first measured value included in the measurement data of a plurality of measured objects whose difference between the first measured value and the second measured value is less than a threshold into the second model; and a step (S126A) to calculate the difference (Rs_d2-g(Rdc4_av2)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0041]

[16] The inspection program of a representative embodiment of the present invention is characterized in that a computer is made to execute each step of the inspection method according to any one of

[10] to

[15] above.

[0042] 〔17〕A learning-completed model generation device (3) according to a representative embodiment of the present invention is characterized by comprising: a learning measurement data acquisition unit (31) that acquires learning measurement data (34, 34_1 to 34_n), where the learning measurement data (34, 34_1 to 34_n) establishes a correspondence between a third measurement value of the AC resistance of a measurement object measured by a two-terminal method, a first measurement value (Rdc4) of the DC resistance of the measurement object measured by a four-terminal method, a second measurement value (Rdc2) of the DC resistance of the measurement object measured by a two-terminal method, and a fourth measurement value (Ls) of the inductance of the measurement object measured by a two-terminal method; and a learning-completed model generation unit (32) that generates a learning-completed model (35) for causing a computer to function in a manner of calculating the third measurement value based on input data including the first measurement value, the second measurement value, and the fourth measurement value by performing machine learning on the learning measurement data. The learning-completed model is represented by the sum of a first model (f(Rdc4, Rdc2)) and a second model (g(Ls)). The first model (f(Rdc4, Rdc2)) uses the first measurement value and the second measurement value as explanatory variables and uses a first resistance value (Rc), which is a value of the resistance component generated by a measurement system using a two-terminal method, as the target variable. The second model (g(Ls)) uses the fourth measurement value as the explanatory variable and uses a second resistance value (Rs0), which is a value of the resistance component generated by the measurement object, as the target variable. The learning-completed model generation unit has: a first model generation unit (321) that generates the first model by performing machine learning on the first measurement value and the second measurement value included in the learning measurement data of the measurement object; a second model generation unit (322) that generates the second model by performing machine learning on the fourth measurement value included in the learning measurement data of the measurement object for which the difference between the first measurement value and the second measurement value is less than the threshold (|Rdc2 - Rdc4| < Rdth); and an adjustment value calculation unit (323) that calculates an adjustment value (Rtng1) corresponding to the difference between a value of the AC resistance as a reference and an estimated value of the second resistance value calculated by inputting the fourth measurement value included in at least one of the learning measurement data into the second model. The first model generation unit adjusts the parameters of the first model in a state where the parameters of the second model are fixed in a model represented by the sum of the first model, the second model, and the adjustment value.

[0043]

[18] In the learning-completed model generation apparatus described in

[17] above, the value of the AC resistance used as a reference may be the average value (Rs_av1) of the third measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The adjustment value calculation unit calculates the estimated value (g(Ls_av1)) of the second resistance value by inputting the average value (Ls_av1) of the fourth measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold into the second model, and calculates the difference (Rs_av1-g(Ls_av1)) between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value.

[0044]

[19] In the learning completed model generation apparatus described in

[17] above, the value of the AC resistance used as a reference may be the defined value (Rs_d1) of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value (g(Ls_av1)) of the second resistance value by inputting the average value (Ls_av1) of the fourth measurement value included in the learning measurement data of a plurality of measurement objects whose difference between the first measurement value and the second measurement value is less than the threshold value into the second model, and calculates the difference (Rs_d1-g(Ls_av1)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0045] 〔20〕A learning-completed model generation device (3A) according to a representative embodiment of the present invention is characterized by comprising: a learning measurement data acquisition unit (31) that acquires learning measurement data (34A, 34A_1 to 34A_n), wherein the learning measurement data (34A, 34A_1 to 34A_n) establishes a correspondence between a third measurement value (Rs) of an AC resistance of a measurement object measured by a two-terminal method, a first measurement value (Rdc4) of a DC resistance of the measurement object measured by a four-terminal method, and a second measurement value (Rdc2) of the DC resistance of the measurement object measured by a two-terminal method; and a learning-completed model generation unit (32A) that generates a learning-completed model (35A) for causing a computer to function in such a manner as to calculate the third measurement value based on input data including the first measurement value and the second measurement value through machine learning of the learning measurement data. The learning-completed model is represented by the sum of a first model (f(Rdc4, Rdc2)) and a second model (g(Rdc4)). The first model (f(Rdc4, Rdc2)) uses the first measurement value and the second measurement value as explanatory variables and uses a first resistance value (Rc), which is a value of a resistance component generated by a measurement system using a two-terminal method, as a target variable. The second model (g(Rdc4)) uses the first measurement value as an explanatory variable and uses a second resistance value (Rs0), which is a value of a resistance component generated by the measurement object, as a target variable. The learning-completed model generation unit includes: a first model generation unit (321) that generates the first model through machine learning of the first measurement value and the second measurement value included in the learning measurement data of the measurement object; a second model generation unit (322A) that generates the second model through machine learning of the first measurement value included in the learning measurement data of the measurement object for which the difference between the first measurement value and the second measurement value is less than the threshold (|Rdc2 - Rdc4| < Rdth); and an adjustment value calculation unit (323A) that calculates an adjustment value (Rtng1) corresponding to the difference between a reference AC resistance value and an estimated value of the second resistance value calculated by inputting the first measurement value included in at least one of the learning measurement data into the second model. The first model generation unit adjusts the parameters of the first model in a state where the parameters of the second model are fixed in a model (f(Rdc4, Rdc2) + g(Rdc4) + Rtng1) represented by the sum of the first model, the second model, and the adjustment value.

[0046]

[21] In the learning-completed model generation apparatus described in

[20] above, the value of the AC resistance used as a reference may be the average value (Rs_av1) of the third measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The adjustment value calculation unit calculates the estimated value (g(Rdc4_av1)) of the second resistance value by inputting the average value (Rdc4_av1) of the first measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold into the second model, and calculates the difference (Rs_av1-g(Rdc4_av1)) between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value.

[0047]

[22] In the learning completed model generation apparatus described in

[20] above, the value of the AC resistance used as a reference may be the defined value (Rs_d1) of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value (g(Rdc4_av1)) of the second resistance value by inputting the average value (Rdc4_av1) of the first measured value included in the learning measurement data of a plurality of measured objects whose difference between the first measured value and the second measured value is less than the threshold into the second model, and calculates the difference (Rs_d1-g(Rdc4_av1)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0048]

[23] A learning-completed model generation method according to a representative embodiment of the present invention is characterized by comprising: a first step (S1), acquiring learning measurement data (34, 34_1 to 34_n), wherein the learning measurement data (34, 34_1 to 34_n) establishes a correspondence between the third measured value of the AC resistance of the object to be measured by the two-terminal method and the first measured value (Rdc4) of the DC resistance of the object to be measured by the four-terminal method, the second measured value (Rdc2) of the DC resistance of the object to be measured by the two-terminal method, and the fourth measured value (Ls) of the inductance of the object to be measured by the two-terminal method; And in the second step (S3, S4), machine learning is performed on the learning measurement data obtained in the first step to generate a learned model (35) that enables the computer to calculate the third measurement value based on input data including the first measurement value, the second measurement value, and the fourth measurement value. The learned model is represented by the sum of a first model (f(Rdc4,Rdc2)) and a second model (g(Ls)). The first model (f(Rdc4,Rdc2)) uses the first measurement value and the second measurement value as explanatory variables, and uses the third measurement value as the value of the resistance component generated by the measurement system using the two-terminal method. The second model (g(Ls)) uses a resistance value (Rc) as the target variable, takes the fourth measured value as the explanatory variable, and takes a second resistance value (Rs0) as the value of the resistance component generated by the measured object as the target variable. The second step includes: a third step (S32), generating the second model by performing machine learning on the fourth measured value included in the learning measurement data of the measured object where the difference between the first measured value and the second measured value is less than a threshold; and a fourth step (S33-S35), calculating the value of the AC resistance as a reference and inputting at least one of the fourth measured values ​​included in the learning measurement data into the target model. The adjustment value (Rtng1) corresponding to the difference between the estimated value of the second resistance value calculated by the second model; and the fifth step (S37, S38), generating the first model by performing machine learning on the first and second measured values ​​included in the learning measurement data of the measurement object where the difference between the first and second measured values ​​is greater than a threshold, the fifth step includes adjusting the parameters of the first model while fixing the parameters of the second model in the model (f(Rdc4,Rdc2)+g(Ls)+Rtng1) represented by the sum of the first model, the second model, and the adjustment value (S38).

[0049]

[24] In the learning-completed model generation method described in

[23] above, the value of the AC resistance used as a reference may be the average value (Rs_av1) of the third measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The fourth step includes: a step (S35) to calculate the estimated value (g(Ls_av1)) of the second resistance value by inputting the average value (Ls_av1) of the fourth measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold into the second model; and a step (S35) to calculate the difference between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value (Rs_av1-g(Ls_av1)).

[0050]

[25] In the learning-completed model generation method described in

[23] above, the value of the AC resistance used as a reference may be the defined value (Rs_d1) of the AC resistance of the reference device in the calibration process. The fourth step includes: a step (S35) to calculate the estimated value (g(Ls_av1)) of the second resistance value by inputting the average value (Ls_av1) of the fourth measurement value included in the learning measurement data of a plurality of measurement objects whose difference between the first measurement value and the second measurement value is less than the threshold into the second model; and a step (S35) to calculate the difference (Rs_d1-g(Ls_av1)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0051]

[26] Another representative embodiment of the present invention provides a method for generating a completed learning model, characterized in that it includes: a first step (S1), acquiring learning measurement data (34A, 34A_1 to 34A_n), wherein the learning measurement data (34A, 34A_1 to 34A_n) establishes a correspondence between the third measured value (Rs) of the AC resistance of the object being measured by the two-terminal method and the first measured value (Rdc4) of the DC resistance of the object being measured by the four-terminal method and the second measured value (Rdc2) of the DC resistance of the object being measured by the two-terminal method; and a second step (S3, S4). By performing machine learning on the learning measurement data obtained in the first step, a fully learned model is generated that enables the computer to function in a manner that calculates the third measurement value based on input data including the first and second measurement values. The fully learned model is represented by the sum of a first model (f(Rdc4,Rdc2)) and a second model (g(Rdc4)). The first model (f(Rdc4,Rdc2)) uses the first and second measurement values ​​as explanatory variables and the first resistance value (Rc), which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model... The model (g(Rdc4)) uses the first measured value as an explanatory variable and the second resistance value (Rs0), which is the value of the resistance component generated by the measured object, as the target variable. The second step includes: a fourth step (S32A), generating the second model by performing machine learning on the first measured value included in the learning measurement data of the measured object where the difference between the first measured value and the second measured value is less than a threshold; and a fifth step (S33, S34A, S35A), calculating the value of the AC resistance as a reference and inputting at least one of the first measured values ​​included in the learning measurement data into the second model. The adjustment value corresponding to the difference between the estimated value of the second resistance value calculated by the model; and the sixth step (S37A, S38A), generating the first model by performing machine learning on the first and second measured values ​​included in the learning measurement data of the measurement object where the difference between the first and second measured values ​​is greater than a threshold, the sixth step includes adjusting the parameters of the first model while fixing the parameters of the second model in the model (f(Rdc4,Rdc2)+g(Rdc4)+Rtng1) represented by the sum of the first model, the second model, and the adjustment value (S38A).

[0052]

[27] In the learning-completed model generation method described in

[26] above, the value of the AC resistance used as a reference may be the average value (Rs_av1) of the third measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The fifth step includes: a step (S35A) to calculate the estimated value (g(Rdc4_av1)) of the second resistance value by inputting the average value (Rdc4_av1) of the first measurement value included in the learning measurement data of a plurality of measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold into the second model; and a step (S35A) to calculate the difference (Rs_av1-g(Rdc4_av1)) between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value.

[0053]

[28] In the learning-completed model generation method described in

[26] above, the value of the AC resistance used as a reference may be the defined value (Rs_d1) of the AC resistance of the reference device in the calibration process. The fifth step includes: a step (S35A) to calculate the estimated value (g(Rdc4_av1)) of the second resistance value by inputting the average value (Rdc4_av1) of the first measured value included in the learning measurement data of a plurality of measured objects whose difference between the first measured value and the second measured value is less than the threshold into the second model; and a step (S35A) to calculate the difference (Rs_d1-g(Rdc4_av1)) between the defined value and the estimated value of the second resistance value as the adjustment value.

[0054]

[29] A program for generating a learned model according to a representative embodiment of the present invention is characterized in that a computer is made to execute each step of the learned model generation method according to any one of

[23] to

[28] above.

[0055] 2. Specific examples of implementation methods Hereinafter, specific examples of embodiments of the present invention will be described with reference to the accompanying drawings. It should be noted that in the following description, common components in various embodiments will be labeled with the same reference numerals, and repeated descriptions will be omitted.

[0056] <Implementation Method 1> Figure 1 This is a diagram showing the configuration of the inspection system 1 in embodiment 1.

[0057] Figure 1 The inspection system 1 shown is a system for inspecting whether the object to be inspected (hereinafter also referred to as "DUT") is in good condition. Figure 1As shown, the inspection system 1 includes: a learning-complete model generation device 3, which learns from multiple measurement results of the DUT using machine learning to generate a learning-complete model; and an inspection device 2, which uses the generated learning-complete model to inspect the DUT.

[0058] Inspection device 2 is a device that measures the electrical characteristics of the DUT and checks whether the DUT is good based on the measurement results. For example, inspection device 2 is a device that checks whether small electronic components (chip components) are good and packages the chip components that are determined to be qualified into a factory-ready state (so-called chip packaging machine).

[0059] In the following description, as an example, the DUT is assumed to be an inductor element (e.g., a chip inductor element), but is not limited thereto.

[0060] The testing device 2 uses the learned model described later to measure the electrical characteristics of the inductor element used as the DUT. Specifically, as follows... Figure 1 As shown, the inspection device 2 includes: a data processing and control device 10, a first measuring unit 11, a second measuring unit 12, an operation unit 13, an output unit 14, and a conveying mechanism 15.

[0061] The first measuring unit 11 is a device that measures the electrical characteristics of an inductor element that serves as a DUT using a four-terminal method. Examples of the first measuring unit 11 include impedance measuring instruments such as ohmmeters and LCR meters that can perform impedance measurement using the four-terminal method.

[0062] The second measuring unit 12 is a device that measures the electrical characteristics of the inductor element of the DUT using the two-terminal method. Examples of the second measuring unit 12 include impedance measuring instruments such as LCR meters capable of measuring impedance using the two-terminal method.

[0063] It should be noted that the first measuring unit 11 and the second measuring unit 12 are not limited to the above examples, as long as they are devices capable of measuring the impedance and other electrical characteristics of the DUT.

[0064] The first measuring unit 11 measures the DC resistance of the inductor element, which is the DUT, using a four-terminal method according to the instruction from the data processing control device 10. For example, the first measuring unit 11 includes a moving mechanism (not shown) that moves probes 61a to 61d, a current output unit, a voltage detection unit (not shown), and a measurement value calculation unit (not shown) that calculates the measured value based on the detection result.

[0065] For example, when the first measuring unit 11 receives an instruction to perform a measurement from the data processing control device 10, its moving mechanism contacts probes 61a and 61c with one terminal of an inductor element that has been transported to a predetermined measuring position, and contacts probes 61b and 61d with the other terminal of the inductor element. Then, the current output unit of the first measuring unit 11 supplies direct current to the inductor element via probes 61a and 61b. The voltage detection unit of the first measuring unit 11 detects the voltage value between the inductor terminals when direct current is supplied to the inductor element via probes 61c and 61d. Based on the detected voltage value and the current value of the direct current supplied to the inductor element, the measurement value calculation unit of the first measuring unit 11 calculates the measured value Rdc4 of the direct current resistance of the inductor element.

[0066] The second measuring unit 12 measures the DC resistance, AC resistance, and inductance of the inductor element of the DUT using a two-terminal method, according to the instruction from the data processing control device 10. For example, the second measuring unit 12 has a moving mechanism (not shown) for moving probes 62a and 62b, a current output unit, a voltage detection unit (not shown), and a measurement value calculation unit (not shown) for calculating the measured value based on the detection result.

[0067] For example, when the second measuring unit 12 receives an instruction to perform a measurement from the data processing control device 10, its moving mechanism contacts probe 62a with one terminal of the inductor element that has been transported to a predetermined measuring position, and contacts probe 62b with the other terminal of the inductor element. Then, the current output unit of the second measuring unit 12 supplies DC current to the inductor element via probes 62a and 62b, and the voltage detection unit of the second measuring unit 12 detects the voltage value between the two terminals of the inductor element via probes 62a and 62b. Based on the detected voltage value and the current value of the DC current supplied to the inductor element, the measurement value calculation unit of the second measuring unit 12 calculates the measured value Rdc2 of the DC resistance of the inductor element.

[0068] Furthermore, when the probes 62a and 62b are in contact with the two terminals of the inductor element via the moving mechanism of the second measuring unit 12, the current output unit of the second measuring unit 12 supplies alternating current to the inductor element via the probes, and the voltage detection unit of the second measuring unit 12 detects the alternating voltage value between the two terminals of the inductor element via the probes 62a and 62b. Based on the detected alternating voltage value (effective voltage value), the alternating current value (effective current value) supplied to the inductor element, and the phase difference between the alternating voltage and the alternating current, the measurement value calculation unit of the second measuring unit 12 calculates the measured value Rs of the alternating resistance and the measured value Ls of the inductor element.

[0069] It should be noted that part of the functions of the first measurement unit 11 and the second measurement unit 12 can also be implemented by the data processing control device 10. For example, the above operations performed by the measurement value calculation units of the first measurement unit 11 and the second measurement unit 12 can also be executed by the data processing control device 10.

[0070] The operation unit 13 is an input interface for the user to operate the inspection device 2. As the operation unit 13, various buttons, touch panels, etc. can be exemplified. For example, by the user operating the operation unit 13, various inspection conditions for inspecting the inductor element as the DUT can be set in the inspection device 2, and the execution and stop of inspections, etc. can be instructed to the inspection device 2.

[0071] The output unit 14 is a functional unit for outputting various information such as inspection conditions and inspection results in the inspection device 2. The output unit 14 is, for example, a display device equipped with an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence). For example, when the execution of the inspection of the DUT is instructed by the user operating the operation unit 13, the output unit 14 displays information such as the inspection result on the screen according to the control performed by the data processing control device 10.

[0072] It should be noted that the output unit 14 can also be a display device equipped with a touch panel that realizes part of the functions of the operation unit 13. In addition, the output unit 14 can also include a communication circuit, etc. that outputs data such as inspection results to the outside by wire or wirelessly.

[0073] The conveying mechanism 15 is a device that conveys the inductor element to be inspected to an appropriate position inside the inspection device 2 according to the control of the data processing control device 10. For example, when measuring by the first measurement unit 11, the conveying mechanism 15 conveys the inductor element to be inspected to a prescribed measurement position measured by the first measurement unit 11. In addition, for example, when measuring by the second measurement unit 12, the conveying mechanism 15 conveys the inductor element to be inspected to a prescribed measurement position measured by the second measurement unit 12. Moreover, the conveying mechanism 15 conveys the inductor elements determined to be qualified among the inductor elements after inspection to a position for encapsulation, and conveys the encapsulated inductor elements to a prescribed place in the next process.

[0074] The data processing control unit 10 is a functional unit that comprehensively controls the various functional units within the inspection unit 2 and performs various data processing tasks for the inspection of the DUT. For example, the data processing control unit 10 is a program processing unit that includes a processor such as a CPU (Central Processing Unit), storage devices such as ROM (Read-Only Memory), RAM (Random Access Memory), and flash memory, as well as peripheral circuits such as timers. Examples of program processing units include, for example, MCU (Microcontroller Unit) and FPGA (Field Programmable Gate Array).

[0075] The data processing control device 10 acquires the measurement results measured by the first measurement unit 11 and the second measurement unit 12, calculates an index representing the performance of the inductor element based on the acquired measurement results, and determines whether the inductor element under inspection is in good condition based on the calculated index. Here, the index representing the performance of the inductor element is, for example, the Q value.

[0076] As described above, when measuring the AC resistance of an inductor element using the two-terminal method, the measured value is affected by the resistance component generated by the measurement system using the two-terminal method. Therefore, when the data processing control device 10 of the inspection apparatus 2 in Embodiment 1 calculates an index (Q value) representing the performance of the inductor element under inspection based on the measurement results measured by the first measurement unit 11 and the second measurement unit 12, it uses a pre-generated learned model 35 as needed to correct the measured value Rs of the AC resistance measured by the second measurement unit 12.

[0077] Here, before providing a detailed explanation of the correction of the measured value Rs of the AC resistance using the learned model 35 performed by the data processing control device 10, the learned model 35 will be explained.

[0078] Model 35, once learned, is a model used to estimate the measured value Rs of the AC resistance of the inductor element of the object under inspection. For example, Model 35 is generated by a machine learning program based on a prescribed algorithm. Examples of such algorithms include multinomial regression, multiple regression, etc.

[0079] The learning completion model 35 of Implementation 1 is a function that enables a computer (MPU, etc.) to estimate the measured value Rs (third measured value) of the AC resistance of the DUT measured by the two-terminal method based on input data, wherein the input data includes the measured value Rdc4 (first measured value) of the DC resistance of the DUT measured by the four-terminal method, the measured value Rdc2 (second measured value) of the DC resistance of the DUT measured by the two-terminal method, and the measured value Ls (fourth measured value) of the inductance of the DUT measured by the two-terminal method.

[0080] In other words, the learned model 35 is a program that enables an information processing device (computer) to perform calculations on the input measurement data (the measured values ​​of DC resistance Rdc4, Rdc2 and the measured value of inductance Ls) based on prescribed learned parameters, and to output a value (estimated value) of AC resistance quantized based on the measurement data.

[0081] For example, the learned model 35 includes a first model representing a first resistance value Rc as the value of the resistance component generated by the measurement system using the two-terminal method, and a second model representing a second resistance value Rs0 as the value of the resistance component generated by the object under inspection (DUT).

[0082] Here, the first resistance value Rc includes, for example, the resistance component of the line consisting of cables, probes, etc., existing between the second measuring unit 12 and the DUT, and the resistance component generated by the contact state between the probe and the DUT (so-called contact resistance).

[0083] The first model representing the first resistance value Rc is, for example, a regression model that uses the measured DC resistance value Rdc4 (first measured value) obtained by the four-terminal method and the measured DC resistance value Rdc2 (second measured value) obtained by the two-terminal method as explanatory variables, and the first resistance value Rc of the resistance component generated by the two-terminal method measurement system as the target variable. In other words, the first model is a function that calculates the first resistance value Rc of the resistance component generated by the two-terminal method measurement system based on the measured DC resistance values ​​Rdc4 and Rdc2.

[0084] The second model representing the second resistance value Rs0 of the resistive component generated by the DUT is, for example, a regression model that uses the measured value Ls (fourth measured value) of the inductance determined by the two-terminal method as the explanatory variable and the second resistance value Rs0 of the resistive component (AC resistance) generated by the DUT as the target variable. In other words, the second model is a function that estimates the second resistance value Rs0 of the resistive component (AC resistance) generated by the DUT based on the measured value of the inductance determined by the two-terminal method.

[0085] When the first model representing the resistance component (Rc) generated by the measurement system using the two-terminal method is set as "Rc=f(Rdc4,Rdc2)" and the second model representing the resistance component (AC resistance Rs0) generated by the DUT is set as "Rs0=g(Ls)", the learned model 35, which is a function of the estimated value Rse used to calculate the measured value of AC resistance, can be represented by Rse=Rc+Rs0=f(Rdc4,Rdc2)+g(Ls).

[0086] That is, the estimated value Rse of the measured AC resistance is represented by the sum of the first resistance value Rc of the resistance component generated by the measurement system using the two-terminal method, obtained by the first model (Rc=f(Rdc4,Rdc2)) and the second resistance value Rs0 of the resistance component (AC resistance) generated by the DUT, obtained by the second model (Rs0=g(Ls)).

[0087] In the following description, the first model (Rc=f(Rdc4,Rdc2)) is denoted as "model f(Rdc4,Rdc2)" and the second model (Rs0=g(Ls)) is denoted as "model g(Ls)".

[0088] Models f(Rdc4, Rdc2) and g(Ls) include parameters. Parameters, for example, refer to the coefficients (weighting coefficients) of models f(Rdc4, Rdc2) and g(Ls). In the following explanation, the parameters mechanically adjusted in a manner that uses the learning measurement data 34_1 to 34_n described later as input to the learning program (a program based on the algorithm specified above) to calculate the estimated value Rse of the measured AC resistance will also be referred to as "learning-complete parameters".

[0089] Next, the learning-completed model generation device 3 that generates the learning-completed model 35 will be described.

[0090] Figure 2 This is a diagram illustrating an example of the configuration of the learning completed model generation device 3 in the inspection system 1 of Embodiment 1.

[0091] The learning-complete model generation device 3 is implemented by an information processing device (computer) such as a server or a personal computer (PC). It generates multiple learning measurement data 34_1 to 34_n (n is an integer of 2 or more) according to the installed learning-complete model generation program, and performs machine learning on the generated learning measurement data based on a prescribed algorithm, thereby generating the learning-complete model 35.

[0092] It should be noted that, for ease of explanation, in Figure 1The learning-complete model generation device 3 and the inspection device 2 are arranged side by side, but the learning-complete model generation device 3 and the inspection device 2 do not necessarily need to be located in the same place. For example, the inspection device 2 and the learning-complete model generation device 3 can also be located in different places and connected via a communication network such as LAN or the Internet. In this case, the inspection device 2 and the learning-complete model generation device 3 can also send and receive various data such as the measurement data measured by the inspection device 2 and the learning-complete model 35 via the communication network.

[0093] Furthermore, during DUT inspection, the inspection device 2 and the learned model generation device 3 may not be electrically connected to each other. For example, various data such as measurement data measured by the inspection device 2 and the learned model 35 may be exchanged via a storage medium such as a memory card before or after inspection.

[0094] Once Model 35 is learned, it can be circulated via the network or written to a computer-readable storage medium such as a memory card.

[0095] The learning-completed model generation device 3, as a functional block for generating the learning-completed model 35, includes, for example, a learning measurement data acquisition unit 31, a learning-completed model generation unit 32, and a storage unit 33. These functional blocks are implemented through the cooperation of hardware resources such as the CPU and memory that constitute the information processing device 3 as the learning-completed model generation device 3, and software (including various programs for generating the learning-completed model) installed on the aforementioned information processing device.

[0096] The learning measurement data acquisition unit 31 is a functional unit that acquires the learning measurement data 34_1 to 34_n required to generate the learned model 35.

[0097] Here, the learning measurement data 34_1 to 34_n are data pairs that establish corresponding values ​​for the AC resistance Rs (third measurement value) of the DUT measured by the two-terminal method, the DC resistance Rdc4 (first measurement value) of the DUT measured by the four-terminal method, the DC resistance Rdc2 (second measurement value) of the DUT measured by the two-terminal method, and the inductance Ls (fourth measurement value) of the DUT measured by the two-terminal method. In the following description, without distinguishing between each learning measurement data 34_1 to 34_n, it will only be referred to as "learning measurement data 34".

[0098] The learning measurement data acquisition unit 31 acquires data pairs, including data 41 (measured value Rdc4 of the DC resistance of the inductor element measured by the four-terminal method), data 42 (measured value Rdc2 of the DC resistance of the inductor element measured by the two-terminal method), data 44 (measured value Ls of the inductance of the DUT measured by the two-terminal method), and data 43 (measured value Rs of the AC resistance measured by the two-terminal method), via a wireless or wired communication medium (not shown) or a storage medium such as a memory card. The learning measurement data acquisition unit 31 acquires data pairs for each inductor element being inspected. For example, measurement results of the inductor element obtained in the past by the inspection device 2, or by measurements using a retainer performed manually, can be used as data pairs.

[0099] The learning measurement data acquisition unit 31 generates learning measurement data 34 by establishing a correspondence between the measured value Rs of the AC resistance included in the acquired data pair as the positive solution value and the measured values ​​Rdc4, Rdc2 of the DC resistance included in the data pair and the measured value Ls of the inductance included in the data pair. The learning measurement data acquisition unit 31 generates learning measurement data 34_1 to 34_n according to the measurement results of each inspected inductor element and stores them in the storage unit 33.

[0100] It should be noted that the learning measurement data acquisition unit 31 can also acquire the learning measurement data 34 that is measured manually and stored in another information processing device via a communication or storage medium instead of generating the learning measurement data 34 itself as described above.

[0101] The learning-completed model generation unit 32 is a functional unit that generates a learning-completed model 35 by performing machine learning on multiple learning measurement data 34_1 to 34_n acquired by the learning measurement data acquisition unit 31.

[0102] After learning is complete, the model generation unit 32 adjusts the parameters of model f(Rdc4,Rdc2) and model g(Ls) by performing machine learning on the measurement data 34_1 to 34_n used for learning, and obtains the learning parameters.

[0103] Specifically, after learning, the model generation unit 32 generates the learned model 35 (Rse = f(Rdc4, Rdc2) + g(Ls)) by performing machine learning on the models f(Rdc4, Rdc2) and g(Ls) separately. That is, the learned model generation unit 32 separately executes the process of adjusting the parameters of the model f(Rdc4, Rdc2) by machine learning and the process of adjusting the parameters of the model g(Ls) by machine learning. In addition, in order to reduce the influence caused by the error included in the defined value of the reference device used in the calibration process, the learned model generation unit 32 generates an adjustment value Rtng1 for adjusting the value estimated by the learned model 35.

[0104] As Figure 2 shown, the learned model generation unit 32 includes a first model generation unit 321, a second model generation unit 322, and an adjustment value calculation unit 323.

[0105] The second model generation unit 322 performs machine learning on the learning measurement data 34 with a small deviation in contact resistance when measured by the two-terminal method, thereby generating the model g(Ls). For example, the second model generation unit 322 performs machine learning on the learning measurement data 34 (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth to generate the model (Rs = g(Ls)). Here, preferably, the threshold value Rdth is a value as close to "0" as possible.

[0106] For example, by manually operating the holder on the table without using the inspection device 2, measurement data of the measured value Rdc4 of the DC resistance, the measured value Rdc2 of the DC resistance, the measured value Ls of the inductance, and the measured value Rs of the AC resistance of a plurality of DUTs with different inductance values are obtained. These measurement data are stored as learning measurement data 34 in the storage unit 33 of the learned model generation device 3. At this time, the learning measurement data 34 in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth is included in the learning measurement data 34.

[0107] At this time, the number of learning measurement data 34 used in the machine learning when generating the model g(Ls) can be smaller than the number of learning measurement data 34 used in the machine learning when generating the model f(Rdc4, Rdc2).

[0108] For example, it is only necessary to measure about 10 each of samples with different inductances (for example, L = 1 nH, 5 nH, 10 nH, etc.), and store the measurement data in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth as the learning measurement data 34 in the storage unit 33.

[0109] The second model generation unit 322 calculates a second resistance value Rs0 of the resistance component (alternating current resistance) generated by the DUT by inputting the measured value Rs of the alternating current resistance included in the learning measurement data 34, where the difference between the measured value Rdc4 of the direct current resistance obtained by using the above method and the measured value Rdc2 of the direct current resistance is less than the threshold value Rdth, into the regression model (g(Ls)). Next, the second model generation unit 322 calculates the difference (error) between the measured value Rs of the alternating current resistance, which is the correct solution value included in the learning measurement data 34, and the second resistance value Rs0 calculated by the regression model (g(Ls)). Then, the learned model generation unit 32, for example, uses the error backpropagation method to sequentially update the parameters of the regression model (Rs0 = g(Ls)) to make the calculated error smaller, thereby generating the learned parameters of the model g(Ls) and storing them in the storage unit 33.

[0110] The first model generation unit 321 generates a model f(Rdc4, Rdc2) by performing machine learning on the learning measurement data 34. Here, in the machine learning performed by the first model generation unit 321, it is preferable to use the learning measurement data 34 with a large deviation in contact resistance when measuring by the two-terminal method. For example, it is preferable that not only the learning measurement data 34 where the difference between the measured value Rdc4 of the direct current resistance and the measured value Rdc2 of the direct current resistance is less than the threshold value (|Rdc2 - Rdc4| < Rdth) is used for the machine learning performed by the first model generation unit 321, but also the learning measurement data 34 where the difference between the measured value Rdc4 of the direct current resistance and the measured value Rdc2 of the direct current resistance is greater than the threshold value (|Rdc2 - Rdc4| > Rdth) is used for the machine learning performed by the first model generation unit 321.

[0111] The first model generation unit 321 generates a model f(Rdc4, Rdc2) by performing machine learning on the measured value Ls of the inductance included in the learning measurement data 34.

[0112] For example, after performing the calibration process of the inspection device 2 using the reference device of the DUT, the inspection device 2 is used to measure a plurality of DUTs, and measurement data of the measured value Rdc4 of the direct current resistance, the measured value Rdc2 of the direct current resistance, the measured value Ls of the inductance, and the measured value Rs of the alternating current resistance are obtained. For example, hundreds to thousands of samples with different inductances are measured. These measurement data are stored as learning measurement data 34 in the storage unit 33 of the learned model generation device 3.

[0113] After adjusting the parameters of the model g(Ls), the first model generation unit 321 uses the adjusted parameters of the model g(Ls) to adjust the parameters of f(Rdc4, Rdc2).

[0114] For example, preferably, the first model generation unit 321 performs machine learning on the learning measurement data 34 in the model (f(Rdc4,Rdc2)+g(Ls)) represented by the sum of model f(Rdc4,Rdc2) and model g(Ls), while fixing the learned parameters of model g(Ls), thereby adjusting the parameters of f(Rdc4,Rdc2).

[0115] However, sometimes the relationship between inductance (Ls) and AC resistance (Rs) when using measurement data to obtain model g (Ls) differs from the relationship between inductance (Ls) and AC resistance (Rs) when inspected by inspection device 2 after calibration. For example, if the measurement error included in the measurement data used in the machine learning of model g (Ls) differs from the measurement error included in the defined value (AC resistance Rs) of the reference device in the calibration process of inspection device 2, or if the characteristics of the DUT used in the machine learning differ from the characteristics of the DUT during inspection, a deviation may occur between the estimated value of AC resistance calculated by inputting the measured value of the inductance of the DUT obtained by the calibrated inspection device 2 into model g (Ls) and the measured value of the AC resistance of the DUT.

[0116] Figure 3 This is a graph used to illustrate the deviation between the second resistance value Rs0 of the AC resistance calculated by model g(Ls) and the defined value (AC resistance Rs) of the DUT in the calibration process.

[0117] exist Figure 3 In the diagram, the horizontal axis represents the inductance value (Ls) of the DUT, and the vertical axis represents the AC resistance Rs of the DUT. Figure 3 In the figure, reference numeral 301 represents model g (Ls), and reference numeral 302 represents the definition value of the reference device (the definition value of the load reference).

[0118] For example, as described above, if the measurement error included in the measurement data used in the machine learning of model g(Ls) differs from the measurement error included in the defined value (AC resistance Rs) of the reference device in the calibration process of inspection device 2, such as Figure 3 As shown, a difference (offset ΔRs) occurs between the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by model g(Ls) and the defined value Rs of the reference device. The estimation accuracy of the learned model may decrease.

[0119] Therefore, in order to reduce Figure 3The influence of the shown offset ΔRs on the accuracy of the learned model. The learned model generation device 3 of Embodiment 1 calculates an adjustment value Rtng1 corresponding to the offset ΔRs and uses the adjustment value Rtng1 for machine learning of the model f(Rdc4, Rdc2).

[0120] Specifically, first, the adjustment value calculation unit 323 calculates the adjustment value Rtng1 based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value Rs0 calculated by inputting the measured value Ls of the inductance included in at least one learning measurement data 34 into the model g(Ls).

[0121] Here, the value of the AC resistance as a reference is the average value Rs_av1 of the measured values Rs of the AC resistance included in the learning measurement data 34 of multiple DUTs (|Rdc2 - Rdc4| < Rdth) where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth.

[0122] The adjustment value calculation unit 323 calculates the estimated value g(Ls_av1) of the second resistance value Rs0 by inputting the average value Ls_av1 of the measured values Ls of the inductance included in the learning measurement data 34 of multiple DUTs (|Rdc2 - Rdc4| < Rdth) where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Ls). Then, the difference between the average value Rs_av1 of the measured values Rs of the AC resistance and the estimated value g(Ls_av1) of the second resistance value Rs0 is calculated as the adjustment value Rtng1 (= Rs_av1 - g(Ls_av1)) and stored in the storage unit 33.

[0123] Alternatively, the value of the AC resistance as a reference can also be the defined value Rs_d1 of the AC resistance of the reference device in the calibration process. Information on the defined value Rs_d1 of the AC resistance only needs to be pre-stored in the storage unit 33 of the learned model generation device 3 after the calibration process of the above inspection system 1.

[0124] In this case, the adjustment value calculation unit 323 calculates the estimated value g(Ls_av1) of the second resistance value Rs0 by inputting the average value Ls_av1 of the measured values Ls of the inductance included in the learning measurement data 34 of multiple DUTs (|Rdc2 - Rdc4| < Rdth) where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Ls). Then, the difference between the defined value Rs_d1 of the AC resistance and the estimated value of the second resistance value Rs0 is calculated as the adjustment value Rtng1 (= Rs_d1 - g(Ls_av1)) and stored in the storage unit 33.

[0125] The first model generation unit 321 adjusts the parameters of model f(Rdc4,Rdc2) in a model represented by the sum of model f(Rdc4,Rdc2), model g(Ls), and adjustment value Rtng1, while keeping the parameters of the already learned model g(Ls) fixed.

[0126] Specifically, the first model generation unit 321 calculates the estimated value Rse of the AC resistance measurement by inputting the measured values ​​Rdc4, Rdc2, and Ls of the DC resistance included in the learning measurement data 34 into the regression model (Rse=f(Rdc4,Rdc2)+g(Ls)+Rtng1). Next, the first model generation unit 321 calculates the difference (error) between the calculated estimated value Rse of the AC resistance measurement and the measured value Rs of the AC resistance included in the learning measurement data 34 as the positive solution value. Next, the first model generation unit 321 updates the parameters of model f(Rdc4,Rdc2) in the above regression model (Rse=f(Rdc4,Rdc2)+g(Ls)+Rtng1) successively by error backpropagation method to reduce the calculated error, generates the learned parameters of model f(Rdc4,Rdc2) and stores them in the storage unit 33.

[0127] Therefore, a model f(Rdc4,Rdc2) that reduces the influence of offset ΔRs can be generated.

[0128] The learning-complete model generation unit 32 stores the regression model (Rse=f(Rdc4,Rdc2)+g(Ls)) including the learning-complete parameters adjusted by the above method as the learning-complete model 35 in the storage unit 33.

[0129] It should be noted that the adjustment value Rtng1 is not included in the model registered as the learned model 35 (Rse=f(Rdc4,Rdc2)+g(Ls)). That is, the adjustment value Rtng1 is only considered when adjusting the parameters of model f(Rdc4,Rdc2).

[0130] The storage unit 33 is a functional unit for storing various data such as the learning measurement data 34_1 to 34_n required to generate the learned model 35 and the generated learned model 35.

[0131] The storage unit 33 is configured to be accessible from the outside. For example, by communicating with the learning-completed model generation device 3, the inspection device 2 can read and obtain the learning-completed model 35 from the storage unit 33. Furthermore, by communicating with the learning-completed model generation device 3, the inspection device 2 can write measurement result data, etc., into the storage unit 33.

[0132] Figure 4 This is a flowchart illustrating the process of generating a learned model 35 by the learned model generation device 3 in Implementation 1.

[0133] like Figure 4 As shown, firstly, in the model generation device 3 after learning, the learning measurement data acquisition unit 31 acquires learning measurement data 34_1 to 34_n (step S1). Specifically, as described above, the learning measurement data acquisition unit 31 acquires data pairs including the measured values ​​of DC resistance Rdc4 and Rdc2 of inductor elements that have been previously checked and the measured value of AC resistance Rs, and based on the acquired data pairs, assigns the measured value of AC resistance Rs to the measured values ​​of DC resistance Rdc4 and Rdc2, thereby generating the learning measurement data 34.

[0134] The learning measurement data 34, as described above, includes measurement data of the DUT measured manually using a retainer and measurement data of the DUT measured by the calibrated inspection device 2. Preferably, each learning measurement data 34 is assigned information that can identify whether it is measurement data measured manually using a retainer or measurement data measured by the inspection device 2.

[0135] Next, the learning-completed model generation device 3 determines whether the required number of learning measurement data 34 for generating the learning-completed model 35 has been generated (step S2). For example, in the learning-completed model generation device 3, the number of data points 34 required for generating the learning-completed model 35 is preset, and the learning-completed model generation device 3 increments the generation count by 1 each time it generates the learning measurement data 34. Then, the learning-completed model generation device 3 determines whether the required number of learning measurement data 34 has been generated by checking whether the counted generation count has reached the preset number of data points.

[0136] If the required number of learning measurement data 34 is not generated (step S2: NO), the learning model generation device 3 returns to step S1 after learning is completed, and repeatedly acquires data pairs related to the measurement results of the new inductor element to generate learning measurement data 34 related to the inductor element (steps S1, S2).

[0137] On the other hand, after the required number of learning measurement data 34 has been generated (step S2: YES), the learning model generation device 3 uses the multiple learning measurement data 34 generated in step S1 to perform machine learning (step S3).

[0138] Figure 5 This is a flowchart illustrating the process of machine learning (step S3) in implementation method 1.

[0139] Here, as an example, a case will be described where the average value Rs_av1 of the measured values Rs of the AC resistance included in the learning measurement data 34 (|Rdc2 - Rdc4| < Rdth) of a plurality of DUTs in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth is set as the reference AC resistance value.

[0140] First, the second model generation unit 322 extracts the learning measurement data 34 with a small deviation in the DC resistance from the learning measurement data 34_1 to 34_n (step S31). For example, as described above, the learning measurement data 34 in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance in the learning measurement data 34 measured manually using the holder is less than the threshold value (|Rdc2 - Rdc4| < Rdth) is extracted.

[0141] Next, the second model generation unit 322 generates the model g(Ls) using the learning measurement data 34 extracted in step S31 by the method described above (step S32).

[0142] Next, the adjustment value calculation unit 323 calculates the average value Rs_av1 of the measured values Rs of the AC resistance (step S33). Specifically, the adjustment value calculation unit 323 extracts the learning measurement data 34 in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance in the learning measurement data 34 measured by the inspection system 1 that has undergone the calibration process is less than the threshold value (|Rdc2 - Rdc4| < Rdth). The adjustment value calculation unit 323 calculates the average value Rs_av1 of the measured values Rs of the AC resistance included in the extracted plurality of learning measurement data 34 (|Rdc2 - Rdc4| < Rdth).

[0143] Next, the adjustment value calculation unit 323 calculates the average value Ls_av1 of the measured values Ls of the inductance (step S34). Specifically, the adjustment value calculation unit 323 calculates the average value Ls_av1 of the measured values Ls of the inductance included in the plurality of learning measurement data 34 (|Rdc2 - Rdc4| < Rdth) used in the calculation of the average value Rs_av1 of the measured values Rs of the AC resistance in step S33.

[0144] Next, the adjustment value calculation unit 323 calculates the adjustment value Rtng1 (step S35). Specifically, the adjustment value calculation unit 323 generates the adjustment value Rtng1 (=Rs_av1-g(Ls_av1)) by subtracting the value g(Ls_av1) obtained by inputting the average value of the measured value of AC resistance Rs calculated in step S33, Rs_av1, into the model g(Ls) generated in step S32.

[0145] Next, the first model generation unit 321 extracts the learning measurement data 34 with large deviations in DC resistance from the learning measurement data 34_1 to 34_n (step S36). For example, the first model generation unit 321 extracts at least the learning measurement data 34 obtained by the calibration-processed inspection system 1, where the difference between the measured value of DC resistance Rdc4 and the measured value of DC resistance Rdc2 is greater than a threshold (|Rdc2-Rdc4|>Rdth).

[0146] Next, the first model generation unit 321 sets a model that takes into account the adjustment value Rtng1 (step S37). Specifically, as described above, the first model generation unit 321 sets a model represented by the sum of model f(Rdc4,Rdc2), model g(Ls) generated in step S32, and adjustment value Rtng1 calculated in step S35 (Rse=f(Rdc4,Rdc2)+g(Ls)+Rtng1).

[0147] Next, the first model generation unit 321 uses the model (Rse=f(Rdc4,Rdc2)+g(Ls)+Rtng1) set in step S37, and adjusts the parameters of the model f(Rdc4,Rdc2) while keeping the parameters of the already learned model g(Ls) fixed.

[0148] Through the above processing flow, the parameters of model g(Ls) and model f(Rdc4,Rdc2) are adjusted separately to generate the learned model 35 (Rse=f(Rdc4,Rdc2)+g(Ls)).

[0149] Next, as Figure 4As shown, the learned model 35 generated in step S3 is registered to the inspection device 2 (step S4). For example, based on the user's operation of the operation unit 13 of the inspection device 2 or the input device (e.g., touch panel, keyboard, mouse, etc.) of the learned model generation device 3, the learned model generation device 3 sends the learned model 35 stored in the storage unit 33 to the inspection device 2, and the inspection device 2 stores the received learned model 35 in the storage unit within the data processing control device 10. It should be noted that, as described above, the registration of the learned model 35 to the inspection device 2 can also be performed using a storage medium such as a memory card.

[0150] It should be noted that the learning-complete model generation program used to enable the computer (information processing device) that serves as the learning-complete model generation device 3 to execute the above steps (S1 to S4) can be distributed via a network or written to a non-transitory computer readable medium such as a memory card.

[0151] Using the methods described above, a learned model 35 is generated for checking inductor components.

[0152] Next, the correction of the measured value Rs of the AC resistance of the model 35 after learning, performed by the inspection device 2, will be explained in detail.

[0153] Similar to the case of acquiring the measurement data 34 for learning, a calibration process is performed in the inspection device 2 before the automatic inspection of the DUT, which is equipped with the learned model 35, begins. Therefore, in the inspection device 2, a decrease in inspection accuracy may occur due to errors included in the defined values ​​of the reference devices. Therefore, the inspection device 2, like the learned model generation device 3, calculates an adjustment value Rtng2 and uses this adjustment value Rtng2 to adjust the value estimated by the learned model 35.

[0154] Figure 6 This is a diagram illustrating an example of the configuration of the data processing control device 10 in the inspection device 2 of Embodiment 1.

[0155] like Figure 6 As shown, the data processing control unit 10 of the inspection device 2 includes, for example, a data acquisition unit 21, a storage unit 22, an estimation unit 23, a correction unit 24, a judgment unit 25, and an adjustment value calculation unit 26. These functional units are implemented, for example, in the program processing unit that serves as the data processing control unit 10, by the CPU performing various operations according to the program stored in the memory and controlling peripheral circuits such as counters.

[0156] The data acquisition unit 21 is a functional unit that acquires various data required to calculate an index (Q value) representing the performance of an inductor element to be inspected.

[0157] The data acquisition unit 21 acquires, for example, the measured value Rdc4 of the DC resistance of the DUT measured by the first measurement unit 11 by the four-terminal method and stores it in the storage unit 22. The data acquisition unit 21 acquires, for example, the measured value Rdc2 of the DC resistance of the DUT measured by the second measurement unit 12 by the two-terminal method, the measured value Rs of the AC resistance of the DUT measured by the second measurement unit 12 by the two-terminal method, and the measured value Ls of the inductance of the DUT measured by the second measurement unit 12 by the two-terminal method, and stores them in the storage unit 22 as the measurement data 50 of the inspection object. In addition, the data acquisition unit 21 acquires, for example, the learned model 35 generated by the learned model generation device 3 and stores it in the storage unit 22.

[0158] The storage unit 22 is a functional unit for storing various data required to calculate an index (Q value) representing the performance of an inductor element to be inspected and the calculated Q value and the like.

[0159] As described above, in the storage unit 22, the measured values Rdc4, Rdc2 of the DC resistance of the inductor element acquired by the data acquisition unit 21, the measured value Rs of the AC resistance of the inductor element, the measured value Ls of the inductance of the inductor element, and the learned model 35 are respectively stored. In addition, in the storage unit 22, for example, the estimated value Rse of the measured value of the AC resistance described later, the estimated value Rc of the resistance component generated by the measurement system using the two-terminal method, the value Rsr of the AC resistance, the Q value, and the adjustment value Rtng1 are respectively stored.

[0160] The adjustment value calculation unit 26 calculates an adjustment value Rtng2 corresponding to Figure 3 the offset ΔRs shown. Specifically, the adjustment value calculation unit 26 calculates an adjustment value Rtng2 corresponding to the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value Rs0 calculated by inputting the measured value Ls of the inductance included in at least one measurement data 50 into the model g(Ls).

[0161] Here, the value of the AC resistance as a reference is the average value Rs_av2 of the measured values Rs of the AC resistance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth.

[0162] The adjustment value calculation unit 26 calculates the estimated value g(Ls_av2) of the second resistance value Rs0 by inputting the average value Ls_av2 of the measured values Ls of the inductance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) for which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Ls). The adjustment value calculation unit 26 calculates the difference between the average value Rs_av2 of the measured value Rs of the AC resistance and the estimated value g(Ls_av2) of the second resistance value Rs0 as the adjustment value Rtng2 (= Rs_av2 - g(Ls_av2)), and stores it in the storage unit 22.

[0163] Alternatively, the value of the AC resistance as a reference may also be the defined value Rs_d2 of the AC resistance of the reference device in the calibration process of the inspection device 2. Information on the defined value Rs_d2 of the AC resistance may be pre-stored in the storage unit 22 of the data processing control device 10 after the calibration process of the inspection device 2 described above.

[0164] In this case, the adjustment value calculation unit 26 calculates the estimated value g(Ls_av2) of the second resistance value Rs0 by inputting the average value Ls_av2 of the measured values Ls of the inductance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) for which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Ls). The adjustment value calculation unit 26 calculates the difference between the defined value Rs_d2 of the AC resistance and the estimated value g(Ls_av2) of the second resistance value Rs0 as the adjustment value Rtng2 (= Rs_d2 - g(Ls_av2)), and stores it in the storage unit 22.

[0165] The estimation unit 23 is a functional unit that estimates the measured value of the AC resistance of the inductor element to be inspected. The estimation unit 23 calculates the estimated value Rse of the measured value of the AC resistance corresponding to the measured values Rdc4 and Rdc2 of the DC resistance of the inductor element to be inspected obtained by the data acquisition unit 21 based on the learned model 35 and the adjustment value Rtng2 stored in the storage unit 22.

[0166] Specifically, the estimation unit 23 adjusts the value obtained by inputting the measured values Rdc4 and Rdc2 of the DC resistance and the measured value Ls of the inductance of the inductor element to be inspected obtained by the data acquisition unit 21 into the learned model 35 (function) by the adjustment value Rtng2, and stores the adjusted value as the estimated value Rse of the AC resistance in the storage unit 22.

[0167] For example, the estimation unit 23 stores the estimated value of AC resistance Rse as the storage unit 22 by adding the adjustment value Rtng2 to the model (f(Rdc4,Rdc2)+g(Ls)+Rtng2) obtained by inputting the measurement data 50 including the measured values ​​of DC resistance Rdc4, Rdc2 and inductance Ls.

[0168] The calibration unit 24 is a functional unit used to calibrate the measured value Rs of AC resistance. The calibration unit 24 performs the following calibration process: based on the estimated value Rse of AC resistance measurement, and based on the measured values ​​Rdc4 and Rdc2 of DC resistance obtained by the data acquisition unit 21, the AC resistance measurement value Rs is calibrated, and the calibrated AC resistance measurement value Rs is output as the AC resistance value Rsr of the DUT.

[0169] More specifically, the calibration unit 24 first calculates the resistance component Rc generated by the measurement system using the two-terminal method based on the measured values ​​Rdc4 and Rdc2 of the DC resistance of the inductor element of the inspection object obtained by the data acquisition unit 21, according to the above model f(Rdc4,Rdc2). The calculated resistance component Rc data is stored, for example, in the storage unit 22.

[0170] Next, the calibration unit 24 corrects the measured value Rs of the AC resistance of the inductor element under test based on the estimated value Rse of the measured AC resistance and the resistance component Rc generated by the measurement system using the two-terminal method, and outputs the corrected measured value Rs as the value Rsr of the AC resistance of the inductor element under test.

[0171] For example, the calibration unit 24 calculates the error |Rse-Rs| between the estimated value Rse of the AC resistance measured by the estimation unit 23 and the measured value Rs (the third measured value) of the AC resistance acquired by the data acquisition unit 21, and evaluates the error |Rse-Rs|. Specifically, the calibration unit 24 compares the error |Rse-Rs| with a threshold value Rth. The threshold value Rth is an arbitrary value that is preset.

[0172] Here, when the error |Rse-Rs| is less than the threshold Rth, it can be considered that the estimation accuracy of the AC resistance based on the learned model 35 is high relative to the measurement results of the inductor element of the object under inspection. That is, it can be considered that the estimation accuracy of the resistance component Rc generated by the measurement system using the two-terminal method based on the model f(Rdc4,Rdc2) included in the learned model 35 is high.

[0173] Therefore, when the error |Rse-Rs| is less than the threshold Rth, the correction unit 24 performs correction processing. Specifically, the correction unit 24 calculates the resistance component Rc generated by the measurement system using the two-terminal method using model f(Rdc4,Rdc2), and uses the calculated resistance component Rc generated by the measurement system using the two-terminal method to correct the measured value Rs (third measured value) of AC resistance.

[0174] For example, the calibration unit 24 first inputs (substitutes) the measured values ​​Rdc4 and Rdc2 of the DC resistance of the inductor element under test into the model f(Rdc4,Rdc2) as the resistance component Rc generated by the measurement system using the two-terminal method. Next, the calibration unit 24 outputs the AC resistance value Rsr (=Rs-Rc) as the value obtained by subtracting the resistance component Rc generated by the measurement system using the two-terminal method from the measured value Rs (third measured value) of AC resistance obtained by the data acquisition unit 21.

[0175] On the other hand, when the error |Rse-Rs| is greater than the threshold Rth, it can be considered that the estimation accuracy of the AC resistance based on the learned model 35 is low relative to the inductor element of the object under inspection. That is, it can be considered that the estimation accuracy of the resistance component Rc generated by the measurement system using the two-terminal method based on the model f(Rdc4,Rdc2) included in the learned model 35 is low.

[0176] In this case, if the resistance component Rc generated by the measurement system using the two-terminal method is calculated using model f(Rdc4,Rdc2), and the measured value Rs (the third measured value) of the AC resistance is corrected using the calculated resistance component Rc generated by the measurement system using the two-terminal method, then it will be an incorrect correction and the value Rsr of the AC resistance may not be properly obtained.

[0177] Therefore, if the error |Rse-Rs| is greater than the threshold Rth, the correction unit 24 does not perform correction processing, but outputs the measured value Rs of the AC resistance obtained by the data acquisition unit 21 as the value Rsr (=Rs) of the AC resistance of the inductor element under inspection.

[0178] The determination unit 25 is a functional unit used to determine whether the DUT (inductor element) is in good condition. The determination unit 25 calculates the Q value (Q=ωL / Rsr) as an indicator of the performance of the inductor element under test, based on the AC resistance value Rsr of the inductor element under test output from the calibration unit 24 and the measured inductance value Ls of the inductor element under test.

[0179] The determination unit 25 determines whether the inductor element of the inspection object is in good condition by, for example, comparing the calculated Q value with a predetermined reference value. The determination unit 25 controls the conveyance mechanism 15 to encapsulate the DUT determined to be a qualified product into a state ready for shipment by a packaging device (not shown).

[0180] Next, the process flow of the inspection process of the DUT by the inspection device 2 will be described.

[0181] Figure 7 It is a flowchart showing the process flow of the inspection by the inspection device 2 of Embodiment 1.

[0182] First, the user performs a calibration process of the inspection device 2 (step S11). In the calibration process, the user determines a reference device (gold device) from the DUTs to be inspected. Next, the user manually measures the electrical characteristics of the reference device using a holder in an environment with less noise, and determines the measured value as the defined value of the reference device. Next, the user measures the reference device by the inspection device 2 under multiple references (for example, open circuit reference, short circuit reference, and load reference), thereby obtaining the measured values under each reference. Then, the inspection device 2 generates a calibration formula for correcting the measured values to reduce the error between the measured values under each reference and the defined values obtained in advance, and stores it in the storage device inside the inspection device 2. Thus, a series of calibration processes are completed. At this time, the defined value Rs_d2 of the AC resistance of the reference device used in the calibration process is stored in the storage unit 22 of the data processing control device 10.

[0183] After the calibration process, for example, when the user operates the operation unit 13 of the inspection device 2 to instruct the execution of the inspection of the DUT (inductor element), the data processing control device 10 starts an adjustment value calculation process for obtaining an adjustment value Rtng2 as a preparation process for the inspection of the DUT (step S12).

[0184] Figure 8 It is a flowchart showing the process flow of the adjustment value calculation process (step S12) of Embodiment 1.

[0185] Here, as an example, the case where the average value Rs_av2 of the measured values Rs of the AC resistance included in the measurement data 50 of multiple DUTs (|Rdc2 - Rdc4| < Rdth) with the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance less than the threshold value Rdth is used as the AC resistance value set as a reference will be described.

[0186] In step S12, firstly, the data processing control device 10 acquires measurement data of multiple inductor elements of the object under inspection (step S121). For example, the data processing control device 10 acquires measurement data 50, including the measured values ​​of DC resistance Rdc4 and Rdc2, AC resistance Rs, and inductance Ls of the inductor elements of the object under inspection, by controlling the first measurement unit 11 and the second measurement unit 12.

[0187] The data processing control device 10 stores the measured values ​​of DC resistance Rdc4, Rdc2, AC resistance Rs, and inductance Ls obtained by the above method as measurement data 50 of the inductor element to be inspected in the storage unit 22.

[0188] Next, the data processing control device 10 extracts multiple measurement data 50 with small deviations in contact resistance from the measurement data 50 acquired in step S121 (step S122). Specifically, the adjustment value calculation unit 26 extracts multiple DUT measurement data 50 from the measurement data 50 stored in the storage unit 22, where the difference between the measured value Rdc4 and the measured value Rdc2 of DC resistance is less than the threshold value Rdth (|Rdc2-Rdc4|). <Rdth)。

[0189] Next, the adjustment value calculation unit 26 calculates the average value Rs_av2 of the AC resistance measurement values ​​Rs included in the plurality of measurement data 50 obtained in step S122 (step S123). In addition, the adjustment value calculation unit 26 calculates the average value Ls_av2 of the inductance measurement values ​​Ls included in the plurality of measurement data 50 obtained in step S122 (step S124).

[0190] Next, the adjustment value calculation unit 26 calculates the estimated value g(Ls_av2) of the second resistance value Rs0 by inputting the average value Ls_av2 of the measured inductance value Ls calculated in step S124 into the model g(Ls) (step S125).

[0191] Next, the adjustment value calculation unit 26 calculates the adjustment value Rtng2 (=Rs_av2-g(Ls_av2)) by subtracting the estimated value g(Ls_av2) of the second resistance value Rs0 calculated in step S125 from the average value Rs_av2 of the measured AC resistance value Rs calculated in step S123 (step S126). The adjustment value Rtng2 can be obtained through the above processing flow.

[0192] Furthermore, when the value of the AC resistance used as a reference is the defined value Rs_d2 of the AC resistance of the reference device in the calibration process of the inspection device 2, in step S123, the adjustment value calculation unit 26 reads the defined value Rs_d2 of the AC resistance of the reference device from the storage unit 22. Then, in step S126, the adjustment value calculation unit 26 only needs to subtract the estimated value g(Ls_av2) of the second resistance value Rs0 calculated in step S125 from the defined value Rs_d2 of the AC resistance of the reference device to calculate the adjustment value Rtng2 (=Rs_d2-g(Ls_av2)).

[0193] After step S12, the data processing control device 10 begins the inspection process of the inductor element of the object to be inspected (step S13).

[0194] Figure 9 This is a flowchart illustrating the process of inspection (step S13) performed by inspection device 2 in Embodiment 1.

[0195] In step S13, firstly, the data processing control device 10 controls the first measuring unit 11 to measure the DC resistance of the inductor element of the test object using the four-terminal method (step S131). For example, the data processing control device 10 controls the conveying mechanism 15 according to the instruction signal from the operation unit 13 to convey the inductor element of the test object to a predetermined measuring position in the first measuring unit 11. Then, the data processing control device 10 controls the first measuring unit 11 to measure the DC resistance of the inductor element of the test object using the four-terminal method, obtaining the measured value Rdc4 of the DC resistance.

[0196] Next, the data processing control device 10 controls the second measuring unit 12 to measure the DC resistance of the inductor element of the test object using a two-terminal method (step S132). For example, the data processing control device 10 controls the conveying mechanism 15 to convey the inductor element of the test object to a predetermined measuring position in the second measuring unit 12. Afterward, the data processing control device 10 controls the second measuring unit 12 to measure the DC resistance of the inductor element of the test object using a two-terminal method, and obtains the measured value Rdc2 of the DC resistance.

[0197] Next, the data processing control device 10 controls the second measurement unit 12 to measure the AC resistance Rs and inductance Ls of the inductor element under test using a two-terminal method (step S133). For example, with the inductor element under test positioned in the same measurement position as in step S132, the data processing control device 10 controls the second measurement unit 12 to measure the AC resistance of the inductor element under test and obtain the AC resistance Rs and inductance Ls respectively.

[0198] In steps S131 to S133, the measured values ​​of DC resistance Rdc4 and Rdc2, AC resistance Rs and inductance Ls obtained by the data processing control device 10 are stored in the storage unit 22 as measurement data 50 of the inductor element to be inspected.

[0199] Next, the data processing control device 10 uses the adjustment value Rtng2 calculated in step S12 to calculate the estimated value Rse of the measured AC resistance of the inductor element under inspection (step S134). Specifically, the estimation unit 23 inputs the measured values ​​Rdc4 and Rdc2 of the DC resistance obtained in steps S131 and S132 to the learning model 35 using the method described above, and outputs the estimated value Rse (= f(Rdc4, Rdc2) + g(Ls) + Rtng2) of the measured AC resistance as the value obtained by adding the adjustment value Rtng2 calculated in step S12 to the value output from the learning model 35.

[0200] It should be noted that, after step S133, the DC resistance of the inductor element under test can also be measured using the two-terminal method to obtain the measured value Rdc2 of the DC resistance again. In this case, the estimation unit 23 can also compare the measured value Rdc2 of the DC resistance obtained in step S132 with the measured value Rdc2 of the DC resistance obtained again after step S133, and use the smaller measured value Rdc2 of the DC resistance to calculate the estimated value Rse of the measured value of the AC resistance in step S134.

[0201] Next, the calibration unit 24 determines whether the difference |Rse-Rs| between the estimated value Rse of the AC resistance calculated in step S134 and the measured value Rs of the AC resistance obtained in step S133 is less than the threshold value Rth (step S135).

[0202] If the difference |Rse-Rs| is less than the threshold Rth (step S135: Yes), the correction unit 24 calculates the resistance component Rc generated by the measurement system using the two-terminal method using model f(Rdc4, Rdc2) (step S136). Specifically, the correction unit 24 obtains the resistance component Rc generated by the measurement system using the two-terminal method by inputting the measured values ​​Rdc4 and Rdc2 of the DC resistance obtained in steps S131 and S132 into model f(Rdc4, Rdc2).

[0203] Next, the calibration unit 24 performs the following calibration process: using the resistance component Rc calculated in step S136 by the measurement system using the two-terminal method, the measured value Rs of the AC resistance obtained in step S133 is corrected, and the corrected value is output as the value Rsr of the AC resistance (step S137). Specifically, the calibration unit 24 outputs the value Rsr (=Rs-Rc) of the AC resistance obtained by subtracting the resistance component Rc calculated in step S136 from the measured value Rs of the AC resistance obtained in step S133.

[0204] On the other hand, if the difference |Rse-Rs| is greater than the threshold Rth (step S135: no), the correction unit 24 does not perform correction processing, but outputs the measured value Rs of the AC resistance obtained in step S133 as the value Rsr (=Rs) of the AC resistance (step S138).

[0205] Next, the determination unit 25 calculates the Q value of the inductor element under inspection based on the AC resistance value Rsr output from the calibration unit 24 in step S137 or S138 and the measured inductance value Ls obtained in step S133 (step S139). Then, based on the Q value calculated in step S139, the determination unit 25 determines whether the inductor element under inspection is good or not (step S140). Inductor elements determined to be qualified are transported by the conveying mechanism 15 and packaged.

[0206] It should be noted that the inspection program used to make the computer (information processing device) that serves as the data processing control device 10 execute the above steps (S131 to S140) can be transmitted via a network or written to a non-transitory computer readable medium such as a memory card.

[0207] In the inspection system 1 of Embodiment 1, the learning-completed model generation device 3 generates a learning-completed model 35 by performing machine learning on multiple learning measurement data 34_1 to 34_n. The learning-completed model 35 includes: a model f (Rdc4, Rdc2) that calculates a first resistance Rc as a resistance component generated by a measurement system using the two-terminal method, and a model g (Ls) that calculates a second resistance value Rs0 as a resistance component generated by a DUT (inductor element). The multiple learning measurement data 34_1 to 34_n are generated by marking the measured value Rs of the AC resistance of the DUT measured by the two-terminal method with the measured value Rdc4 of the DC resistance of the DUT measured by the four-terminal method, the measured value Rdc2 of the DC resistance of the DUT measured by the two-terminal method, and the measured value Ls of the inductance of the DUT measured by the two-terminal method.

[0208] Therefore, even if the relationship between the measured DC resistance values ​​Rdc4, Rdc2 and the measured AC resistance value Rs of the inductor element is nonlinear, a learning model 35 (function) can be obtained that appropriately represents the relationship between the measured DC resistance values ​​Rdc4, Rdc2 and the measured inductance value Ls and the measured AC resistance value Rs.

[0209] Furthermore, after learning Model 35, Model f(Rdc4,Rdc2) is a regression model that uses the measured values ​​of DC resistance Rdc4 and Rdc2 as explanatory variables and the values ​​of the resistance components generated by the measurement system using the two-terminal method as the target variable. Model g(Ls) is a regression model that uses the measured value of inductance Ls as explanatory variable and the values ​​of the resistance components generated by the DUT as the target variable.

[0210] Therefore, the learned model 35 can be represented by a simpler function, thus avoiding the black box effect of the learned model 35 that is a concern in machine learning. Furthermore, in the regression model (g(Ls)) representing the value of the resistive component generated by the DUT, the resistive component generated by the DUT can be estimated with higher accuracy by using the measured value of inductance Ls as the explanatory variable.

[0211] Furthermore, as described above, after the learning process is complete, the model generation device 3 calculates the offset ΔRs (refer to the above). Figure 3The adjustment value Rtng1 is used to perform machine learning on model f(Rdc4,Rdc2). That is, after learning, the model generation device 3 calculates the adjustment value Rtng1 corresponding to the offset ΔRs based on the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured value Ls of the inductance included in at least one learning measurement data into model g(Ls). Then, after learning, the model generation device 3 adjusts the parameters of model f(Rdc4,Rdc2) while keeping the parameters of model g(Ls) fixed in the model (f(Rdc4,Rdc2)+g(Ls)+Rtng1) represented by the sum of model f(Rdc4,Rdc2), model g(Ls), and adjustment value Rtng1.

[0212] Therefore, even if there is a difference (offset ΔRs) between the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured data into the model g(Ls) and the defined value Rs of the reference device, the learned model 35 can be generated in a way that eliminates this difference. Thus, a learned model with higher estimation accuracy can be generated.

[0213] Furthermore, as the reference value of AC resistance, the adjustment value Rtng1 is calculated by using the average value Rs_av1 of the AC resistance measurements Rs included in the learning measurement data 34 of multiple DUTs where the difference between the measured DC resistance Rdc4 and the measured DC resistance Rdc2 is less than the threshold Rdth, thereby further reducing the error of model g(Ls).

[0214] Furthermore, deviations from the estimated value Rse of the AC resistance based on the learned model relative to the defined value of the aforementioned reference device may also occur during inspection by the inspection device 2. Therefore, the inspection device 2 of Embodiment 1 calculates an adjustment value Rtng2 based on the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured value Ls of the inductance included in at least one measurement data 50 into the model g (Ls). Based on the adjustment value Rtng2, the value calculated by the learned model 35 is adjusted, and the estimated value Rse of the measured value of the AC resistance is output.

[0215] Therefore, even if the aforementioned offset ΔRs occurs due to differences between the measurement environment when measuring the learning measurement data 34 used in the machine learning of the completed model 35 and the measurement environment when performing the calibration process of the inspection device 2, the estimated value Rse of the AC resistance can be calculated in a way that eliminates this difference. As a result, the reliability of the inspection of electronic components can be improved.

[0216] Furthermore, as the reference value of AC resistance, the adjustment value Rtng2 is calculated by using the average value Rs_av2 of the AC resistance measurements Rs included in the measurement data 50 of multiple DUTs where the difference between the measured DC resistance Rdc4 and the measured DC resistance Rdc2 is less than the threshold Rdth, thereby further reducing the error of model g(Ls).

[0217] Furthermore, the adjustment value Rtng2 is calculated using the defined value of the AC resistance of the reference device in the calibration process, which serves as the reference AC resistance value. Thus, the contact resistance can be corrected from the chip of the first object being inspected during inspection.

[0218] <Implementation Method 2> Figure 10 This is a diagram illustrating an example of the configuration of the learning completed model generation device 3A in the inspection system 1A of Embodiment 2.

[0219] In the second embodiment, the inspection system 1A differs from the inspection system 1 of embodiment 1 in that, in the regression model (second model) representing the resistance component generated by the DUT, the DC resistance Rdc4 measured by the four-terminal method is used as the explanatory variable instead of the measured value Ls of the inductance. In other aspects, it is the same as the inspection system 1 of embodiment 1.

[0220] In the learning completed model generation device 3A, the learning measurement data acquisition unit 31A, similar to the learning measurement data acquisition unit 31 in Embodiment 1, acquires data pairs for each inspected inductor element via wireless or wired communication (not shown) or storage medium such as a memory card. These data pairs include data 41, which is the measured value of the DC resistance of the inductor element measured by the four-terminal method Rdc4; data 42, which is the measured value of the DC resistance of the inductor element measured by the two-terminal method Rdc2; and data 43, which is the measured value of the AC resistance measured by the two-terminal method Rs.

[0221] The learning measurement data acquisition unit 31A generates a learning measurement data 34A by establishing a correspondence between the measured value Rs of the AC resistance included in the acquired data pair as the positive solution value and the measured values ​​Rdc4 and Rdc2 of the DC resistance included in the data pair. The learning measurement data acquisition unit 31A generates learning measurement data 34A_1 to 34A_n according to the measurement results of each inspected inductor element and stores them in the storage unit 33.

[0222] The learned model generation unit 32A, similar to the learned model generation unit 32 in Embodiment 1, generates a learned model 35A (Rse = f(Rdc4, Rdc2) + g(Rdc4)) by separately learning a first model (Rc = f(Rdc4, Rdc2)) and a second model (Rs0 = g(Rdc4)). That is, the learned model generation unit 32A separately performs the process of adjusting the parameters of the first model (Rc = f(Rdc4, Rdc2)) by machine learning and the process of adjusting the parameters of the second model (Rs0 = g(Rdc4)) by machine learning. It should be noted that the second model (Rs0 = g(Rdc4)) is also denoted as "model g(Rdc4)".

[0223] The learned model generation unit 32A generates the model g(Rdc4) by performing machine learning on the learning measurement data 34A with a small deviation (contact resistance deviation) of the measured value of the DC resistance, and generates f(Rdc4, Rdc2) by performing machine learning on the learning measurement data 34A with a large deviation of the measured value of the DC resistance.

[0224] Specifically, the second model generation unit 322A of the learned model generation unit 32A generates a model (Rs = g(Rdc4)) by performing machine learning on the learning measurement data 34A (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth. Here, preferably, the threshold value Rdth is a value as close to "0" as possible.

[0225] For example, the second model generation unit 322A uses the measured value Rdc4 of the DC resistance and the measured value Rs of the AC resistance in the measurement data (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance in the learning measurement data 34A measured by manual operation using a holder on the table without using the inspection device 2A is less than the threshold value Rdth to generate a model (Rs = g(Rdc4)).

[0226] The learned model generation device 3A, similar to the learned model generation device 3 in Embodiment 1, calculates an adjustment value Rtng1 by the adjustment value calculation unit 323A, and the first model generation unit 321 uses the adjustment value Rtng1 to perform machine learning on the model f(Rdc4, Rdc2).

[0227] Specifically, the adjustment value calculation unit 323A calculates an adjustment value Rtng1 corresponding to the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value Rs0 calculated by inputting the measured value Rdc4 of the DC resistance included in at least one learning measurement data 34A into the model g(Rdc4).

[0228] Here, the value of the AC resistance as a reference can also be, as in the first embodiment, the average value Rs_av1 of the measured values Rs of the AC resistance included in the learning measurement data 34A of a plurality of DUTs where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth (|Rdc2 - Rdc4| < Rdth). In this case, the adjustment value calculation unit 323A calculates the estimated value g(Rdc4_av1) of the second resistance value Rs0 by inputting the average value Rdc4_av1 of the measured values Rdc4 of the DC resistance included in the learning measurement data 34A of a plurality of DUTs where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth (|Rdc2 - Rdc4| < Rdth) into the model g(Rdc4). Next, the adjustment value calculation unit 323A calculates the difference between the average value Rs_av1 of the measured values Rs of the AC resistance and the estimated value g(Rdc4_av1) of the second resistance value Rs0 as the adjustment value Rtng1 (= Rs_av1 - g(Rdc4_av1)), and stores it in the storage unit 33.

[0229] Alternatively, the value of the AC resistance as a reference can also be, as in the first embodiment, the defined value Rs_d1 of the AC resistance of the reference device in the calibration process. In this case, the adjustment value calculation unit 323A calculates the estimated value g(Rdc4_av1) of the second resistance value Rs0 by inputting the average value Rdc4_av1 of the measured values Rdc4 of the DC resistance included in the learning measurement data 34A of a plurality of DUTs where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth (|Rdc2 - Rdc4| < Rdth) into the model g(Rdc4). Next, the difference between the defined value Rs_d1 of the AC resistance and the estimated value of the second resistance value Rs0 is calculated as the adjustment value Rtng1 (= Rs_d1 - g(Rdc4_av1)), and stored in the storage unit 33.

[0230] The first model generation unit 321 adjusts the parameters of the model f(Rdc4, Rdc2) in a state where the parameters of the already learned model g(Rdc4) are fixed in the model represented by the sum of the model f(Rdc4, Rdc2), the model g(Rdc4), and the adjustment value Rtng1.

[0231] Specifically, the first model generation unit 321 calculates an estimated value Rse of the measured value of the AC resistance by inputting the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance included in the learning measurement data 34A (for example, the learning measurement data 34A where at least |Rdc2 - Rdc4| > Rdth) into the regression model (Rse = f(Rdc4, Rdc2) + g(Rdc4) + Rtng1). Next, the first model generation unit 321 calculates the difference (error) between the calculated estimated value Rse of the measured value of the AC resistance and the measured value Rs of the AC resistance as the correct value included in the above learning measurement data 34A. Next, the first model generation unit 321 successively updates the parameters of the model f(Rdc4, Rdc2) in the above regression model (Rse = f(Rdc4, Rdc2) + g(Rdc4) + Rtng1) by, for example, the error backpropagation method to make the calculated error smaller, generates the learned parameters of the model f(Rdc4, Rdc2), and stores them in the storage unit 33. At this time, the learned parameters of the model g(Rdc4) are not updated.

[0232] The learned model generation unit 32A stores the regression model (Rse = f(Rdc4, Rdc2) + g(Rdc4)) including the learned parameters adjusted by the above method as the learned model 35A in the storage unit 33.

[0233] Here, similar to the learned model generation device 3 of the first embodiment, the adjustment value Rtng1 is not included in the model (Rse = f(Rdc4, Rdc2) + g(Rdc4)) registered as the learned model 35A. That is, the adjustment value Rtng1 is only considered when adjusting the parameters of the model f(Rdc4, Rdc2).

[0234] The overall process of the learned model generation method of the second embodiment is the same as that of the learned model generation method of the first embodiment (steps S1 to step S4). A part of the processing in the machine learning (step S3) of the second embodiment is different from the processing of the machine learning of the first embodiment.

[0235] Figure 11 It is a flowchart showing the process of the machine learning (step S3) of the second embodiment.

[0236] Here, as an example, a case where the average value Rs_av1 of the measured values Rs of the AC resistance included in the learning measurement data 34A (|Rdc2 - Rdc4| < Rdth) of a plurality of DUTs where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth is used as the reference AC resistance value is described.

[0237] First, the second model generation unit 322A extracts the learning measurement data 34A with a small deviation in the DC resistance from the learning measurement data 34A_1 to 34A_n (step S31). For example, as described above, the learning measurement data 34A in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance in the learning measurement data 34A measured manually using the retainer is less than the threshold value is extracted (|Rdc2 - Rdc4| < Rdth).

[0238] Next, the second model generation unit 322A generates the model g(Rdc4) using the learning measurement data 34A extracted in step S31 by the above method (step S32A).

[0239] Next, the adjustment value calculation unit 323A calculates the average value Rs_av1 of the measured value Rs of the AC resistance by the same method as the adjustment value calculation unit 323 in the first embodiment (step S33).

[0240] Next, the adjustment value calculation unit 323A calculates the average value Rdc4_av1 of the measured value Rdc4 of the DC resistance (step S34A). Specifically, the adjustment value calculation unit 323A calculates the average value Rdc4_av1 of the measured value Rdc4 of the DC resistance included in the plurality of learning measurement data 34A (|Rdc2 - Rdc4| < Rdth) used in the calculation of the average value Rs_av1 of the measured value Rs of the AC resistance in step S33.

[0241] Next, the adjustment value calculation unit 323A calculates the adjustment value Rtng1 (step S35A). Specifically, the adjustment value calculation unit 323A generates a value obtained by subtracting the value g(Rdc4_av1) obtained by inputting the average value Rdc4_av1 of the measured value Rdc4 of the DC resistance calculated in step S34A into the model g(Rdc4) generated in step S32A from the average value Rs_av1 of the measured value Rs of the AC resistance calculated in step S33 as the adjustment value Rtng1 (= Rs_av1 - g(Rdc4_av1)).

[0242] Next, the first model generation unit 321 extracts the learning measurement data 34A with a large deviation in the DC resistance from the learning measurement data 34A_1 to 34A_n (step S36). For example, the first model generation unit 321 extracts the learning measurement data 34A measured by the inspection system 1 that has undergone the calibration process (for example, the learning measurement data 34A with at least |Rdc2 - Rdc4| > Rdth).

[0243] Next, the first model generation unit 321 sets a model considering the adjustment value Rtng1 (step S37A). Specifically, as described above, the first model generation unit 321 sets a model represented by the sum of the model f(Rdc4, Rdc2), the model g(Rdc4) generated in step S32A, and the adjustment value Rtng1 calculated in step S35A (Rse = f(Rdc4, Rdc2) + g(Rdc4) + Rtng1).

[0244] Next, the first model generation unit 321 uses the method described above to adjust the parameters of the model f(Rdc4, Rdc2) while fixing the parameters of the already learned model g(Rdc4) using the model (Rse = f(Rdc4, Rdc2) + g(Rdc4) + Rtng1) set in step S37A.

[0245] Through the above processing flow, the parameters of the model g(Rdc4) and the parameters of the model f(Rdc4, Rdc2) are adjusted separately to generate the learned model 35A (Rse = f(Rdc4, Rdc2) + g(Rdc4)).

[0246] Next, the correction of the measured value Rs of the AC resistance using the learned model 35A performed by the inspection device 2A of Embodiment 2 will be described in detail.

[0247] Figure 12 FIG. is an example showing the configuration of the data processing control device 10A in the inspection device 2A of Embodiment 2.

[0248] As Figure 12 shown, the data processing control device 10A of the inspection device 2A stores the learned model 35A (Rse = f(Rdc4, Rdc2) + g(Rdc4)) in the storage unit 22. The data processing control device 10A generates an adjustment value Rtng2 through the adjustment value calculation unit 26A in the same manner as the data processing control device 10 of Embodiment 1.

[0249] Specifically, the adjustment value calculation unit 26A calculates an adjustment value Rtng2 corresponding to the difference between the value of the AC resistance as a reference and the estimated value of the second resistance value Rs0 calculated by inputting the measured value Rdc4 of the DC resistance included in at least one measurement data 50 into the model g(Rdc4).

[0250] Here, the value of the AC resistance as a reference is the average value Rs_av2 of the measured values Rs of the AC resistance included in the measurement data 50 (|Rdc2 - Rdc4| < Rdth) of multiple DUTs where the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth.

[0251] The adjustment value calculation unit 26A calculates the estimated value g(Rdc4_av2) of the second resistance value Rs0 by inputting the average value Rdc4_av2 of the measured values Rdc4 of the DC resistance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) for which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Rdc4). The adjustment value calculation unit 26A calculates the difference between the average value Rs_av2 of the measured value Rs of the AC resistance and the estimated value g(Rdc4_av2) of the second resistance value Rs0 as the adjustment value Rtng2 (= Rs_av2 - g(Rdc4_av2)), and stores it in the storage unit 22.

[0252] Alternatively, the value of the AC resistance as a reference may also be the defined value Rs_d2 of the AC resistance of the reference device in the calibration process of the inspection device 2A. Information on the defined value Rs_d2 of the AC resistance only needs to be pre-stored in the storage unit 22 of the data processing control device 10A after the above-mentioned calibration process of the inspection device 2A.

[0253] In this case, the adjustment value calculation unit 26A calculates the estimated value g(Rdc4_av2) of the second resistance value Rs0 by inputting the average value Rdc4_av2 of the measured values Rdc4 of the DC resistance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) for which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth into the model g(Rdc4). The adjustment value calculation unit 26A calculates the difference between the defined value Rs_d2 of the AC resistance and the estimated value g(Rdc4_av2) of the second resistance value Rs0 as the adjustment value Rtng2 (= Rs_d2 - g(Rdc4_av2)), and stores it in the storage unit 22.

[0254] Similar to the estimation unit 23 in Embodiment 1, the estimation unit 23A adjusts the value obtained by inputting the measured values Rdc4 and Rdc2 of the DC resistance of the inductor element to be inspected, which are acquired by the data acquisition unit 21, into the learned model 35A (function) by the adjustment value Rtng2, and stores the adjusted value as the estimated value Rse of the AC resistance in the storage unit 22.

[0255] For example, the estimation unit 23A stores, in the storage unit, as the estimated value Rse of the AC resistance, the value calculated by inputting the measured values Rdc4 and Rdc2 of the DC resistance included in the measurement data 50 into the model (f(Rdc4, Rdc2) + g(Rdc4) + Rtng2) obtained by adding the adjustment value Rtng2 to the learned model 35A.

[0256] Next, the process of the inspection process of the DUT by the inspection device 2A of Embodiment 2 will be described.

[0257] The overall process of the inspection process of the inspection device 2A of Embodiment 2 is the same as the inspection process of Embodiment 1 (Steps S11 to Step S13). A part of the adjustment value calculation process (Step S12) of Embodiment 2 is different from the adjustment value calculation process of Embodiment 1.

[0258] Figure 13 It is a flowchart showing the process of the adjustment value calculation process (Step S12) in the inspection process of Embodiment 2.

[0259] Here, as an example, the case where the average value Rs_av2 of the measured values Rs of the AC resistance included in the measurement data 50 of a plurality of DUTs (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value Rdth is used as the reference AC resistance value will be described.

[0260] In Step S12, first, the data processing control device 10A acquires the measurement data of a plurality of inductor elements to be inspected (Step S121). For example, the data processing control device 10A acquires the measurement data 50 including the measured values Rdc4, Rdc2 of the DC resistance and the measured value Rs of the AC resistance of the inductor element to be inspected by controlling the first measurement unit 11 and the second measurement unit 12.

[0261] The measured values Rdc4, Rdc2 of the DC resistance and the measured value Rs of the AC resistance acquired by the data processing control device 10A by the above method are stored in the storage unit 22 as the measurement data 50 of the inductor element to be inspected.

[0262] Next, the data processing control device 10A extracts a plurality of measurement data 50 with small contact resistance deviation from the measurement data 50 acquired in Step S121 (Step S122). Specifically, the adjustment value calculation unit 26A extracts the measurement data 50 of a plurality of DUTs in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance stored in the storage unit 22 is less than the threshold value Rdth (|Rdc2 - Rdc4| < Rdth).

[0263] Next, the adjustment value calculation unit 26A calculates the average value Rs_av2 of the measured values Rs of the AC resistance included in the plurality of measurement data 50 acquired in Step S122 (Step S123). In addition, the adjustment value calculation unit 26A calculates the average value Rdc4_av2 of the DC resistance Rdc4 included in the plurality of measurement data 50 acquired in Step S122 (Step S124A).

[0264] Next, the adjustment value calculation unit 26A calculates an estimated value g(Rdc4_av2) of the second resistance value Rs0 by inputting the average value Rdc4_av2 of the DC resistance Rdc4 calculated in step S124A into the model g(Rdc4) (step S125A).

[0265] Next, the adjustment value calculation unit 26A calculates an adjustment value Rtng2 (= Rs_av2 - g(Rdc4_av2)) by subtracting the estimated value g(Rdc4_av2) of the second resistance value Rs0 calculated in step S125A from the average value Rs_av2 of the measured value Rs of the AC resistance calculated in step S123 (step S126A). Through the above processing flow, the adjustment value Rtng2 can be obtained.

[0266] Here, when the value of the AC resistance as a reference is the defined value Rs_d2 of the AC resistance of the reference device in the calibration process of the inspection device 2A, in step S123, the adjustment value calculation unit 26A reads the defined value Rs_d2 of the AC resistance of the reference device from the storage unit 22. Then, in step S126A, the adjustment value calculation unit 26A only needs to calculate the adjustment value Rtng2 (= Rs_d2 - g(Rdc4_av2)) by subtracting the estimated value g(Rdc4_av2) of the second resistance value Rs0 calculated in step S125A from the defined value Rs_d2 of the AC resistance of the reference device.

[0267] It should be noted that the inspection process (step S13) performed by the data processing control device 10A is the same as the inspection process performed by the data processing control device 10 in the first embodiment, so the description thereof is omitted.

[0268] As described above, the learned model generation device 3A according to the second embodiment generates a model g(Rdc4) by performing machine learning on the learning measurement data 34A (|Rdc2 - Rdc4| < Rdth) in which the difference between the measured value Rdc4 of the DC resistance and the measured value Rdc2 of the DC resistance is less than the threshold value, and generates a model f(Rdc4, Rdc2) by performing machine learning on the learning measurement data 34A.

[0269] Thereby, similarly to the learned model generation device 3 in the first embodiment, a model g(Rdc4) and a model f(Rdc4, Rdc2) with higher accuracy can be obtained.

[0270] Furthermore, after learning, the model generation device 3A calculates an adjustment value Rtng1 corresponding to the offset ΔRs based on the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured value Rdc4 of the DC resistance included in at least one learning measurement data into the model g (Rdc4). Then, after learning, the model generation device 3A adjusts the parameters of model f (Rdc4,Rdc2) in the model (f(Rdc4,Rdc2)+g(Rdc4)+Rtng1) represented by the sum of model f (Rdc4,Rdc2), model g (Rdc4), and adjustment value Rtng1, while keeping the parameters of model g (Rdc4) fixed.

[0271] Therefore, even when there is a difference (offset ΔRs) between the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured data into model g (Rdc4) and the defined value Rs of the reference device, the learned model 35A can be generated in a way that eliminates this difference. Thus, a learned model with higher estimation accuracy can be generated.

[0272] Furthermore, the deviation of the estimated value Rse of the AC resistance based on the learned model 35A from the defined value of the aforementioned reference device may also occur during inspection by the inspection device 2A. Therefore, the inspection device 2A of Embodiment 2 calculates an adjustment value Rtng2 based on the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance value Rs0 of the AC resistance Rs calculated by inputting the measured value Rdc4 of the DC resistance included in at least one measurement data 50 into the model g (Rdc4). The value calculated by the learned model 35A is adjusted based on the adjustment value Rtng2, and the estimated value Rse of the measured value of the AC resistance is output as the estimated value Rse of the measured value of the AC resistance.

[0273] Therefore, similar to the inspection device 2 in Embodiment 1, the reliability of inspection of electronic components can be improved.

[0274] <Extended Implementation Methods> The invention described above is based on the implementation methods and has been specifically described by the inventors of this application. However, the invention is not limited thereto, and various modifications can be made without departing from its spirit.

[0275] For example, the model (learned model 35) used by the inspection device 2 during inspection can be any function representing the correspondence between the measured values ​​of DC resistance Rdc4 and Rdc2 and the measured value of AC resistance Rs, or it can be a model generated by a method other than machine learning. For example, the inspection device 2 can also use a model that includes models f(Rdc4, Rdc2) and g(Ls) (or g(Rdc4)) whose coefficients have been adjusted by a method other than machine learning, and perform the inspection of the inductor element in the same way as described above.

[0276] Furthermore, in the above embodiment, the inspection device 2 (2A) is exemplified as a device that integrates the data processing control device 10 (10A), the first measuring unit 11, the second measuring unit 12, the operation unit 13, the output unit 14, and the conveying mechanism 15 into one unit. However, some of the components constituting the inspection device 2 (2A) can also be separately configured from other components. For example, the data processing control device 10 (10A), the operation unit 13, and the output unit 14 can be implemented by a first device (e.g., an information processing device such as a PC), and the first measuring unit 11, the second measuring unit 12, and the conveying mechanism 15 can be implemented by a second device different from the first device. In this case, the first device and the second device can be connected via a wired or wireless network.

[0277] The flowchart above is an example used to illustrate an action and is not limited to it. That is, the steps shown in each diagram are specific examples and are not limited to this process. For example, the order of some processes can be changed, other processes can be inserted between processes, and some processes can be performed in parallel.

[0278] Explanation of reference numerals in the attached figures 1. 1A: Inspection system; 2. 2A: Inspection device; 3. 3A: Model generation device after learning; 10. 10A: Data processing and control device; 11: First measurement unit; 12: Second measurement unit; 13: Operation unit; 14: Output unit; 15: Conveying mechanism; 21: Data acquisition unit; 22: Storage unit; 23: Estimation unit; 24: Calibration unit; 25: Judgment unit; 26. 26A: Adjustment value calculation unit; 31. 31A: Measurement data acquisition unit for learning; 32. 32A: Model generation unit after learning; 33: Storage unit; 34. 34_1~34_n, 34A, 34A_1~34A_n: Measurement data for learning; 35. 35A: Model after learning; 50: Measurement data; 3 21: First model generation unit; 322, 322A: Second model generation unit; 323, 323A: Adjustment value calculation unit; Rc: Resistance component generated by the measurement system using the two-terminal method; Rdc2: Measured value of DC resistance measured by the two-terminal method; Rdc4: Measured value of DC resistance measured by the four-terminal method; Rs: Measured value of AC resistance measured by the two-terminal method; Rse: Estimated value of the measured value of AC resistance measured by the two-terminal method; Rs0: Value of the resistance component generated by the DUT (second resistance value); Rs_d1, Rs_d2: Defined values ​​of AC resistance; Rsr: Value of AC resistance; Rtng1, Rtng2: Adjustment values; Rth, Rdth: Thresholds.

Claims

1. An inspection device comprising: The data acquisition unit acquires measurement data, which includes a first measured value of the DC resistance of the test object measured by the four-terminal method, a second measured value of the DC resistance of the test object measured by the two-terminal method, a third measured value of the AC resistance of the test object measured by the two-terminal method, and a fourth measured value of the inductance of the test object measured by the two-terminal method. The storage unit stores the learned model, which enables the computer to function in a manner that calculates the third measurement value based on the first measurement value, the second measurement value, and the fourth measurement value. The estimation unit calculates an estimated value of the third measurement value corresponding to the first measurement value, the second measurement value, and the fourth measurement value obtained by the data acquisition unit, based on the learning completed model stored in the storage unit. as well as The adjustment value calculation unit calculates an adjustment value to adjust the value calculated by the learned model. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the fourth measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The adjustment value calculation unit calculates the adjustment value based on the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the fourth measurement values ​​included in the measurement data into the second model. The estimation unit adjusts the value calculated by the learned model based on the adjustment value, and calculates it as the estimated value of the third measurement value.

2. The inspection device according to claim 1, wherein, The reference AC resistance value is the average of the third measurement value included in multiple measurement data of the object being measured, where the difference between the first and second measurement values ​​is less than a threshold. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average of the fourth measurement value, which is included in a plurality of measurement data of the measured object where the difference between the first measurement value and the second measurement value is less than the threshold, into the second model, and calculates the difference between the average of the third measurement value and the estimated value of the second resistance value as the adjustment value.

3. The inspection device according to claim 1, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the fourth measurement value included in the measurement data of a plurality of the measurement objects whose difference between the first measurement value and the second measurement value is less than a threshold into the second model, and calculates the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

4. An inspection device comprising: The data acquisition unit acquires measurement data, which includes a first measured value of the DC resistance of the object being measured by the four-terminal method, a second measured value of the DC resistance of the object being measured by the two-terminal method, and a third measured value of the AC resistance of the object being measured by the two-terminal method. The storage unit stores the learned model, which enables the computer to function in a manner that calculates the third measurement value based on the input first and second measurement values. The estimation unit calculates an estimated value of the third measurement value corresponding to the first measurement value and the second measurement value obtained by the data acquisition unit, based on the learning completed model stored in the storage unit. as well as The adjustment value calculation unit calculates an adjustment value to adjust the value calculated by the learned model. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the first measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The adjustment value calculation unit calculates the adjustment value based on the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the first measured values ​​included in the measurement data into the second model. The estimation unit adjusts the value calculated by the learned model based on the adjustment value, and calculates it as the estimated value of the third measurement value.

5. The inspection device according to claim 4, wherein, The reference AC resistance value is the average of the third measurement value included in multiple measurement data of the object being measured, where the difference between the first and second measurement values ​​is less than a threshold. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the first measured value, which is included in a plurality of measured data of the measured object where the difference between the first measured value and the second measured value is less than the threshold, into the second model, and calculates the difference between the average value of the third measured value and the estimated value of the second resistance value as the adjustment value.

6. The inspection device according to claim 4, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the first measured value included in the measurement data of a plurality of the measured objects whose difference between the first measured value and the second measured value is less than a threshold into the second model, and calculates the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

7. The inspection device according to claim 1 or 4, wherein, The inspection device also includes a calibration unit, which performs the following calibration process: based on the estimated value of the third measured value, and based on the first measured value and the second measured value obtained by the data acquisition unit, the third measured value is calibrated, and the calibrated third measured value is output as the value of the AC resistance of the object being measured.

8. The inspection device according to claim 7, wherein, The correction unit calculates the resistance component generated by the measurement system using the two-terminal method according to the first model based on the first measurement value and the second measurement value obtained by the data acquisition unit, and corrects the third measurement value based on the resistance component generated by the measurement system using the two-terminal method as the correction process.

9. The inspection device according to claim 8, wherein, The correction unit calculates the error between the estimated value of the third measured value calculated by the estimation unit and the third measured value obtained by the data acquisition unit, and performs the correction process if the error is less than a predetermined threshold.

10. An inspection method, comprising: The first step is to acquire measurement data, which includes a first measured value of the DC resistance of the object being measured by the four-terminal method, a second measured value of the DC resistance of the object being measured by the two-terminal method, a third measured value of the AC resistance of the object being measured by the two-terminal method, and a fourth measured value of the inductance of the object being measured by the two-terminal method. The second step involves calculating an estimated value of the third measurement corresponding to the first, second, and fourth measurement values ​​obtained in the first step, based on a learned model that enables the computer to function in a manner that infers the third measurement value based on the first, second, and fourth measurement values ​​input. The third step is to calculate the adjustment value used to adjust the value calculated by the learned model. The fourth step is to adjust the estimated value of the third measurement calculated in the second step based on the adjustment value. as well as The fifth step involves correcting the third measured value based on the estimated value of the third measured value adjusted in the fourth step, using the first measured value and the second measured value obtained in the first step, and outputting the corrected third measured value as the AC resistance value of the object being measured. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the fourth measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The third step includes calculating the adjustment value based on the difference between the value of the AC resistance used as a reference and an estimated value of the second resistance calculated by inputting at least one of the fourth measurement values ​​included in the measurement data into the second model.

11. The inspection method according to claim 10, wherein, The reference AC resistance value is the average of the third measurement value included in multiple measurement data of the object being measured, where the difference between the first and second measurement values ​​is less than a threshold. The third step includes: calculating an estimated value of the second resistance value by inputting the average of the fourth measurement value, which is included in a plurality of measurement data of the object being measured where the difference between the first measurement value and the second measurement value is less than the threshold, into the second model; and calculating the difference between the average of the third measurement value and the estimated value of the second resistance value as the adjustment value.

12. The inspection method according to claim 10, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The third step includes: calculating an estimated value of the second resistance value by inputting the average of the fourth measurement value included in the measurement data of a plurality of the measurement objects whose difference between the first measurement value and the second measurement value is less than a threshold into the second model; and calculating the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

13. An inspection method, comprising: The first step is to obtain measurement data, which includes a first measured value of the DC resistance of the object being measured by the four-terminal method, a second measured value of the DC resistance of the object being measured by the two-terminal method, and a third measured value of the AC resistance of the object being measured by the two-terminal method. The second step involves calculating an estimated value of the third measurement corresponding to the first and second measurement values ​​obtained in the first step, based on a learned model that enables the computer to function in a manner that infers the third measurement value based on the first and second measurement values ​​input. The third step is to calculate the adjustment value used to adjust the value calculated by the learned model. The fourth step is to adjust the estimated value of the third measurement calculated in the second step based on the adjustment value. as well as The fifth step involves correcting the third measured value based on the estimated value of the third measured value adjusted in the fourth step, using the first measured value and the second measured value obtained in the first step, and outputting the corrected third measured value as the AC resistance value of the object being measured. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the first measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The third step includes calculating the adjustment value based on the difference between the value of the AC resistance used as a reference and an estimated value of the second resistance calculated by inputting at least one of the first measured values ​​included in the measurement data into the second model.

14. The inspection method according to claim 13, wherein, The reference AC resistance value is the average of the third measurement value included in multiple measurement data of the object being measured, where the difference between the first and second measurement values ​​is less than a threshold. The third step includes: calculating an estimated value of the second resistance value by inputting the average of the first measured values ​​included in a plurality of measured data of the measured object, where the difference between the first measured value and the second measured value is less than the threshold, into the second model; and calculating the difference between the average of the third measured value and the estimated value of the second resistance value as the adjustment value.

15. The inspection method according to claim 13, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The third step includes: calculating an estimated value of the second resistance value by inputting the average of the first measured values ​​included in the measurement data of a plurality of the measured objects for which the difference between the first measured value and the second measured value is less than a threshold into the second model; and calculating the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

16. An inspection program that causes a computer to perform the steps of the inspection method according to claim 10 or 13.

17. A learning-completed model generation device, comprising: The learning measurement data acquisition unit acquires learning measurement data, which establishes a correspondence between the third measured value of the AC resistance of the test object measured by the two-terminal method and the first measured value of the DC resistance of the test object measured by the four-terminal method, the second measured value of the DC resistance of the test object measured by the two-terminal method, and the fourth measured value of the inductance of the test object measured by the two-terminal method. as well as After the learning process is complete, the model generation unit performs machine learning on the learning measurement data to generate a completed learning model that enables a computer to calculate the third measurement value based on input data including the first measurement value, the second measurement value, and the fourth measurement value. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the fourth measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The learned model generation unit has: The first model generation unit generates the first model by performing machine learning on the first measurement value and the second measurement value included in the learning measurement data of the measurement object; The second model generation unit generates the second model by performing machine learning on the fourth measurement value included in the learning measurement data of the measurement object where the difference between the first measurement value and the second measurement value is less than a threshold. as well as The adjustment value calculation unit calculates an adjustment value corresponding to the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the fourth measurement values ​​included in the learning measurement data into the second model. The first model generation unit adjusts the parameters of the first model while fixing the parameters of the second model in a model represented by the sum of the first model, the second model, and the adjustment value.

18. The learning-completed model generation apparatus according to claim 17, wherein, The value of the AC resistance used as a reference is the average of the third measurement values ​​included in the learning measurement data of a plurality of the measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average of the fourth measurement value included in the learning measurement data of a plurality of measurement objects whose difference between the first measurement value and the second measurement value is less than the threshold into the second model, and calculates the difference between the average of the third measurement value and the estimated value of the second resistance value as the adjustment value.

19. The learning-completed model generation apparatus according to claim 17, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the fourth measurement value included in the learning measurement data of a plurality of measurement objects whose difference between the first measurement value and the second measurement value is less than the threshold into the second model, and calculates the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

20. A learning-completed model generation device, comprising: The learning measurement data acquisition unit acquires learning measurement data, which establishes a correspondence between the third measured value of the AC resistance of the test object measured by the two-terminal method, the first measured value of the DC resistance of the test object measured by the four-terminal method, and the second measured value of the DC resistance of the test object measured by the two-terminal method; and After the learning process is complete, the model generation unit performs machine learning on the learning measurement data to generate a learned model that enables a computer to function in a manner that calculates the third measurement value based on input data including the first and second measurement values. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the first measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The learned model generation unit has: The first model generation unit generates the first model by performing machine learning on the first and second measurement values ​​included in the learning measurement data of the measurement object where the difference between the first and second measurement values ​​is greater than a threshold. The second model generation unit generates a second model by performing machine learning on the first measurement values ​​included in the learning measurement data of the measurement object where the difference between the first measurement value and the second measurement value is less than the threshold; and The adjustment value calculation unit calculates an adjustment value corresponding to the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the first measured values ​​included in the learning measurement data into the second model. The first model generation unit adjusts the parameters of the first model while keeping the parameters of the second model fixed in a model represented by the sum of the first model, the second model, and the adjustment value.

21. The learning-completed model generation apparatus according to claim 20, wherein, The value of the AC resistance used as a reference is the average of the third measurement values ​​included in the learning measurement data of a plurality of the measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the first measurement value included in the learning measurement data of a plurality of measurement objects where the difference between the first measurement value and the second measurement value is less than the threshold into the second model, and calculates the difference between the average value of the third measurement value and the estimated value of the second resistance value as the adjustment value.

22. The learning-completed model generation apparatus according to claim 20, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The adjustment value calculation unit calculates the estimated value of the second resistance value by inputting the average value of the first measured value included in the learning measurement data of a plurality of the measured objects whose difference between the first measured value and the second measured value is less than the threshold into the second model, and calculates the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

23. A method for generating a learned model, comprising: The first step is to acquire measurement data for learning purposes. The measurement data for learning purposes establishes a correspondence between the third measured value of the AC resistance of the test object measured by the two-terminal method, the first measured value of the DC resistance of the test object measured by the four-terminal method, the second measured value of the DC resistance of the test object measured by the two-terminal method, and the fourth measured value of the inductance of the test object measured by the two-terminal method. as well as The second step involves performing machine learning on the learning measurement data obtained in the first step to generate a fully learned model that enables the computer to function by calculating the third measurement value based on input data including the first measurement value, the second measurement value, and the fourth measurement value. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the fourth measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The second step includes: The third step involves performing machine learning on the fourth measurement value included in the learning measurement data of the measurement object where the difference between the first measurement value and the second measurement value is less than a threshold, to generate the second model. The fourth step involves calculating an adjustment value corresponding to the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the fourth measurement values ​​included in the learning measurement data into the second model; and The fifth step involves performing machine learning on the first and second measurement values ​​included in the learning measurement data of the measurement object where the difference between the first and second measurement values ​​is greater than a threshold, to generate the first model. The fifth step includes adjusting the parameters of the first model while keeping the parameters of the second model fixed in the model represented by the sum of the first model, the second model, and the adjustment value.

24. The method for generating a learned model according to claim 23, wherein, The value of the AC resistance used as a reference is the average of the third measurement values ​​included in the learning measurement data of a plurality of the measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The fourth step includes: calculating an estimated value of the second resistance value by inputting the average of the fourth measurement values ​​included in the learning measurement data of a plurality of the measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold value into the second model; and calculating the difference between the average of the third measurement value and the estimated value of the second resistance value as the adjustment value.

25. The method for generating a learned model according to claim 23, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The fourth step includes: calculating an estimated value of the second resistance value by inputting the average of the fourth measurement values ​​included in the learning measurement data of a plurality of the measurement objects whose difference between the first measurement value and the second measurement value is less than the threshold value into the second model; and calculating the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

26. A method for generating a learned model, comprising: The first step is to acquire the measurement data for learning purposes. The measurement data for learning purposes establishes a correspondence between the third measured value of the AC resistance of the object being measured by the two-terminal method, the first measured value of the DC resistance of the object being measured by the four-terminal method, and the second measured value of the DC resistance of the object being measured by the two-terminal method. as well as The second step involves performing machine learning on the learning measurement data obtained in the first step to generate a fully learned model that enables the computer to function in a manner that calculates the third measurement value based on input data including the first and second measurement values. The learned model is represented by the sum of a first model and a second model. The first model uses the first and second measured values ​​as explanatory variables and the first resistance value, which is the value of the resistance component generated by the measurement system using the two-terminal method, as the target variable. The second model uses the first measured value as an explanatory variable and the second resistance value, which is the value of the resistance component generated by the measured object, as the target variable. The second step includes: The fourth step is to generate the second model by performing machine learning on the first measurement value included in the learning measurement data of the measurement object where the difference between the first measurement value and the second measurement value is less than a threshold. The fifth step involves calculating an adjustment value corresponding to the difference between the value of the AC resistance used as a reference and the estimated value of the second resistance calculated by inputting at least one of the first measured values ​​included in the learning measurement data into the second model; and The sixth step involves performing machine learning on the first and second measurement values ​​included in the learning measurement data of the measurement object where the difference between the first and second measurement values ​​is greater than a threshold, to generate the first model. The sixth step includes adjusting the parameters of the first model while keeping the parameters of the second model fixed in the model represented by the sum of the first model, the second model, and the adjustment value.

27. The method for generating a learned model according to claim 26, wherein, The value of the AC resistance used as a reference is the average of the third measurement values ​​included in the learning measurement data of a plurality of the measurement objects for which the difference between the first measurement value and the second measurement value is less than the threshold. The fifth step includes: calculating an estimated value of the second resistance value by inputting the average of the first measured values ​​included in the learning measurement data of a plurality of the measured objects, where the difference between the first measured value and the second measured value is less than the threshold, into the second model; and calculating the difference between the average of the third measured value and the estimated value of the second resistance value as the adjustment value.

28. The method for generating a learned model according to claim 26, wherein, The value of the AC resistance used as a reference is the defined value of the AC resistance of the reference device in the calibration process. The fifth step includes: calculating an estimated value of the second resistance value by inputting the average of the first measured values ​​included in the learning measurement data of a plurality of the measured objects, where the difference between the first measured value and the second measured value is less than the threshold, into the second model; and calculating the difference between the defined value and the estimated value of the second resistance value as the adjustment value.

29. A program for generating a learned model, which causes a computer to perform the steps of the method for generating a learned model according to claim 23 or 26.

Images