JP2007329329A - Device, method and program for extracting defective factor, and recording medium storing program for extracting defective factor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for extracting a defective factor, which judges the defective factor of a product in a manufacturing device from a wide view point. <P>SOLUTION: When a change in feature quantity data is outputted to judge the defective factor of a substrate of an entire electronic component packaging device, data varying in time about one feature quantity are outputted (S21). Next, about one feature quantity, a change in data on one substrate is outputted (S22). Then, about one feature quantity, a change in data on a plurality of substrates is outputted (S23). <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a defect factor extraction device, a defect factor extraction method, a defect factor extraction program, and a recording medium on which a defect factor extraction program is recorded, and in particular, an electronic component that mounts a component on a substrate through a plurality of steps The present invention relates to a defect factor extracting device, a defect factor extracting method, a defect factor extracting program, and a recording medium on which a defect factor extracting program is recorded, which can extract feature quantities for quality improvement in each process of the mounting apparatus.

For example, Japanese Laid-Open Patent Publication No. 11-298200 (Patent Document 1), Japanese Laid-Open Patent Publication No. 10-284900 (Patent Document 2), and the like are methods for extracting the cause of a substrate defect for improving the quality of a conventional surface mounting apparatus. And it is disclosed by Unexamined-Japanese-Patent No. 7-212098 (patent document 3) etc. According to Patent Document 1, quality control is performed for each process of the surface mounting device, and measurement data in each process when a defect measure is taken is stored in a database in advance. It is disclosed that a corresponding treatment is performed by referring to a database using measurement data at the time. According to Patent Document 2, a failure cause estimation method for estimating a failure cause is disclosed based on a result of counting suction / mounting errors in a mounting process and a change record of equipment mounting conditions. According to Patent Document 3, an evaluation table is created on the basis of a measurement value database that manages measurement values, a set of mounting positions corresponding to the evaluation range is obtained, a common part of the set of mounting positions is obtained, and between evaluation ranges The correlation frequency is obtained.
JP 11-298200 A (summary) JP 10-284900 A (summary) JP-A-7-212098 (summary)

  The estimation of the cause of the defect of the substrate in the conventional electronic component mounting apparatus has been performed as described above. In Patent Documents 1 and 2, since factor extraction is not automated, it takes time to extract the factor and only monitors whether there is an abnormality in the measurement data due to temporal changes. It may not be possible to identify. Further, according to Patent Document 3, although the factor is automatically estimated, there is a problem in that the failure factor cannot be extracted in the entire electronic component mounting apparatus because only the inspection in the mounting process is performed. . Therefore, in the prior art, there is a problem that when a certain defect occurs, it is not known where to start, or how important the defect is from the whole.

  The present invention has been made paying attention to the above-described problems, and a failure factor extraction device that can determine a factor of a failure of a product such as a substrate in a manufacturing device such as an electronic component mounting device from a wide viewpoint. An object of the present invention is to provide a failure factor extraction method, a failure factor extraction program, and a recording medium storing the failure factor extraction program.

  According to the present invention, in a manufacturing apparatus that manufactures a product through a plurality of steps, a defect factor extraction device that extracts a defect factor of a product is obtained from an inspection machine provided in at least one of a plurality of steps. One of the input means for inputting feature quantity data related to the cause of the above, the temporal change detection means for detecting temporal changes in the feature quantity data inputted by the input means, and the production of the feature quantity data inputted by the input means Positional change detecting means for detecting a change according to a position on a product, and a positional change on a plurality of products for detecting a change according to a position on a plurality of products of feature amount data inputted by the input means At least two of the detection means, and at least any two of the temporal change detection means, the positional change detection means, and the on-product positional change detection means. Or And an output means for outputting the feature data.

  At least any two of the temporal change, the change according to the position on one product, and the change according to the position on a plurality of products in the feature amount data related to the cause of the defect Since the feature amount data is output, the user can know not only the data viewed from one point of view, but also the change tendency of the feature amount data from a plurality of points of view.

  As a result, it is possible to provide a failure factor extraction device that can determine a factor of a product defect in a manufacturing device from a wide viewpoint.

  Preferably, the manufacturing apparatus is an electronic component mounting apparatus, and the product is a substrate.

  More preferably, the output means displays the feature amount data as a graph.

  More preferably, it includes normalizing means for normalizing the feature data, and the output means outputs the data normalized by the normalizing means.

  More preferably, an extraction means for extracting a factor of failure based on the normalized data is included.

  The normalizing means normalizes the feature amount data with a plurality of different indices, and the extracting means extracts defect factors to be dealt with on the basis of the normalized data based on the plurality of different indices.

The feature amount data may include device information used when a component is mounted on the board.
According to another aspect of the present invention, a defect factor extraction method for extracting a defect factor of a product of a manufacturing apparatus that manufactures a product through a plurality of steps is performed by an inspection machine provided in each of the plurality of steps. A step of inputting feature value data related to a cause of a defect in the product; a step of detecting temporal changes in the input feature value data; and a position of the input feature value data on one product. A step of detecting a change in accordance with the step, a step of detecting a change in the input feature value data in accordance with a position on a plurality of products, and a detected temporal change, a position on one product. Detecting a change in any two of the change in accordance with the position and the change in accordance with the position on the plurality of products, and a step of outputting the change in the detected feature amount data.

  Preferably, the manufacturing apparatus is an electronic component mounting apparatus, and the product is a substrate.

  Preferably, the method further includes a step of normalizing the feature amount data and a step of extracting a factor of failure based on the normalized feature amount data.

  According to another aspect of the present invention, a failure factor extraction program causes a computer to execute the above-described failure factor extraction method.

  The defect factor extraction program may be stored in a computer-readable recording medium.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Here, the defect of the substrate in the electronic component mounting apparatus will be described as an example of the defect of the manufacturing apparatus and the product to which the defect factor extracting apparatus according to the present invention is applied. FIG. 1 is a diagram illustrating a configuration around an electronic component mounting apparatus. Referring to FIG. 1, electronic component mounting apparatus 10 includes a printing process, a mounting process, and a reflow process arranged from the upstream side to the downstream side of the substrate on which the electronic component is mounted. Each process is connected by a conveyor, a robot, and other transfer devices. Each process is provided with an apparatus for performing the process.

  In the printing process, a printing machine 11 for printing lands on a substrate and a post-printing inspection machine 12 for performing an inspection after printing are provided. In the mounting process, a mounter 13 for mounting a component on a substrate and a post-mount inspection machine 14 for performing an inspection after mounting are provided. In the reflow process, a reflow furnace 15 for soldering the terminal of the component to the land and an after-reflow inspection machine 16 for performing an inspection after soldering are provided.

  The printing machine 11, the post-printing inspection machine 12, the mounter 13, the post-mounting inspection machine 14, the reflow furnace 15, and the post-reflow inspection machine 16 are connected to the defect factor extraction device 20 via the network 17.

  From the post-printing inspection machine 12, inspection data in the printing machine 11 is received. From the mounter 13, apparatus information relating to devices used by the mounter 13 such as a used head, a used nozzle and a used feeder is received from the post-mounting inspection machine 14. The data indicating the inspection result in the mounter 13 and the data indicating the inspection result after reflow are input from the post-reflow inspection machine 16 to the defect factor extracting device 20, respectively.

  Here, there are various types of inspection data measured by each of the inspection machines 12, 14, and 16, but these inspection data, a combination of a position state on the substrate, a substrate ID for specifying the substrate, and the like are characteristic. It becomes quantity data.

  Accordingly, the feature amount data measured by the post-printing inspection machine 12 includes a combination of print area and print deviation data with substrate position information. The feature amount data measured by the post-mounting inspection machine 14 includes a combination of device information used by the mounter 13 and positional deviation information of mounted components.

  Next, the configuration of the failure factor extraction device 20 will be described. FIG. 2 is a block diagram illustrating a configuration of the failure factor extraction device 20. Referring to FIG. 2, failure factor extraction device 20 is basically the same as a personal computer (hereinafter referred to as “personal computer”). The failure factor extraction device 20 includes a CPU 21 that controls the entire device, and a ROM, RAM (not shown), a display device 23, a hard disk 24, a keyboard 25, a mouse 26, and a network connected to the CPU 21 via an interface 22. Including the connecting device 27 and the like. By this network connection device 27, the failure factor extraction device 20 is connected to each of the printing machines 11, the post-printing inspection machine 12, the mounter 13, the post-mounting inspection machine 14, the reflow furnace 15, and the post-reflow inspection machine 16. To input feature data. Therefore, it functions as the network connection device 27 input means.

Next, a basic concept in the defect factor extraction device 20 in this embodiment will be described. In this embodiment, the failure factor extraction device 20 analyzes data collected from the inspection machines 12, 14, and 16 from the following viewpoints.
(1) A change in a certain feature value is seen on a plurality of substrates.
(2) A change in a certain feature amount is seen over the entire board.
(3) A change in one characteristic amount is observed on a plurality of substrates.

  In this way, by analyzing the feature amount data that can be a cause of substrate failure from a wide viewpoint, it is possible to grasp the cause of failure from a wider viewpoint.

  Note that it is preferable to analyze the feature data from all three viewpoints described here, but the feature data may be analyzed from at least two of these viewpoints.

  Next, a specific processing procedure performed by the CPU 21 of the failure factor extraction device 20 will be described. FIG. 3 is a flowchart showing the specific processing contents for knowing the change of the feature amount data as the basis when collecting the feature amount data from the above three viewpoints.

  Referring to FIG. 3, first, component selection calculated for each feature amount is automatically or manually performed (step S <b> 11, hereinafter “step” is omitted). Specifically, a plurality of processed component data are stored in advance in the hard disk 24, and the contents are displayed on the display device 23. The user selects a desired part from among them using the mouse 26 or the like.

  Next, the inspection data related to the feature data is read from each of the inspection machines 12, 14, 16 stored in the hard disk 24 (S12). Next, the change state of the selected feature amount data is calculated and graphed, and output by, for example, displaying on the display device 23 (S13). Therefore, the display device functions as output means.

  The user sees the change in the feature amount data, and determines whether or not it is necessary to extract a defect factor (S14). When it is determined that it is necessary, a failure factor is extracted (S15). Specifically, after the display device 23 displays the change in the feature amount data, the display device 23 displays an inquiry as to whether or not to extract the defect factor. In response, the user answers. If the defect factor extraction is not performed in S14, the process is terminated as it is.

  Here, the reason why the change output of the feature amount data and the extraction of the defect factor are separated is to first allow the user to determine the degree of change of the feature amount data.

  It should be noted that a predetermined threshold value may be set to automatically determine whether or not a defect factor needs to be extracted.

  Next, the processing contents of the change data output shown in S13 of FIG. 3 will be described with reference to FIG. Referring to FIG. 4, in this process, first, temporal change data for one feature amount is output (S21). The output of this feature amount corresponds to the above (1). Next, a change in data on one substrate is output for one feature amount (S22). This corresponds to the above (2). Next, a change in data on a plurality of substrates for one feature amount is output (S23). This corresponds to the above (3).

  Therefore, the CPU 21 functions as a temporal change detection unit, a positional change detection unit, and a multiple substrate position change detection unit.

  Next, the points of interest in the temporal data change, the data change on one substrate, and the data change on a plurality of substrates described in S21 to S23 will be described. When viewing the output of the temporal change data, it is checked whether the variation of the measurement value of a certain feature amount has occurred suddenly or has changed gradually, and the degree of variation of the measurement value is also checked.

  When looking at changes in data on a single substrate, the degree of size by location, the degree of variation in measured values, the degree of difference when compared to similar items, and the comparison by location with the same Check the degree of difference when you shoot, the degree of outbreak depending on the location, etc.

  When viewing changes in data on a plurality of substrates, check the degree of difference in measured values and the degree of size according to location when comparing the same with each other.

  Next, the content of the defect factor extraction process shown in S15 of FIG. 3 will be described. FIG. 5 is a flowchart showing the operation performed by the CPU 21 in this case. Referring to FIG. 5, in the failure factor extraction processing content, first, the selected feature data is normalized by a statistical method (S31). Next, the presence / absence of abnormal data outside a predetermined range is determined from the normalized data (S32). If the data is outside the predetermined range (YES in S32), the feature amount is displayed based on the data and the factor is extracted (S33). This factor extraction is performed with reference to a database prepared in advance. Therefore, the CPU 21 functions as a normalizing unit and an extracting unit.

  Here, from the normalized data, whether the data is within a predetermined range or out of the range is performed by using a standard statistical method, for example, a standard deviation σ. That is, if the detected feature amount data is within 3σ, it can be determined as an allowable value. If the detected feature amount data is outside the predetermined range, the display device 23 outputs an output indicating that the degree of abnormality is high, assuming that it is outside the predetermined range. And extract factors.

  As described above, since the cause of failure is extracted using not a measured value but a value obtained by normalizing it, it can be easily determined whether or not it is an important failure factor. Further, in factor extraction, the degree of contribution as a defect factor can be determined by obtaining normalized values for a plurality of defect factors and comparing them with each other. Therefore, it is possible to easily determine from which part it is more efficient to deal with when a substrate defect occurs.

  Next, a specific example of the temporal change data shown in S21 of FIG. 4 will be described taking as an example the change in the print area detected by the print inspection machine 12 in the printing process. FIG. 6 is a graph showing the change in the print area displayed by the display device 23. In FIG. 6, the horizontal axis represents the number of substrates, and the vertical axis represents the printing area. By looking at such a graph, the user determines whether or not it is necessary to extract a defect factor.

  In the example shown in FIG. 6, since the printing area suddenly increases at a certain point, the user wants to know how much this change is abrupt in time. At this time, the user causes the failure factor extraction apparatus 10 to extract the failure factor. The defect factor extraction apparatus 10 performs the process described with reference to FIG. 5 to diagnose temporal suddenness of “printing area”. As described above, the failure factor extraction apparatus 10 determines whether or not the detected data is temporally abrupt using a statistical method. Specifically, the failure factor extraction device 10 calculates the temporal suddenness according to the following equation (1).

Suddenness = | Average value−Measured value | / σ (1)
Where σ is the standard deviation.

  As described above, for example, if the degree of suddenness exceeds 3, it is determined that there is an abnormality, and a failure factor is extracted.

  FIG. 7 shows a graph when the data is gradually changing as another example of the temporal change data. Also in FIG. 7, the horizontal axis represents the number of substrates, and the vertical axis represents the printing area. Here, the degree of gradual change is detected by obtaining the slope of the approximate line by the least square method. This value is also normalized based on other data, and if it is 3 times or more of the standard deviation as described above, it is determined to be abnormal, and that fact is displayed on the display device 23 and Perform extraction.

  FIG. 8 is a diagram illustrating the degree of data variation as an example of temporal change data. Referring to FIG. 8, here again, the degree of variation is calculated based on the standard deviation σ of the data. Similarly to the above, when the standard deviation is three times or more of the standard deviation, it is determined that there is an abnormality, and defect factors are extracted. Do.

  Next, the degree of change depending on the position on the substrate shown in S22 of FIG. 4 will be described. In FIG. 9, the horizontal axis is the position of the substrate (X coordinate or Y coordinate of the substrate), and the vertical axis is the same as the above, for example, the printing area. Referring to FIG. 9, here, the printing area gradually increases according to the position on the substrate. This increase amount is also obtained by, for example, obtaining the slope of the approximate line by the least square method.

  Here, the position on the substrate will be described. FIG. 10 is a diagram illustrating the substrate 30. In FIG. 10, for example, the position on the substrate corresponds to the position indicated by the arrow X in the figure as seen from the direction of the thick arrow A. That is, in FIG. 10, the direction indicated by arrow X corresponds to the X axis in FIG. 9 (the smaller value in FIG. 10 is smaller). This position is not limited to the X axis, and may be detected in the direction indicated by Y as viewed from the direction of the thick arrow B in FIG.

  FIG. 11 is a diagram showing variations in measured values as the degree of change depending on the position on the substrate. Again, the degree of variation is determined using the standard deviation σ.

  Next, a description will be given of a case where the print areas as feature amount data at different positions on the substrate are compared with probability density. FIG. 12 shows that when the substrate is divided into three equal parts in the vertical direction (Y direction in FIG. 10), the portion at the upper third of the substrate (the side indicated as small in FIG. 10) and the lower portion of the substrate It is a graph which shows the probability density showing distribution of the printing area in each position when it divides | segments into the part which exists in 1/3 (the side described as large value in FIG. 10).

  Referring to FIG. 12, here, the respective data are normally distributed, the printing area of the lower part of the substrate is generally larger than that of the upper part of the substrate, and the respective center values are separated by a difference d. Recognize. When such data is output, it is determined how much the difference d is statistically abnormal, and defective factors causing such a phenomenon are extracted.

  Next, change data on a plurality of substrates shown in S23 of FIG. 4 will be described. FIG. 13 is an example when the print area data shown in FIG. 12 is obtained for a plurality of substrates. Referring to FIG. 13, similar results are obtained even when a plurality of substrates are obtained. In this case, the same processing as in FIG. 12 is performed.

  Next, FIG. 14 shows a result when the inclination representing the change of the printing area depending on the substrate position in one substrate shown in FIG. 9 is viewed for a plurality of substrates. Referring to FIG. 14, the position of the substrate is the same as in FIG. The Y axis indicates the slope obtained in FIG. Even in this case, the slope of the approximate straight line is obtained by the method of least squares.

  Next, an example will be described in which it is determined whether or not a sudden change in position has occurred using feature value change data when the positions on the substrate are different. This example corresponds to the example of FIG. 6 in which it is determined whether or not a sudden change occurs in time. An example of output data in this case is shown in FIG. Again, whether or not there is a sudden outage is calculated by the above equation (1).

  Next, a case where the printing area for each position on the substrate shown in FIG. 9 is output for a plurality of substrates will be described. An output example in this case is shown in FIG. Referring to FIG. 16, the X axis is the same as in FIG. 9, and the Y axis is a plot of the print area at each position for each substrate. Referring to FIG. 16, when the feature amount data is arranged in this way, it is possible to grasp the change in the print area according to the position on the plurality of substrates.

  Next, another example of normalizing each feature amount shown in S31 of FIG. 5 will be described. In the above embodiment, normalization is performed using a statistical method. In this embodiment, the expert determines a normalized numerical value based on the knowledge.

  FIG. 17 is a diagram for explaining a procedure for an expert to normalize data using, for example, a print area as feature amount data. Referring to FIG. 17, here, for example, the degree of variation in the printing area is shown on the X axis, and the normalized value is shown on the Y axis. When the variation is distributed from a1 to a3, the expert determines b1 and b3 as normalized values corresponding to a1 and a3. Based on this determination, if the variation measurement data is a2, for example, b2 is obtained as a normalized value. That is, if the graph of FIG. 17 defined by the expert is used, the corresponding value after normalization can be calculated from the degree of variation.

  Next, a method for normalization using the degree of divergence will be described. The divergence degree is an index for normalization that is determined to be the same if it is 0 and is completely different if it is 1. FIG. 18A is a diagram showing the variation of the printing area divided into an upper part of the substrate (this position is the same as described in FIG. 10 and indicated by a solid line in the figure) and a lower part (indicated by a dotted line). FIG. 18B shows a graph normalized based on the data shown in FIG.

  Referring to FIG. 18A, it is assumed that the degree of deviation of the feature amount between the upper part and the lower part of the substrate is 0.188. When this is normalized by MF (membership function), the degree of divergence becomes 0.384. When normalization is performed in this way, as shown in FIG. 18B, a portion different from the normal portion is clarified, and if a portion having the largest divergence is found, this is the feature quantity most related to the defect. It can be seen that it is.

  Furthermore, when a plurality of data normalized with different indexes such as standard deviation and divergence are compared in this way, it is possible to extract a failure factor to be dealt with from a wider range of feature amount data.

  Next, the factor extraction method shown in S33 of FIG. 5 will be described. As described above, the failure factor extraction device 10 displays a value indicating a change in the feature amount data together with the graph based on the measurement data obtained from the inspection machines 12, 14, and 16. Normalize the value to determine the improvement target. Specifically, the person estimates the degree of abnormality and considers where to start. On the other hand, if a threshold value is set in advance and the threshold value is exceeded, data is automatically normalized, the largest failure factor is extracted for a plurality of indicators, and where to start is displayed. You may do it.

  A specific factor extraction example is performed using a separately prepared database. Here, some examples will be described. First, there are the following cases as examples based on temporal changes. For example, the measured value (printing area) gradually increases because the printing pressure is too high. It is determined that the sudden increase in the printing area is caused by sudden failure such as dust.

  As an example when viewing the data for the entire substrate, for example, the degree of divergence of the measured value (for example, print area) is calculated by the upper or lower portion of the substrate. If there is an inclination in the approximate straight line of the measured value from the upper part to the lower part of the substrate, an unexpected imbalance of the printing pressure, for example, an inclination caused by warpage of the substrate, dust, or the like can be considered.

  When looking at the data of the entire plurality of substrates, continuous defects can be confirmed. For example, the degree of deviation of the measured value (printing area) is calculated by the upper part or the lower part of the substrate. The inclination of the approximate straight line of the measured value from the upper part to the lower part of the substrate is considered to be caused by imbalance (for example, the lower mold inclination) within the lot of printing pressure.

  Difference in deviation amount with the same component type (chip resistance etc. 0603 size etc.) with different constants. In the case where a difference in deviation amount between different nozzles is continuously generated, it is possible to consider factors such as poor feeder settings and nozzle wear.

  In the above embodiment, the feature area data has been described by taking the printing area in the printing process as an example. However, the same applies to the positional deviation between the device information and the mounted component in the mounting process and the data in the reflow process. Is performed.

  Moreover, in the said embodiment, although the case where the inspection machine provided in each of three processes was used was demonstrated, you may use the inspection machine provided in not only this but at least 1 process.

  In the above embodiment, feature data is normalized by standard deviation and divergence to detect “abnormality” that is a failure factor. However, the present invention is not limited to this, and a reference value for each feature is used. , And “high degree of abnormality” may be presented based on the feature amount that has changed significantly compared to the reference value.

  In the above embodiment, the statistical method used to extract the feature amount has been described using the standard deviation and the degree of divergence. However, the present invention is not limited to this, and the average difference, the correlation ratio, As the deviation amount, an arbitrary amount such as a ratio of the distance from the origin of the deviation degree of deviation may be used.

  In the above-described embodiment, the case where the user selects a desired part for determining the failure factor has been described. However, the present invention is not limited to this, and the CPU may be related from the normalized data. Specific parts that are likely to be displayed may be displayed as analysis targets.

  Further, in the above-described embodiment, the case where the defect factor extraction device is applied to the extraction of the cause of the substrate defect in the electronic component mounting apparatus has been described. However, the present invention is not limited to this, and is applicable to the factor extraction of the manufacturing defect in an arbitrary manufacturing apparatus. May be.

  In the above embodiment, the case where the failure factor extraction device is the dedicated device has been described. However, the present invention is not limited to this, and the failure factor extraction device is a general-purpose personal computer, and all the above operations are performed as a failure factor extraction program. The personal computer may be operated by the program. In this case, the defect factor extraction program may be provided on a recording medium such as an optical disk or a hard disk, or may be downloaded from a server on the network via a network.

  As mentioned above, although embodiment of this invention was described with reference to drawings, this invention is not limited to the thing of embodiment shown in figure. Various modifications and variations can be made to the illustrated embodiment within the same range or equivalent range as the present invention.

It is a figure which shows the whole structure of an electronic component mounting apparatus. It is a block diagram which shows the structure of a defect factor extraction apparatus. It is a flowchart which shows the operation | movement which CPU of a defect factor extraction apparatus performs. It is a flowchart which shows the operation | movement performed in the output process of change data. It is a flowchart which shows the operation | movement performed in a defect factor extraction process. It is a graph which shows the sudden time change of a printing area. It is a graph which shows the time change of a printing area. It is a graph which shows the time change of a printing area. It is a graph which shows the positional change of the printing area on one board | substrate. It is a figure which shows the position on a board | substrate. It is a graph which shows the positional change of the printing area on one board | substrate. It is a figure which shows the state which compared the printing area in the different position on a board | substrate with probability density. It is a figure which shows the state which compared the printing area in the different position on a board | substrate with the probability density using the data of several board | substrates. It is a graph which shows the positional change of the printing area on a some board | substrate. It is a graph which shows the sudden positional change of a printing area. It is a graph at the time of outputting the printing area for every position on a substrate about a plurality of substrates. It is a figure for demonstrating normalization of data when a printing area is used as feature-value data. It is a figure for demonstrating the method normalized using a deviation degree.

Explanation of symbols

10 electronic component mounting apparatus, 11 printing machine, 12 post-printing inspection machine, 13 mounter, 14 post-mounting inspection machine, 15 reflow furnace, 16 post-reflow inspection machine, 17 network, 20 defect factor extraction device, 21 CPU, 22 interface, 23 display device, 24 hard disk, 25 keyboard, 26 mouse, 27 network communication device.

Claims (12)

In a manufacturing apparatus that manufactures a product through a plurality of processes, a defect factor extracting device that extracts a defect factor of the product,
Input means for inputting feature value data related to the cause of the defect from an inspection machine provided in at least one of the plurality of steps;
Temporal change detecting means for detecting temporal changes in the feature amount data input by the input means, and positional detection for detecting changes in the feature amount data input by the input means according to the position on one product. At least two detection means among a change detection means, and a position change detection means on the plurality of products for detecting a change of the feature amount data input by the input means according to the position on the plurality of products; and With
Including: temporal change detection means; positional change detection means; and output means for outputting feature data from at least any two of the plurality of product positional change detection means. , Defect factor extraction device. The defect factor extraction device according to claim 1, wherein the manufacturing apparatus is an electronic component mounting apparatus, and the product is a substrate. The defect factor extraction device according to claim 1, wherein the output unit displays the feature amount data as a graph. 4. The failure factor extraction device according to claim 1, further comprising normalization means for normalizing the feature amount data, wherein the output means outputs data normalized by the normalization means. 5. The failure factor extraction device according to claim 4, further comprising an extraction unit that extracts the cause of the failure based on data normalized by the normalization unit. The normalization means normalizes the feature amount data with a plurality of different indexes,
6. The failure factor extraction device according to claim 5, wherein the extraction unit extracts failure factors to be dealt with on the basis of normalized data based on the plurality of different indexes. The defect factor extraction device according to claim 1, wherein the feature amount data includes device information used when a component is mounted on the substrate. A defect factor extraction method for extracting a defect factor of a product of a manufacturing apparatus that manufactures a product through a plurality of steps,
Inputting feature value data related to the cause of product defects from an inspection machine provided in each of the plurality of processes;
Detecting a temporal change in the input feature data;
Detecting a change in input feature data according to a position on one product;
Detecting a change in input feature data according to positions on a plurality of products;
Detecting any two of the detected temporal change, a change according to a position on one product, and a change according to a position on a plurality of products;
Outputting a change in the detected feature value data. The defect factor extraction method according to claim 8, wherein the manufacturing apparatus is an electronic component mounting apparatus, and the product is a substrate. Furthermore, normalizing the feature data,
The failure factor extraction method according to claim 8, further comprising a step of extracting a failure factor based on the normalized feature amount data. A failure factor extraction program for causing a computer to execute the failure factor extraction method according to claim 7. A computer-readable recording medium storing the defect factor extraction program according to claim 11.

Images