JP2017005228A - Quality judgment device, wire bonder device, and quality judgment method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To judge a bonding quality of wire bonding with high accuracy.SOLUTION: A pass/fail judging device 2 for judging whether or not bonding of wire bonding is good in accordance with an output signal at the time of joining wires obtained from a wire bonder device comprises an acquiring unit (31) for acquiring a captured image of the joining portion from the wire bonder device, a classification unit (32) for classifying the captured image into a plurality of categories on the basis of a predetermined index, a selection unit (33) for selecting a reference waveform associated with a category to which the captured image belongs, and a judgment unit (36) for judging whether the junction is good or bad according to a result of comparison between the target waveform of the output signal at the time of joining the joining portion and the reference waveform.SELECTED DRAWING: Figure 5

Description

  The present invention relates to a pass / fail judgment device, a wire bonder device, and a pass / fail judgment method for judging whether or not a wire bond is bonded, and more particularly to a device for determining whether a bond is good or not based on a signal waveform during a bonding operation of the wire bonder.

  In wire bonding, a wire is bonded to an electrode pad on an IC chip and an electrode on a substrate by a wire bonder device, and a bonded state is determined by a quality determination device connected to the wire bonder device. 2. Description of the Related Art Conventionally, a device for determining whether or not a wire bonding is good based on a signal level of an output signal obtained from a wire bonder device is known as a quality determination device (see, for example, Patent Document 1). In the quality determination device described in Patent Document 1, a plurality of windows are set in the output signal from the wire bonder device, and a bonding failure is determined depending on whether or not the signal level in each window falls within an allowable range.

JP 2000-114321 A

  However, the signal level of the output signal from the wire bonder device varies depending on the joint location where the wire is joined. For this reason, in the quality determination apparatus described in Patent Document 1, it is difficult to determine the quality of each joint from the signal level of the output signal obtained from the wire bonder apparatus. Another possible method is to judge whether or not the wire bonding is good from the waveform of the output signal instead of the signal level. In this method, if the difference in waveform shape is clear between when the bonding state is normal and when it is poor, it is possible to judge the quality with high accuracy. There was a possibility to judge.

  The present invention has been made in view of such a point, and an object thereof is to provide a quality determination device, a wire bonder device, and a quality determination method capable of accurately determining the bonding quality of wire bonding.

  The pass / fail judgment device of the present invention is a pass / fail judgment device for judging the pass / fail of wire bonding in accordance with an output signal at the time of joining wires obtained from a wire bonder device, and according to the joint location where the wire is joined. A selection unit that selects a reference waveform, and a determination unit that determines whether or not the joint is good according to a comparison result between a target waveform of an output signal at the time of joining at the joint portion and the reference waveform.

  The pass / fail judgment method of the present invention is a pass / fail judgment method for judging pass / fail of wire bonding in accordance with an output signal at the time of joining wires obtained from a wire bonder device, according to the joint location where the wires are joined. The method includes a step of selecting a reference waveform and a step of determining whether or not the joint is good according to a comparison result between a target waveform of an output signal at the time of joining at the joint portion and the reference waveform.

  The wire bonder apparatus of this invention is equipped with said quality determination apparatus.

  According to the present invention, a reference waveform suitable for determining the quality of a joint at each joint location is selected, and the quality of the joint is determined by comparing the target waveform of the output signal with the reference waveform. Therefore, even if the output signal varies depending on the joint location, the quality of the joint can be determined with high accuracy for each joint location by using the reference waveform in consideration of the variation.

It is a schematic diagram of the quality determination apparatus and wire bonder apparatus which concern on 1st Embodiment. It is a figure which shows an example of the output signal at the time of joining operation | movement which concerns on 1st Embodiment. It is a figure which shows an example of the VCO voltage at the time of the junction operation | movement which concerns on 1st Embodiment. It is a figure which shows an example of the VCO voltage for every junction location which concerns on 1st Embodiment. It is a figure which shows an example of the control block diagram of the arithmetic unit which concerns on 1st Embodiment. It is a figure which shows an example of the classification process by the classification | category part which concerns on 1st Embodiment. It is a figure which shows an example of the selection process by the selection part which concerns on 1st Embodiment. It is a figure which shows an example of the calculation process by the calculation part which concerns on 1st Embodiment. It is a flowchart of the judgment process of the joining quality based on 1st Embodiment. It is a figure which shows an example of the control block diagram of the calculator which concerns on 2nd Embodiment. It is a figure which shows the image of the basic statistic which concerns on 2nd Embodiment. It is a figure which shows an example of the integrated judgment process by the judgment part which concerns on 2nd Embodiment. It is a figure which shows an example of the determination process by the determination part which concerns on 2nd Embodiment. It is a figure which shows the example of a display of the target waveform by the indicator which concerns on 2nd Embodiment. It is a figure which shows the example of a display of the feature-value by the indicator which concerns on 2nd Embodiment. It is a figure which shows the example of a display of the judgment information by the indicator which concerns on 2nd Embodiment. It is a flowchart of the judgment process of the quality of joining which concerns on 2nd Embodiment.

  Hereinafter, the quality determination device according to the first embodiment will be described with reference to the accompanying drawings. FIG. 1 is a schematic diagram of a pass / fail judgment device and a wire bonder device according to the first embodiment. FIG. 2 is a diagram illustrating an example of an output signal during the bonding operation according to the first embodiment. FIG. 3 is a diagram illustrating an example of the VCO voltage during the junction operation according to the first embodiment. FIG. 4 is a diagram illustrating an example of the VCO voltage for each joint location according to the first embodiment. Note that the schematic diagram of the pass / fail judgment device and the wire bonder device shown in FIG. 1 is simplified for the purpose of explaining the present invention, and the configurations that the pass / fail judgment device and the wire bonder device normally have are provided.

  As shown in FIG. 1, the quality determination device 2 according to the present embodiment is configured to determine the quality of wire bonding according to the output signal at the time of wire bonding obtained from the wire bonder device 1. Yes. In the wire bonder apparatus 1, the ultrasonic vibrator 12 is driven by the drive circuit 11, and the wire is ultrasonically bonded to a bonding portion such as a substrate, a lead, and a chip by the bonding tool 13 to which the ultrasonic vibration is transmitted. At the time of ultrasonic bonding, an output signal at the time of bonding operation is output from the drive circuit 11 toward the pass / fail judgment device 2. The drive circuit 11 includes, for example, an oscillation circuit, an amplifier circuit, and a transformer.

  As shown in FIGS. 2A to 2F, the pass / fail judgment device 2 includes, as an output signal from the wire bonder device 1, a displacement amount of the bonding tool 13, a driving current of the ultrasonic transducer 12, and an ultrasonic transducer 12. Drive voltage, current and voltage phase of the ultrasonic transducer 12, input power of the ultrasonic transducer 12, VCO (Voltage Controlled Oscillator) voltage, and the like. The VCO voltage is a voltage that represents a change in the oscillation frequency of the oscillation circuit that follows the change in the resonance frequency of the ultrasonic transducer 12. These various output signals are input to the A / D converter 21 of the pass / fail judgment device 2, A / D converted from analog signals to digital signals, and output to the calculator 22.

  The computing unit 22 determines whether the wire bonding is good or bad according to the waveform shape of the output signal. The output signal has an ideal waveform shape determined by the output of the ultrasonic wave, the application time of the ultrasonic wave, the pressing load of the wire, and the like. For example, as indicated by the solid line in FIG. 3, the waveform shape of a normal VCO voltage rises while drawing a smooth curve at the start of application of ultrasonic waves and decreases to 0 [V] at the end of application. On the other hand, as shown by the broken line in FIG. 3, the waveform shape of the VCO voltage whose bonding state is not good due to dirt on the bonding surface does not rise immediately at the start of ultrasonic application A, but starts to change after a while. .

  Thus, the waveform shape of the VCO voltage differs between when the junction state is normal and when it is defective. Although not described in detail, the waveform shapes of other output signals are similarly different depending on whether the bonding state is normal or defective. The computing unit 22 compares the waveform shape of the output signal with the ideal waveform shape to determine whether or not the joining is good, and the judgment result of the joining quality is displayed on the display unit 23. When the arithmetic unit 22 determines that the bonding is defective, an alarm signal is output from the pass / fail determination device 2 to the wire bonder device 1. The wire bonder apparatus 1 receives the alarm signal from the pass / fail judgment apparatus 2 and stops the joining operation.

  By the way, the waveform shape of the output signal obtained from the wire bonder device 1 is different depending on the joint location. For example, as shown in FIGS. 4A to 4C, the VCO voltage has a variation in the waveform shape according to the joint location of the terminal, the chip, and the substrate. This is because the material, surface roughness, surface shape, and the like of the joint surface at the joint location are different between the terminal, the chip, and the substrate. Therefore, since the waveform shape of the output signal varies at each joint location, it is possible to appropriately determine the joining state even if the waveform shape of the output signal obtained from the wire bonder device 1 is compared with the ideal waveform shape. Is difficult.

  Therefore, in the present embodiment, attention is paid to the fact that the waveform shape of the output signal varies depending on the joint location, and the reference waveform serving as a judgment reference for each joint location is stored in the storage unit 24. Then, a reference waveform suitable for determination of the quality of each joint is selected from the storage unit 24, and the quality of the connection is determined by comparing the target waveform of the output signal with the reference waveform. As a result, even if the waveform shape of the output signal varies depending on the joining location, it is possible to judge the joining quality with high accuracy by selecting a reference waveform that is effective for quality judgment for each joining location. ing.

  Hereinafter, the detailed configuration of the determination process of the bonding quality is described with reference to FIGS. FIG. 5 is a diagram illustrating an example of a control block diagram of the arithmetic unit according to the first embodiment. FIG. 6 is a diagram illustrating an example of classification processing by the classification unit according to the first embodiment. FIG. 7 is a diagram illustrating an example of a selection process by the selection unit according to the first embodiment. FIG. 8 is a diagram illustrating an example of calculation processing performed by the calculation unit according to the first embodiment.

  In the present embodiment, it is assumed that the reference waveform for each joint location is already stored in the storage unit. In the following description, the VCO voltage is exemplified as an output signal, but the same applies to output signals other than the VCO voltage. Moreover, although this Embodiment demonstrates and demonstrates the structure which acquires the picked-up image of a junction location from a wire bonder apparatus and specifies a junction location, it is not limited to this structure. The quality determination device may directly acquire the captured image of the joint portion. In addition, the quality determination device may acquire any information as long as it is information that can identify the joint location. The quality determination device may acquire information from the wire bonder device. May be obtained directly.

  As shown in FIG. 5, the computing unit 22 of the quality determination device 2 includes an acquisition unit 31, a classification unit 32, a selection unit 33, an adjustment unit 34, a calculation unit 35, and a determination unit 36. The acquisition part 31 acquires the captured image of a joining location from the wire bonder apparatus 1 (refer FIG. 1). In this case, in the wire bonder device 1, the joint portion is imaged before the wire is joined by the imaging unit (not shown), and a captured image of the joint portion is output to the pass / fail judgment device 2 using the joining start time as a trigger. Note that the acquisition unit 31 is not limited to a configuration that acquires a captured image directly from the wire bonder device 1, and may acquire a captured image via the A / D converter 21 in the same manner as the output signal.

  The classification unit 32 classifies the captured image into a plurality of categories based on a predetermined index. In this case, since the brightness and the hue are different for each material used for the joint portion, the lightness and the hue are stored in the storage unit 24 as a predetermined index for classifying the captured image of the joint portion. The classification unit 32 extracts brightness and hue from the captured image, and classifies the captured image by comparison with the brightness and hue set as the indices. Note that, for example, an average value, a minimum value, and a maximum value in the image may be used as the brightness as an index. The hue as the index may be indicated by any of RGB, HSV, and color code, for example.

  Specifically, as shown in FIG. 6, the storage unit 24 stores a minimum value, a maximum value, an average value, and a hue for each category as predetermined indexes. The minimum value 100, maximum value 150, average value 120, and white are set for category 1, and the minimum value 50, maximum value 100, average value 80, and red are set for category 2, and category 3 Are set to a minimum value 20, a maximum value 80, an average value 60, and yellow. For example, in the case of a captured image of the minimum value 90, maximum value 130, average value 110, and white of brightness, the captured image is classified into category 1 having the closest brightness and hue.

  In the present embodiment, the category of the captured image of the joint portion is classified using the brightness and the hue. However, the present invention is not limited to this configuration. The predetermined index only needs to be able to classify the captured image of the joint portion into a plurality of categories. Therefore, the captured image may be classified using either one of brightness and hue, or the captured image may be classified using the shape of the joint portion and the surface roughness of the joint portion in the captured image. Since the surface roughness of the joint portion is expressed in black and white in the captured image, it may be indicated by a ratio of black and white. In this way, the captured images are classified into a plurality of categories according to the material of the imaging location.

  The selection unit 33 selects a reference waveform corresponding to the joint location based on the category classified by the classification unit 32. As described above, since the waveform shape of the output signal varies depending on the joint location, the storage unit 24 stores a reference waveform serving as a criterion for pass / fail judgment for each category of the joint location (captured image). For example, as illustrated in FIGS. 7A to 7C, the storage unit 24 stores three types of reference waveforms for categories 1-3 with different waveform rise times. The selection unit 33 selects a reference waveform associated with the category to which the captured image of the joint portion belongs from the storage unit 24. In this way, a reference waveform that is effective for quality determination is selected for each joint location.

  The adjustment unit 34 acquires the output signal from the wire bonder device 1 through the A / D converter 21 and adjusts the level of the target waveform of the output signal to the same level as the reference waveform. As shown in FIG. 3, even if the bonding state is normal, the output value varies due to the difference in resonance frequency caused by the attachment position of the bonding tool 13 (see FIG. 1). For this reason, the adjustment unit 34 calculates a proportionality constant such that the output value of the target waveform becomes a constant value, and normalizes the proportionality constant by multiplying all the target waveforms. As a result, the waveform shapes can be easily compared by superimposing the target waveform of the output signal and the reference waveform.

  The calculation unit 35 compares the target waveform level-adjusted by the adjustment unit 34 with the reference waveform selected by the selection unit 33, and calculates the feature amount of the target waveform from this comparison result. In this case, as shown in FIG. 8A, the difference when the reference waveform indicated by the solid line and the target waveform indicated by the broken line are superimposed is quantified. For example, as shown in FIG. 8B, the deviation between the target waveform and the reference waveform is calculated, and the maximum value of the deviation and the integrated value of the deviation are extracted as the feature amount. As a result, the feature amount of the target waveform of the output signal is calculated in consideration of the variation in the waveform shape due to the joint location. Note that either the maximum deviation value or the integral value of the deviation may be extracted as the feature amount.

  The determination unit 36 determines whether or not the joint is good by using a feature amount obtained from a comparison result between the target waveform and the reference waveform. In this case, for example, ± 3σ of the average value of the output signal accumulated in the past is set as the allowable range. Then, whether or not the joint is good is determined according to whether or not the feature amount of the target waveform is within an allowable range. Further, when the feature amount of the target waveform corresponds to an intermediate value of the joining quality, the operator may be left with the judgment of the joining quality as an intermediate product. Note that the setting method of the allowable range can be changed as appropriate. In addition, when a plurality of feature amounts such as the maximum value of standard deviation and integral value are extracted, an allowable range may be set for each feature amount, or Mahalanobis distance which is one method of multivariate analysis. Pass / fail judgment may be made by threshold judgment.

  When the arithmetic unit 22 determines whether the joining is good or bad, the judgment result of the joining quality is displayed on the display unit 23. When the joining state is determined to be normal, a message indicating that the bonding state is normal is displayed on the display unit 23. When the bonding state is determined to be defective, a defective part is displayed together with the fact that the bonding state is defective. For example, a captured image of a defective part is displayed and the defective part is highlighted. When it is determined that the bonding state is defective, an alarm signal is output from the pass / fail determination device 2 to the wire bonder device 1 (see FIG. 1), and the operation of the wire bonder device 1 is stopped. In this way, the bonding failure during the bonding operation is determined in real time while switching the reference waveform according to the bonding location.

  With reference to FIG. 9, the overall flow of the joint quality judgment process by the quality judgment device will be described. FIG. 9 is a flowchart of the joint quality determination process according to the first embodiment. In addition, the determination process of the quality of joining shown below is only an example to the last, and can be changed suitably. Further, in the description of FIG. 9, description will be made using the reference numerals of FIG. 1 as appropriate.

  As shown in FIG. 9, when the joining operation of the joining portion is started in the wire bonder device 1, a captured image of the joining portion before the joining imaged by the wire bonder device 1 is acquired by the pass / fail judgment device 2 (step S01). . Output signals from the start to the end of the joining operation are input from the wire bonder device 1 to the pass / fail judgment device 2, and are converted from analog signals to digital signals in real time by the A / D converter 21 (step S02). The output signal after A / D conversion is multiplied by a proportionality constant in the arithmetic unit 22, and the signal level of the target waveform of the output signal is adjusted to the same level as the reference waveform (step S03).

  On the other hand, the captured image of the joint portion is classified into a plurality of categories based on a predetermined index by the computing unit 22 (step S04), and a reference waveform associated with the category to which the captured image belongs is selected (step S05). . The computing unit 22 compares the target waveform after the level adjustment with the reference waveform, and calculates a feature amount such as a maximum deviation value or an integrated value of deviation from the comparison result (step S06). Then, whether or not the joint is good is determined according to whether or not the feature amount of the target waveform is within the allowable range (step S07). The determination result of the bonding quality is output to the display 23, and the determination result of the bonding quality is notified to the operator (step S08).

  Moreover, in this Embodiment, although the quality determination apparatus 2 was set as the structure which determines a joining quality using one type of output signal, it is not limited to this structure. As shown in FIG. 2, since the quality determination device 2 obtains six types of output signals from the wire bonder device 1, it is possible to determine the quality of bonding using a plurality of types of output signals. In this case, the selection unit 33 (see FIG. 5) selects a reference waveform corresponding to the junction for each output signal, and the determination unit 36 (see FIG. 5) determines the comparison result between the target waveform and the reference waveform for each output signal. Accordingly, it is determined whether the joining is good or bad. As the plurality of types of output signals, it is preferable to obtain from the wire bonder device 1 the displacement amount of the bonding tool 13 and the VCO voltage with a clear difference in waveform shape according to the determination result of bonding quality.

  For example, when it is determined that the bonding state of both the displacement amount of the bonding tool 13 and the VCO voltage is normal during a bonding operation at a certain bonding point, it is determined that the bonding state of the bonding point is normal as a whole. On the other hand, if it is determined that the bonding state of either the displacement amount of the bonding tool 13 or the VCO voltage is defective during a bonding operation at a certain bonding point, it is determined that the bonding state of the bonding point is defective as a whole. As described above, by using a plurality of comparison results obtained by comparing the target waveform and the reference waveform with respect to a plurality of types of output signals, it is possible to determine the quality of the bonding with higher accuracy.

  As described above, in the pass / fail judgment device 2 according to the first embodiment, the reference waveform is switched according to the joining location of the joining operation by the wire bonder device 1, and based on the comparison result between the target waveform of the output signal and the reference waveform. Whether or not joining is determined in real time. Therefore, even if the output signal varies according to the joint location, the quality of the joint is determined with high accuracy for each joint location using the reference waveform in consideration of the variation.

  In the first embodiment, a configuration is described in which the pass / fail judgment device 2 acquires captured images of joints from the wire bonder device 1, classifies the captured images into a plurality of categories, and selects a reference waveform. However, it is not limited to this configuration. The pass / fail judgment device 2 only needs to acquire information that can identify the joint location from the wire bonder device 1 and select a reference waveform corresponding to the joint location, and the position data that can identify the joint location is used as the wire bonder device 1. May be obtained from In this case, it is not necessary to classify the captured images into a plurality of categories for selection of the reference waveform.

  In the first embodiment, the quality determination device 2 is configured to determine the quality of the joint based on the feature amount of the target waveform of the output signal, but is not limited to this configuration. The quality determination device 2 may be configured to determine the quality of the joint according to the comparison result between the target waveform of the output signal and the reference waveform.

  Incidentally, in the first embodiment described above, a reference waveform must be set in advance in the pass / fail judgment device 2. However, the reference waveform is created from the judgment result of past pass / fail judgment of the output signal, and the reference waveform is not necessarily set at the start of operation of the pass / fail judgment device 2, and the target waveform of the output signal and the reference waveform are obtained. It is not possible to judge pass / fail by comparison. Therefore, in the quality determination device 2 according to the second embodiment, the feature amount can be calculated from the target waveform itself of the output signal without using the reference waveform. Furthermore, by calculating a plurality of types of feature amounts from the target waveform of the output signal, it is possible to determine joint quality in an integrated manner.

  Hereinafter, the quality determination apparatus according to the second embodiment will be described with reference to the accompanying drawings. FIG. 10 is a diagram illustrating an example of a control block diagram of the arithmetic unit according to the second embodiment. FIG. 11 is a diagram illustrating an image of basic statistics according to the second embodiment. FIG. 12 is a diagram illustrating an example of the integrated determination process performed by the determination unit according to the second embodiment. FIG. 13 is a diagram illustrating an example of a determination process by the determination unit according to the second embodiment. FIG. 14 is a diagram illustrating a display example of the target waveform by the display according to the second embodiment. FIG. 15 is a diagram illustrating a display example of feature amounts by the display according to the second embodiment. FIG. 16 is a diagram illustrating a display example of determination information by the display according to the second embodiment. The second embodiment is different from the first embodiment in that the quality of the joint is determined mainly using a plurality of feature amounts. Therefore, in the second embodiment, differences will be mainly described in detail.

  As shown in FIG. 10, the computing unit 22 of the quality determination device 2 includes an acquisition unit 31, a classification unit 32, a selection unit 33, an adjustment unit 34, a calculation unit 35, a determination unit 36, and a determination unit 37. . In addition, about each part of the acquisition part 31, the classification | category part 32, the selection part 33, and the adjustment part 34, since it is the structure same as each part of 1st Embodiment, description is abbreviate | omitted and the characteristic of 2nd Embodiment The calculation unit 35, the determination unit 36, and the determination unit 37, which are parts, will be described in detail. As described above, the calculation unit 35 calculates the feature quantity of the target waveform from the comparison result of the target waveform of the output signal and the reference waveform, and calculates the feature quantity from the target waveform itself of the output signal.

  In this case, the calculation unit 35 calculates the feature amount from the basic statistics and dynamic characteristics of the target waveform of the output signal. The feature quantity based on the basic statistic is obtained by statistical processing on the target waveform of the output signal. For example, the average value, median value (median), minimum value, maximum value, variance, standard deviation, and standard error are calculated as the feature values based on the basic statistics. For example, as shown in FIG. 11, in the case of the target waveform of the VCO voltage, the average value, median value, maximum value, and minimum value are obtained from the voltage value of the target waveform that changes with time.

If the target waveform is x _i , the number of samples of the voltage value is N, and the average value of the samples is m, the variance σ ² is the following equation (1), the standard deviation σ is the following equation (2), and the standard error β is the following: It is calculated | required by Formula (3), respectively.

Next, the feature amount relating to the dynamic characteristics of the target waveform of the output signal is obtained by approximating the target waveform with a temporary delay element and a dead time element. The target waveform is expressed as a transfer function of Equation (4), where K is the gain, T is the time constant, and L is the dead time. Thereby, a gain, a time constant, and a dead time are obtained as a feature amount related to dynamic characteristics. In the case of an output signal other than the VCO voltage, it is approximated by another method that matches the waveform shape.

In addition, the relevance ratio may be obtained as a feature amount related to dynamic characteristics. The precision is the precision between the standard waveform and the target waveform after normalization (after level adjustment). When the reference waveform is y (k), the target waveform is y (k), and the average value of the target waveform is y￣, the precision FIT is obtained by the following equation (5). Here, y ^ (k) indicates that “^” is attached on y, and y￣ indicates that “￣” is attached on y.

  As described above, in the calculation unit 35, in addition to the feature amount based on the comparison result between the target waveform and the reference waveform in the first embodiment, the feature amount is calculated from the basic statistics and dynamic characteristics of the target waveform itself of the output signal. Is done. As a result, even when the reference waveform is not set as when the pass / fail judgment apparatus 2 is started, a plurality of feature amounts can be calculated from the target waveform itself of the output signal. In addition, although the seven types of feature amounts are illustrated as the feature amounts based on the basic statistics and the four types of feature amounts are illustrated as the feature amounts related to the dynamic characteristics, the present invention is not limited to these feature amounts, and the target of the output signal Any feature amount obtained from the waveform itself may be used.

  The determination unit 36 determines whether or not the joint is good or bad by using the feature value obtained from the comparison result of the target waveform and the reference waveform, and also the feature value obtained from the target waveform itself of the output signal, that is, the feature value based on the statistical processing, The quality of the joint is judged using the feature quantity related to the characteristic. The determination unit 36 may determine the quality of joining from one type of feature value per output signal, or may judge the quality of joining from a plurality of types of feature values per output signal. When judging whether or not the joining is good from one type of feature amount, ± 3σ of the average value of the feature amount of the output signal accumulated in the past is set as the allowable range as described above, and whether the feature amount is within the allowable range or not. Whether or not the joining is good is determined. Note that the setting method of the allowable range can be changed as appropriate.

  When determining the joining quality from a plurality of types of feature amounts, the determination unit 36 may determine the joining quality from the ratio of the individual quality results of the plurality of types of feature amounts. For example, a threshold value is set for a plurality of types of feature amounts, and when the feature amount exceeds the threshold value, the joining state is determined to be poor, and when the feature amount does not exceed the threshold value, the joining state is determined to be normal. Then, whether or not joining is individually determined for a plurality of types of feature amounts, and depending on whether or not the ratio of the feature amounts determined to be normal for all types of feature amounts is greater than a predetermined ratio The quality of joining is judged in an integrated manner. For example, with 10 types of feature amounts, when the joint state is normal with seven types of feature amounts, the joint state is determined to be normal as a whole. The threshold setting method can be changed as appropriate.

  Further, the determination unit 36 may integrally determine whether or not the joint is good from the result of clustering a plurality of types of feature quantities as multidimensional data. Output signals accumulated in the past are divided into a cluster to which an output signal with a normal joint state belongs and a cluster to which an output signal with a poor joint state belongs in a multidimensional space having a plurality of types of feature quantities as coordinate axes. When the output signal at the time of joining of the joint location to be judged is arranged in the multidimensional space, the cluster to which the coordinate position of the output signal to be judged belongs is judged. Then, whether or not the joint is good is determined based on the cluster to which the output signal belongs in the multidimensional space.

  For example, as shown in FIG. 12, the output signal accumulated in the past is subjected to cluster analysis, and a multidimensional space having the feature quantity 1-3 as a coordinate axis is clustered into clusters 1-5. Each cluster 1-5 is associated with a judgment result of past joining of past output signals acquired in advance. Then, it is determined that the output signal belongs to the cluster having the shortest distance from the coordinate position of the output signal to be determined in the multidimensional space to the center of gravity of each cluster. When the output signal to be determined belongs to the cluster 5, it is determined that the joining state is normal as a whole from the output signal to be determined. When the output signal to be determined belongs to the cluster 2, it is determined that the joining state is poor as a whole from the output signal to be determined.

  In this cluster analysis of a plurality of types of feature quantities, the joint quality is judged in multiple dimensions using a plurality of feature quantities, so that it is difficult to judge by one-dimensional judgment that judges the joint quality for each feature quantity. In addition, it is possible to determine the quality of the bonding with high accuracy. As a cluster analysis of a plurality of types of feature amounts, for example, a ward method that is a hierarchical clustering method may be used. As described above, in the second embodiment, it is difficult to determine only by comparing the target waveform and the reference waveform by comprehensively determining the joint quality using a plurality of types of feature amounts with respect to the output signal. In addition, it is possible to judge the quality of the joining with high accuracy.

  By the way, when judging the quality of joints using multiple types of feature quantities, unless there is knowledge about the feature quantities in advance, it is not known which feature quantity is effective for judgment of joint quality. Become. On the other hand, when all of a plurality of types of feature amounts are calculated, the load of the feature amount calculation processing and the joining quality determination processing increases. Therefore, in the second embodiment, the determining unit 37 determines in advance the types of feature quantities effective for quality determination from a plurality of types of feature quantities of the output signal accumulated in the past, thereby improving the accuracy of the quality determination. In addition to improving the efficiency of the bonding work.

  The determination unit 37 may determine an effective feature amount for pass / fail judgment using an average value of feature amounts of output signals accumulated in the past. For each of a plurality of types of feature values of the output signal accumulated in the past, an average value of the feature values determined to be normal in the joining state and an average value of the feature values determined to be defective in the joining state Desired. Then, a feature quantity in which the absolute value of the difference between the average value of the normal joining state and the average value of the poor joining state is equal to or greater than a predetermined value is determined as a type of feature quantity effective for pass / fail judgment. It should be noted that the predetermined value serving as a reference when determining the feature amount can be arbitrarily set by the operator with reference to the tendency of each feature amount. In this method, it is also possible to use the median value instead of the average value.

  More specifically, for each type of feature amount, an average value Xa of feature amounts determined to be normal and an average value Xb of feature amounts determined to be poor are determined. When the absolute value of the difference between the average value Xa and the average value Xb is greater than or equal to a predetermined value, there is a high possibility that the feature amount is clearly separated when the joining state is normal and when the joining state is poor. When the absolute value of the difference between the value Xa and the average value Xb is smaller than a predetermined value, it is unlikely that the feature amount is clearly separated when the joining state is normal and when the joining state is poor. Therefore, a feature quantity having an absolute value of a difference greater than or equal to a predetermined value among a plurality of types of feature quantities is used as an effective feature quantity for determining pass / fail.

  Note that the type of feature amount only needs to be an absolute value of the difference greater than or equal to a predetermined value, and feature amounts effective for pass / fail judgment may be determined in descending order of the absolute value of the difference. Considering other matters such as the calculation load, a feature amount effective for pass / fail judgment may be determined. In addition, the types of feature quantities whose absolute value of the difference is equal to or greater than a predetermined value may be displayed on the display 23 in descending order of the absolute value of the difference and arbitrarily determined by the operator. By determining only the types of feature quantities effective for the quality determination obtained from the absolute value of the difference by the determination unit 37, the load of the determination process is reduced along with the accuracy of the quality determination.

  Further, the determination unit 37 may determine a feature amount effective for pass / fail judgment using the frequency distribution of the feature amount of the output signal accumulated in the past. For each of a plurality of types of feature values of the output signal accumulated in the past, a frequency distribution in which the joining state is determined to be normal and a frequency distribution in which the joining state is determined to be defective are obtained. Then, a feature quantity in which the overlapping ratio between the frequency distribution with the normal joining state and the frequency distribution with the poor joining state is equal to or less than a predetermined ratio is determined as the type of feature quantity effective for the pass / fail judgment. It should be noted that the predetermined ratio that serves as a reference when determining the feature amount can be arbitrarily set by the operator with reference to the tendency of each feature amount.

More specifically, as shown in FIG. 13, for each type of feature amount, the frequency distribution of the feature amount determined to be normal and the frequency distribution of the feature amount determined to be defective. Is required. When the frequency distribution overlap ratio is larger than the predetermined ratio, it is unlikely that the feature quantity is clearly separated between the normal state and the poor connection state, and the frequency distribution overlap ratio is low. Is less than a predetermined ratio, there is a high possibility that the feature amount is clearly separated when the joining state is normal and when the joining state is poor. Therefore, a feature quantity whose frequency distribution overlaps a predetermined ratio or less among a plurality of types of feature quantities is used as an effective feature quantity for quality determination. In addition, the ratio of the frequency distribution overlap is obtained by the following equation (6).

  Note that the type of feature amount only needs to be such that the frequency distribution overlap ratio is equal to or less than a predetermined ratio, and the feature amount effective for pass / fail judgment may be determined in descending order of frequency distribution overlap ratio. The feature quantity effective for the pass / fail judgment may be determined in consideration of other matters such as the calculation load of each feature quantity. Further, the types of feature quantities whose frequency distribution overlap ratio is equal to or less than a predetermined ratio may be displayed on the display unit 23 in ascending order of frequency distribution overlap ratio and arbitrarily determined by the operator. By determining only the type of feature quantity effective for the pass / fail judgment obtained from the overlapping ratio of the frequency distribution by the determination unit 37, the load of the judgment process is reduced along with the pass / fail judgment accuracy.

  In addition, the display device 23 according to the second embodiment may display comparison information on the waveform shape of the output signal, comparison information on the feature amount, and determination information on the quality of the joint as support for the quality determination for the operator. As shown in FIG. 14, the display unit 23 displays, as comparison information on the waveform shape of the output signal, the target waveform of the output signal for each joining operation in a manner that can be compared with the judgment result of the joining quality. In this case, the target waveform of the output signal is superimposed and displayed for each joining operation called a batch on the coordinate system with the vertical axis representing the output value and the horizontal axis representing the time. Judgment results are associated. Thereby, the operator can confirm the correlation between the difference in the target waveform of the output signal for each batch and the determination result of the bonding quality.

  For example, when the behavior of the waveform shape of the output signal changes between a batch with a normal joining state and a batch with a poor joining state, the operator can recognize the output signal as an effective output signal. On the other hand, when the behavior of the waveform shape of the output signal does not change between a batch with a normal joining state and a batch with a poor joining state, the operator can recognize the output signal as not being valid for pass / fail judgment. In the example of the figure, the behavior of the waveform shape of the output signal is different between batches 1 and 2 with poor bonding state and batch 3 with normal bonding state, and there is a correlation between the judgment result of bonding quality and the displacement of the target waveform of the output signal. It is recognized by the operator as an output signal.

  In addition, although the target waveform of the several output signal was superimposed and displayed on the indicator 23, the display method is not specifically limited. It is only necessary that the target waveform of the output signal is displayed on the display 23 so as to be comparable. For example, data indicating the behavior of the target waveform may be displayed in a table.

  Further, as shown in FIG. 15, the display unit 23 displays the feature quantity of the target waveform of the output signal associated with the judgment result of the joining quality as comparison information of the feature quantity of the output signal for each batch. The In this case, a plurality of feature values of the target waveform are displayed for each batch in a coordinate system in which the vertical axis represents feature value values and the horizontal axis represents batches, and each batch is associated with a determination result of pass / fail judgment. Yes. Thereby, the operator can confirm the correlation between the feature amount of the target waveform and the determination result of the joining quality. For example, the operator recognizes a feature value that is significantly different between an output signal with a normal joining state and an output signal with a poor joining state as an effective feature amount for necessity determination.

  In the example of the figure, the feature amount 1 is a low value in the batches 1 and 2 in which the joining state is poor, and is a high value in the batch 3 in which the joining state is normal. Since the feature quantity 1 and the joining quality are correlated, the feature quantity 1 is recognized by the operator as an effective feature quantity for necessity determination. The feature amount 2 is a low value in the batch 1 in which the joining state is poor, and is a high value in the batch 2 in which the joining state is poor and the batch 3 in which the joining state is normal. The feature amount 3 is a high value in the batch 1 in which the joining state is poor, and is a low value in the batch 2 in which the joining state is poor and the batch 3 in which the joining state is normal. Since there is no correlation between the feature quantities 2 and 3 and the joining quality, the feature quantities 2 and 3 are recognized by the operator as feature quantities that cannot be used for quality judgment.

  In addition, although the feature-value of the output signal for every batch was displayed with the graph on the indicator 23, the display method is not specifically limited. It is only necessary that the feature value of the target waveform is displayed on the display unit 23 so as to be comparable, and the feature value may be displayed in a table for each batch.

  Further, as shown in FIG. 16, on the assumption that the output signals accumulated in the past in the multidimensional space are clustered, the display 23 displays the output signal to be determined as the determination information of the joint quality. The coordinate position and the distance to the center of gravity of each cluster are displayed. In this case, the distance from the coordinate position of the output signal to be determined for each batch to the center of gravity of each cluster is displayed in a coordinate system with the vertical axis representing distance and the horizontal axis representing batch. Each cluster is divided into a cluster to which an output signal with a normal bonding state belongs and a cluster to which an output signal with a poor bonding state belongs. The operator recognizes whether or not the bonding is good according to the cluster to which the coordinate position of the output signal belongs. The

  In the example of the figure, cluster 1 is a cluster with a normal bonding state, and cluster 2 is a cluster with a poor bonding state. In batches 1, 2, and 4, the distance from the coordinate position of the output signal to the center of gravity of cluster 1 is shorter than the distance to the center of gravity of cluster 2. Thus, the operator recognizes that the output signals of batches 1, 2, and 4 belong to cluster 1 and the joining state is normal. In batches 3 and 5, the distance from the coordinate position of the output signal to the center of gravity of cluster 1 is longer than the distance to the center of gravity of cluster 2. Therefore, the operator recognizes that the output signals of the batches 3 and 5 belong to the cluster 2 and the joining state is defective.

  In addition, although the coordinate position of the output signal for every batch and the distance to the center of gravity of each cluster are displayed in a graph on the display 23, the display method is not particularly limited. The display 23 only needs to be displayed so as to be able to determine whether the joining is good or bad by the distance from the coordinate position of the output signal to the center of gravity of the cluster. For example, the table shows the distance from the coordinate position of the output signal to the center of gravity of each cluster. May be displayed.

  With reference to FIG. 17, the overall flow of the joint quality determination process by the quality determination device will be described. FIG. 17 is a flowchart of a process for determining whether or not the joining is good according to the second embodiment. In addition, the determination process of the quality of joining shown below is only an example to the last, and can be changed suitably. In the description of FIG. 17, description will be made using the reference numerals of FIG. 10 as appropriate.

  As shown in FIG. 17, as the pre-processing before the operation of the wire bonder device 1, the type of feature quantity effective for the pass / fail judgment is determined from the plurality of types of feature quantity of the output signal based on the output signal accumulated in the past ( Step S11). The type of feature quantity may be determined using the average value or the median of the feature quantities of past output signals, or the type of feature quantity may be determined using the frequency distribution of the feature quantities of past output signals. Good. When the wire bonder device 1 is operated, output signals from the start to the end of the bonding operation are input from the wire bonder device 1 to the pass / fail judgment device 2, and are converted from analog signals to digital signals in real time by the A / D converter 21. (Step S12).

  When the reference waveform is set in the pass / fail judgment apparatus 2 (Yes in step S13), the feature amount of the target waveform is calculated from the comparison result between the target waveform of the output signal and the reference waveform (step S14). Although detailed description is omitted here, each processing such as acquisition of the captured image of the joint portion, level adjustment of the target waveform, classification of the captured image of the joint portion, selection of the reference waveform, and the like is performed (see FIG. 9). Steps S01, S03-S05). Thereafter, the target waveform after the level adjustment is compared with the reference waveform, and a feature amount such as a maximum deviation value or an integral value of the deviation is calculated from these comparison results.

  When the reference waveform is not set in the pass / fail judgment device 2 (No in step S13), the feature amount is calculated from the target waveform itself of the output signal (step S15). A feature amount based on a basic statistic may be calculated by statistical processing on the target waveform of the output signal, or a feature amount related to dynamic characteristics may be calculated by approximating the target waveform with a temporary delay element and a dead time element. When the feature amount of the target waveform is obtained in this way, it is determined whether or not the joining portion is good based on the feature amount (step S16). In this case, the joining quality may be determined from one type of feature amount, or the joining quality may be determined from a plurality of types of feature amounts.

  When the quality of the joint is determined from one type of feature amount, the quality of the joint is determined according to whether or not the feature amount of the target waveform is within an allowable range. When joint quality is determined from multiple types of feature quantities, joint quality may be determined from the ratio of individual quality results of multiple types of feature quantities, or multiple types of feature quantities may be used as multidimensional data. The quality of joining may be determined in an integrated manner from the result of clustering. Then, the determination result of bonding quality is output to the display 23, and the determination result of bonding quality is notified to the operator (step S17). Note that the display unit 23 may display comparison information on the waveform shape of the output signal, comparison information on the feature amount, and determination information on whether or not the joint is good as support for the operator in determining whether the product is good or bad.

  It should be noted that step S11 can be omitted when a feature amount to be used for quality determination is determined in advance. Further, when the reference waveform is set in step S13, the feature amount is calculated from the comparison result between the target waveform of the output signal and the reference waveform. However, the present invention is not limited to this configuration. The feature quantity may be calculated from the comparison result between the target waveform of the output signal and the reference waveform, and the feature quantity may be calculated from the target waveform itself of the output signal. In step S16, the joining quality may be determined using a feature amount obtained from the comparison result of the target waveform of the output signal and the reference waveform and a feature amount obtained from the target waveform itself of the output signal as a plurality of feature amounts. .

  As described above, in the quality determination device 2 according to the second embodiment, it is determined only by comparing the target waveform and the reference waveform by comprehensively determining the quality of the joint using a plurality of types of feature amounts. It is possible to determine whether or not the joining is good with high accuracy even for difficult things. Even when the reference waveform is not set as in the case of starting the operation of the quality determination device 2, it is possible to determine the quality of the joining.

  In addition, this invention is not limited to the said embodiment, It can change and implement variously. In the above-described embodiment, the size, shape, and the like illustrated in the accompanying drawings are not limited to this, and can be appropriately changed within a range in which the effect of the present invention is exhibited. In addition, various modifications can be made without departing from the scope of the object of the present invention.

  For example, in the first and second embodiments, the configuration in which the wire bonder device 1 and the pass / fail judgment device 2 are separate is described, but the configuration is not limited thereto. The wire bonder device 1 may include a quality determination device 2. Thereby, in the wire bonder apparatus 1, the quality of joining at the time of joining of a wire can be judged with high precision.

  In the first and second embodiments, the configuration in which the captured image of the joint portion is acquired by the imaging unit of the wire bonder device 1 is illustrated, but the configuration is not limited thereto. An imaging unit may be provided in the quality determination device 2 so that the quality determination device 2 acquires a captured image of the joint portion.

  In the second embodiment, the feature amount is calculated from the target waveform itself when the reference waveform is not set at the start of operation of the pass / fail judgment device 2, but the present invention is not limited to this configuration. Regardless of the presence or absence of the reference waveform at the start of operation of the quality determination device 2, the feature amount may be calculated from the target waveform itself.

The feature points in the various embodiments described above are organized below.
The pass / fail judgment device described in the above embodiment is a pass / fail judgment device for judging pass / fail of wire bonding in accordance with an output signal at the time of joining wires obtained from a wire bonder device, where the wire is joined. A selection unit that selects a reference waveform corresponding to the reference point, and a determination unit that determines whether the joining is good or not according to a comparison result between a target waveform of an output signal at the time of joining at the joining portion and the reference waveform. .

  In addition, the quality determination method described in the above embodiment is a quality determination method for determining the quality of wire bonding according to an output signal at the time of wire bonding obtained from a wire bonder device, and the wire is bonded. A step of selecting a reference waveform corresponding to the joining portion, and a step of judging whether the joining is good or not according to a comparison result between a target waveform of an output signal at the time of joining the joining portion and the reference waveform. .

  According to these configurations, the reference waveform suitable for determining the quality of the joint at each joint location is selected, and the quality of the joint is determined by comparing the target waveform of the output signal with the reference waveform. Therefore, even if the output signal varies depending on the joint location, the quality of the joint can be determined with high accuracy for each joint location by using the reference waveform in consideration of the variation. Further, even if the waveform shape varies between the case where the joining state is normal and the case where it is poor, the joining quality is not erroneously determined by comparing the target waveform with the reference waveform.

  The quality determination device according to the embodiment includes an acquisition unit that acquires a captured image of the joint portion, and a classification unit that classifies the captured image into a plurality of categories based on a predetermined index, The unit selects a reference waveform associated with a category to which the captured image belongs. According to this configuration, the captured image of the joint portion is classified based on a predetermined index, and a reference waveform suitable for determining whether or not the joint portion is good can be selected based on the classification of the captured image.

  In the quality determination apparatus described in the above embodiment, the classification unit classifies the captured images into a plurality of categories using brightness and / or hue as the predetermined index. According to this configuration, the surface state of the joint portion is classified into a plurality of categories based on the brightness and / or hue of the imaging portion, and a reference waveform suitable for determining whether the joint is good or bad in the surface state of the joint portion can be selected.

  In addition, the quality determination device described in the above embodiment has obtained a plurality of types of output signals from the wire bonder device, and the selection unit selects a reference waveform corresponding to the joint location for each of the output signals, The determination unit determines whether or not the bonding is good according to a comparison result between the target waveform and the reference waveform for each output signal. According to this configuration, by using a plurality of comparison results obtained by comparing the target waveform and the reference waveform with respect to each of a plurality of types of output signals, it is possible to determine the quality of the joint with higher accuracy.

  Further, the quality determination device described in the above embodiment obtains a displacement amount of a bonding tool and a VCO (Voltage Controlled Oscillator) voltage from the wire bonder device as the plurality of types of output signals. According to this configuration, it is possible to determine the bonding quality with high accuracy from the displacement amount of the bonding tool and the target waveform of the VCO voltage where the difference in waveform shape according to the determination result of the bonding quality is clear.

  In addition, the pass / fail judgment device according to the embodiment includes an adjustment unit that adjusts a signal level of the target waveform to the same level as the reference waveform, and a comparison result between the target waveform after the level adjustment and the reference waveform. A calculation unit that calculates a feature amount of the target waveform, and the determination unit determines the quality of the joint using the feature amount of the target waveform obtained from the comparison result. According to this configuration, since the target waveform and the reference waveform are adjusted to the same signal level, the difference in waveform shape between the target waveform and the reference waveform can be easily compared. It can be used for judgment.

  Further, in the quality determination device described in the above embodiment, the calculation unit calculates a deviation between the target waveform and the reference waveform as a feature amount of the target waveform from a comparison result between the target waveform after level adjustment and the reference waveform. And / or an integrated value of the deviation is calculated. According to this configuration, the maximum value of the deviation between the target waveform and the reference waveform and / or the integrated value of the deviation can be used as a feature value for determining whether or not the joint is good.

  In the quality determination device described in the above embodiment, the calculation unit calculates a feature amount of the target waveform from statistical processing on the target waveform, and the determination unit uses the feature amount based on the statistical processing. Judging the quality of joining. According to this configuration, by performing statistical processing on the target waveform of the output signal, the feature amount can be calculated from the target waveform itself.

  In the quality determination device according to the embodiment, the calculation unit calculates a feature amount of the target waveform from the dynamic characteristic of the target waveform, and the determination unit calculates a feature amount related to the dynamic characteristic. Use it to judge whether the joint is good or bad. According to this configuration, by analyzing the dynamic characteristics of the target waveform of the output signal, the feature amount can be calculated from the target waveform itself.

  Further, in the quality determination device described in the above embodiment, the determination unit determines the quality of joining based on a plurality of types of feature amounts for one output signal. According to this configuration, it is possible to judge joint quality with high accuracy even if it is difficult to judge only by comparing the target waveform with the reference waveform by judging joint quality using multiple types of feature values in an integrated manner. Can do.

  Further, in the quality determination device described in the above embodiment, the determination unit determines whether or not the plurality of types of feature quantities are individually joined for the output signal, and the joining state is normal for all types of feature quantities. It is determined whether the joining is good or not according to whether or not the ratio of the determined feature amount is larger than a predetermined ratio. According to this configuration, it is possible to determine joint quality based on the judgment results of the quality judgment of each of the plurality of types of feature amounts.

  Further, in the pass / fail judgment apparatus described in the above embodiment, the output signal accumulated in the past is a multidimensional space having the plurality of types of feature quantities as coordinate axes, and the cluster and the joining state to which the output signal having a normal joining state belongs. Are divided into clusters to which the defective output signal belongs, and the determination unit determines whether or not the joint is good based on the cluster to which the coordinate position of the output signal to be determined belongs in the multidimensional space. According to this configuration, since the quality of joining is determined in a multidimensional manner using a plurality of feature quantities, even those that are difficult to judge by a one-dimensional judgment that judges the quality of joining for each feature quantity are joined with high accuracy. Pass / fail can be judged.

  In addition, the quality determination device described in the embodiment determines a type of feature quantity effective for quality determination from the plurality of types of feature quantities based on a plurality of types of feature quantities of output signals accumulated in the past. A determination unit is provided. According to this configuration, since only the feature quantity effective for the pass / fail judgment is used from among a plurality of types of feature quantities, the accuracy of the pass / fail judgment can be improved and the load of the judgment process can be reduced.

  Further, in the quality determination device described in the above embodiment, the determination unit determines an average value of the feature amounts determined to be normal for each of a plurality of types of feature amounts of the output signal accumulated in the past. Alternatively, the median and the average value or median value of the features that are determined to be poor in the joining state are obtained, and the difference between the average value or median that is normal in the joining state and the average value or median that is poor in the joining state. A feature amount having an absolute value of at least a predetermined value is determined as a type of feature amount effective for pass / fail judgment. According to this configuration, the larger the absolute value of the difference between the average value or the median value where the bonding state is normal and the bonding state is poor, and the average value or the median value where the bonding state is bad, the larger the difference between the normal state and the poor bonding state. Features are clearly separated. Therefore, a feature quantity whose absolute value of the difference is equal to or greater than a predetermined value can be used as a feature quantity effective for pass / fail judgment.

  Further, in the quality determination apparatus described in the above embodiment, the determination unit includes a frequency distribution in which a joining state is determined to be normal for each of a plurality of types of feature amounts of the output signal accumulated in the past, and a joining A frequency distribution in which the state is determined to be defective is obtained, and a feature quantity that is less than or equal to a predetermined ratio of the overlap ratio between the frequency distribution in which the bonding state is normal and the frequency distribution in which the bonding state is poor Determine as the type of quantity. According to this configuration, the lower the ratio of the overlap between the frequency distribution in which the bonding state is normal and the frequency distribution in which the bonding state is poor, the more clearly the feature amount is separated when the bonding state is normal and when the bonding state is poor. The Therefore, a feature quantity with an overlap ratio equal to or less than a predetermined ratio can be used as a feature quantity effective for pass / fail judgment.

  Moreover, the quality determination apparatus described in the above-described embodiment includes a display that displays the target waveform of the output signal for each bonding operation in a comparable manner in association with the determination result of the bonding quality. According to this configuration, the operator can confirm the correlation between the difference in the target waveform of the output signal and the determination result of the joining quality. Therefore, the operator can recognize an output signal having a correlation between the determination result of the bonding quality and the displacement of the target waveform of the output signal as an output signal effective for the determination of the bonding quality.

  Moreover, in the quality determination device described in the above embodiment, the display unit displays the feature amount of the target waveform of the output signal associated with the determination result of the bonding quality in a comparable manner. According to this configuration, the operator can confirm the correlation between the feature amount of the target waveform and the determination result of the joining quality. Therefore, it is possible to cause the operator to recognize a feature quantity having a correlation between the judgment result of the joint quality and the feature quantity of the target waveform as a feature quantity effective for judgment of the joint quality.

  Further, in the pass / fail judgment apparatus described in the above embodiment, the output signal accumulated in the past is a multidimensional space having the plurality of types of feature quantities as coordinate axes, and the cluster and the joining state to which the output signal having a normal joining state belongs. Are divided into clusters to which the defective output signal belongs, and the display unit displays the coordinate position of the output signal to be determined in the multidimensional space and the distance to the center of gravity of each cluster. According to this configuration, it is possible to make the operator confirm a cluster having a short distance from the coordinate position of the output signal to the center of gravity of each cluster. Therefore, it is possible to make the operator recognize the determination result of the pass / fail determination based on the cluster having the close coordinate position of the output signal.

  The wire bonder device described in the embodiment includes the above-described pass / fail judgment device. According to this configuration, it is possible to determine with high accuracy whether or not the wire is bonded in the wire bonder device.

  As described above, the present invention has an effect that it is possible to determine the bonding quality of wire bonding with high accuracy, and in particular, the quality determination device that determines the bonding quality based on the signal waveform during the bonding operation of the wire bonder device, This is useful for a wire bonder device and a quality determination method.

DESCRIPTION OF SYMBOLS 1 Wire bonder apparatus 2 Pass / fail judgment apparatus 13 Bonding tool 22 Calculator 23 Display device 24 Storage part 31 Acquisition part 32 Classification part 33 Selection part 34 Adjustment part 35 Calculation part 36 Judgment part 37 Determination part

Claims (20)

According to the output signal at the time of bonding of the wire obtained from the wire bonder device, it is a pass / fail judgment device for judging the bonding quality of wire bonding,
A selection unit that selects a reference waveform according to a joint location where the wire is joined;
A pass / fail judgment device comprising: a judgment unit that judges a joint pass / fail according to a comparison result between a target waveform of an output signal at the time of joining at the joint location and the reference waveform. An acquisition unit for acquiring a captured image of the joint portion;
A classification unit that classifies the captured image into a plurality of categories based on a predetermined index;
The pass / fail judgment apparatus according to claim 1, wherein the selection unit selects a reference waveform associated with a category to which the captured image belongs.   The quality determination apparatus according to claim 2, wherein the classification unit classifies the captured images into a plurality of categories using brightness and / or hue as the predetermined index. A plurality of types of output signals are obtained from the wire bonder device,
The selection unit selects a reference waveform corresponding to the joint location for each output signal,
The quality determination device according to any one of claims 1 to 3, wherein the determination unit determines the quality of the bonding according to a comparison result between the target waveform and a reference waveform for each output signal.   5. The pass / fail judgment device according to claim 4, wherein a displacement amount of a bonding tool and a voltage controlled oscillator (VCO) voltage are obtained from the wire bonder device as the plurality of types of output signals. An adjustment unit for adjusting the signal level of the target waveform to the same level as the reference waveform;
A calculation unit that calculates a feature amount of the target waveform from the comparison result of the target waveform after level adjustment and the reference waveform;
The quality determination device according to claim 1, wherein the determination unit determines the quality of joining using a feature amount of a target waveform obtained from the comparison result.   The calculation unit calculates a maximum value of a deviation between the target waveform and the reference waveform and / or an integrated value of the deviation as a feature amount of the target waveform from a comparison result between the target waveform after the level adjustment and the reference waveform. The pass / fail judgment device according to claim 6. The calculation unit calculates a feature amount of the target waveform from statistical processing for the target waveform,
The quality determination device according to claim 6 or 7, wherein the determination unit determines the quality of joining using a feature amount based on the statistical processing. The calculation unit calculates a feature amount of the target waveform from dynamic characteristics of the target waveform,
The quality determination apparatus according to any one of claims 6 to 8, wherein the determination unit determines the quality of the joint using a feature amount related to the dynamic characteristic.   The quality determination device according to any one of claims 6 to 9, wherein the determination unit determines the quality of joining based on a plurality of types of feature values for one output signal.   The determination unit individually determines whether or not the plurality of types of feature values are connected to the output signal, and the ratio of the feature values determined to be normal for all types of feature values is greater than a predetermined ratio. The quality determination apparatus according to claim 10, wherein the quality determination is made according to whether the size is large or not. The output signals accumulated in the past are divided into a cluster to which the output signal having a normal joining state belongs and a cluster to which an output signal having a poor joining state belongs in a multidimensional space having the plurality of types of feature amounts as coordinate axes. And
The quality determination device according to claim 10, wherein the determination unit determines the quality of joining based on a cluster to which a coordinate position of an output signal to be determined belongs in the multidimensional space.   11. The apparatus according to claim 10, further comprising a determining unit that determines a type of feature quantity effective for quality determination from the plurality of types of feature quantities based on a plurality of types of feature quantities of output signals accumulated in the past. The quality determination device according to claim 12.   The determination unit determines, for each of a plurality of types of feature amounts of the output signal accumulated in the past, an average value or a median value of the feature amounts determined to be normal, and a determination that the connection state is poor. The average value or median value of the obtained feature values is obtained, and the feature value in which the absolute value of the difference between the average value or the median value where the joining state is normal and the mean value or the median value where the joining state is poor is equal to or greater than a predetermined value. The quality determination apparatus according to claim 13, wherein the quality determination apparatus determines the type of feature quantity effective for the quality determination.   For each of the plurality of types of feature values of the output signal accumulated in the past, the determination unit includes a frequency distribution determined to be normal and a frequency distribution determined to be poor. The feature amount that is obtained and the ratio of the overlap between the frequency distribution with a normal joining state and the frequency distribution with a poor joining state is determined to be a predetermined ratio or less is determined as the type of feature amount effective for the pass / fail judgment. Item 14. The quality determination device according to Item 13.   The pass / fail judgment device according to any one of claims 1 to 15, further comprising a display that displays a target waveform of an output signal for each joining operation so as to be able to be compared in association with a judgment result of the pass / fail.   The quality determination apparatus according to claim 16, wherein the display unit displays a feature quantity of a target waveform of an output signal associated with the determination result of the bonding quality so as to be comparable. The output signals accumulated in the past are divided into a cluster to which the output signal having a normal joining state belongs and a cluster to which an output signal having a poor joining state belongs in a multidimensional space having the plurality of types of feature amounts as coordinate axes. And
The pass / fail judgment apparatus according to claim 16 or 17, wherein the display unit displays a coordinate position of an output signal to be judged in the multidimensional space and a distance to the center of gravity of each cluster.   A wire bonder device comprising the pass / fail judgment device according to any one of claims 1 to 18. According to the output signal at the time of bonding of the wire obtained from the wire bonder device, it is a quality determination method for determining the quality of wire bonding bonding,
Selecting a reference waveform according to the joint location where the wire is joined;
And a step of determining whether or not the joint is good according to a comparison result between a target waveform of an output signal at the time of joining the joint portion and the reference waveform.

Images