JP2009115462A - Inspection method of solder electrode by x-ray tomographic image, and substrate inspecting device using this method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve processing efficiency as much as possible, while securing inspection accuracy of a solder electrode. <P>SOLUTION: When inspecting the solder electrode 10 of a BGA (ball grid array) component 11 by an inspection device having a function executing both image reconstitution by tomosynthesis and image reconstitution by X-ray CT, in the case of executing image reconstitution by tomosynthesis, in each inspection region which is a photographing object at each time, it is discriminated by using CAD data of a substrate 1 whether duplication is generated in a projection range of a component 12, on the rear side of the solder electrode 10 in a plurality of fluoroscopic images used for reconstitution. In this case, the image reconstitution by X-ray CT is determined to be executed in an inspection region where it is discriminated that duplication is generated and the image reconstitution by tomosynthesis is determined to be executed in an inspection region where it is discriminated that duplication is not generated, and the determination results are registered. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention uses a tomographic image of X-rays to indicate the state of a solder electrode that connects an electronic component mounted on this board and a board-side electrode (called a land or pad) using a double-sided mounting board as an inspection target. The present invention relates to an inspection method and a substrate inspection apparatus using the method.

  This type of inspection is intended for solder electrodes that are difficult to inspect, such as ball-shaped solder electrode assemblies (generally called BGA (Ball Grid Aray)) provided on the back of the component body. It is suitable for use in determining whether or not there is a hollow portion called “void”. As a specific method for reconstructing a tomographic image of a solder electrode, a method using X-ray CT (see Patent Document 1) and a method using the principle of tomosynthesis (see Patent Document 2) are known. Yes.

Japanese Patent No. 3694833 JP-T-2004-515762

  In image reconstruction by tomosynthesis, an X-ray source and an X-ray detector are arranged so as to sandwich an object, a plurality of projection processes are executed while changing the relationship between the two, and the generated image is digitally processed. Average. In this case, the projection processing is based on one of a plurality of cross-sections (hereinafter referred to as “target cross-sections”) to be generated as tomographic images, and an arbitrary point in the reference cross-section is always detected by X-rays. Since the positions of the X-ray source and the X-ray detector in the hourly projection are adjusted so that they are projected onto the same coordinates of the instrument, the points in the reference section are emphasized by averaging the images, The points of the cross-section are projected at various positions and are blurred in the averaged image. Also, for each target cross section other than the reference cross section, the averaging process is performed by correcting the image generated for the reference cross section so that the coordinates of each point in the cross section are the same. It is possible to obtain an image in which each point is emphasized and other cross-sectional points are blurred. Therefore, although any noise remains in any target cross section, an image in which the target cross section is clarified can be obtained.

  On the other hand, in image reconstruction by X-ray CT, for each target section, the axis line connecting the X-ray source and X-ray detector arranged opposite each other is orthogonal to the normal of the target section. The object is deployed, and the projection is repeated while changing the orientation of the X-ray source and the X-ray detector with respect to the object in units of minute angles. And the calculation which calculates | requires the X-ray absorption coefficient of each point of a target cross section is performed using many produced | generated X-ray fluoroscopic images.

  However, in the method disclosed in Patent Document 1, in consideration of the point of generating a cross section orthogonal to the thickness direction for a thin substrate, the point of generating an enlarged cross section of a partial region, etc. The processing is slightly different from the general method. Briefly, the X-ray source and the X-ray detector are arranged to face each other in a direction obliquely intersecting with the rotation axis while rotating the substrate by a rotary table, and each generated X-ray fluoroscope is photographed. The image is converted into an image that has been transmitted from a direction perpendicular to the rotation axis, and an operation for calculating an absorption coefficient is executed using the converted image.

  Comparing the processing by tomosynthesis and the processing by X-ray CT, X-ray CT is overwhelmingly superior in terms of the accuracy of the generated tomographic image. However, in X-ray CT, a large number of projection processes are required and the calculation is complicated, so that it takes a considerable time to reconstruct a tomographic image. On the other hand, according to tomosynthesis, the number of times of projection processing is much smaller than that of X-ray CT and the calculation is simple, so that the processing can be completed in a short time.

  From this point of view, establishments that manufacture and inspect a large number of substrates desire to perform inspection by tomosynthesis tomography as much as possible in order to increase inspection efficiency. However, it is assumed that the accuracy of inspection can be ensured.

In image reconstruction using the tomosynthesis method, if there is a composition with a high X-ray absorption rate in a relatively large area other than the target cross section, the image quality of the tomographic image may be degraded, and the accuracy of the inspection may not be ensured. There is. In a plurality of images used for reconstruction of tomographic images, each point of the above-described component is projected at a different position for each image. This is because when a point having a high absorption rate is projected, these projection results are emphasized by the averaging process and appear as a clear image.
If noise due to a projected image outside the inspection target as described above is generated within the range in which the solder electrode to be inspected is projected, it is difficult to ensure the accuracy of the inspection. Therefore, in such a case, it is necessary to give up applying tomosynthesis and generate a tomographic image by X-ray CT.

  As a specific example of the noise as described above, the inventor focused attention on noise generated by a component mounted on the back side of the solder electrode to be inspected when the solder electrode on the double-sided mounting board is inspected. Even if there is a component on the back side of the solder electrode, there is no possibility that significant noise will occur if the projection range of this component does not overlap between images. However, depending on the size and mounting position of the component on the back side, the range in which the component is projected on the X-rays from each direction may overlap, and the projected image of this overlapping portion is superimposed, causing significant noise. It can happen.

  In consideration of the above situation, the present invention is based on X-ray CT image reconstruction and tomosynthesis depending on whether or not noise caused by components on the back side occurs when inspecting solder electrodes on a double-sided mounting board. The first problem is to improve the processing efficiency as much as possible while ensuring the accuracy of the inspection by automatically switching and executing the image reconstruction by the technique.

  Furthermore, a second object of the present invention is to enable automatic execution of processing for determining an image reconstruction method suitable for each region to be examined using data indicating the arrangement state of components.

  In the inspection method according to the present invention, an X-ray source and an X-ray detector arranged at different height positions, a substrate support part for horizontally supporting the substrate to be inspected therebetween, and X for the substrate to be inspected A control processing unit that adjusts the relative positions of the X-ray detector and the X-ray detector to generate an X-ray fluoroscopic image of the substrate, and a combination of a pair of orientations that are in conflict with each other as the X-ray source orientation relative to the target cross section A plurality of fluoroscopic images are generated by adjusting the relative positions so that one or more sets of images are generated, and each tomographic image is integrated based on the projection range of the target cross section to generate a tomographic image of the target cross section. Image reconstruction by tomosynthesis and X-ray CT that generates tomographic images by generating more X-ray fluoroscopic images than tomosynthesis and obtaining the X-ray absorption coefficient distribution of the target cross section using these images. And an image reconstruction, to prepare a test apparatus that can perform switching.

  Further, in this method, data indicating the configuration of the double-sided mounting substrate to be inspected is input to the control processing unit prior to the inspection, and the control processing unit after the data input uses the above two types for each predetermined imaging target region. The image to be reconstructed is determined using input data, and the determination result is registered, and then the tomographic image is obtained by the registered image reconstruction for each imaging target region of the substrate to be inspected. Is automatically generated and the inspection is executed.

  In the process of determining the content of image reconstruction for one imaging target area performed by the control processing unit described above, assuming that image reconstruction by tomosynthesis is performed on the imaging target area, a plurality of fluoroscopic images are obtained. Photographed when the X-ray source orientation with respect to the target section at the time of image generation is combined with each other and the X-ray transmission image related to the combination is aligned with respect to the projection range of the target section for each combination. Whether or not there is an overlapping part in the projection range of the back part of the part to be inspected in the target area is photographed by the size of the back part obtained from the input data and the X-ray for the target cross section when each image is generated The determination is made based on the relationship with the size of the range passing through the substrate surface on the back side of the target area. Then, when a determination result is obtained that the projection range of the back part is overlapped for the predetermined number of combinations, it is determined that the image reconstruction by the X-ray CT is executed, and it is determined that the projection range overlap occurs. When the number of combinations is less than the predetermined number, it is determined that image reconstruction by tomosynthesis is executed.

  In a preferred aspect of the above method, in the process of determining the content of image reconstruction for one imaging target region, hourly fluoroscopic imaging by hypothetical tomosynthesis is performed for each of the cases where the X-ray source orientation with respect to the target cross section is in a contradictory relationship. Based on the X-ray irradiation angle and the height of the target cross section from each direction irradiated to a predetermined position of the target cross section for each combination and the direction in which the reciprocity of the X-ray source direction occurs for each combination. Then, the size of the X-ray passing range passing through the substrate surface on the back side of the imaging target region is obtained, and the obtained size is compared with the size of the back side component obtained from the input data. Then, in a predetermined number of combinations, if the part becomes larger than the X-ray passage range, it is determined that the projection range of the back part overlaps, and the image reconstruction by the X-ray CT is executed. decide.

  According to the above method, prior to the inspection, registration is performed so that image reconstruction by X-ray CT is performed on an imaging target region where there is a high possibility that noise caused by components on the back side becomes significant. Settings can be made so that image reconstruction by tomosynthesis is executed for other imaging target regions. Therefore, since an appropriate image generation process is automatically selected and executed for each imaging target region based on the set data at the time of inspection, a tomographic image that does not interfere with the inspection is obtained for any region to be inspected. The accuracy of inspection can be ensured. In addition, since the image reconstruction by tomosynthesis is performed on a region to be inspected where there is no possibility that noise due to the components on the back side becomes noticeable, the efficiency of the inspection can be improved.

  In this method, data indicating the configuration of the double-sided mounting board to be inspected (board thickness, position and size of each component, etc.) is input to the control processing unit to determine whether inspection by tomosynthesis is possible. It is possible to easily determine without performing test shooting.

  In the above method, each solder electrode may be individually photographed. However, when a minute object such as a solder electrode constituting a BGA is to be inspected, an area including a plurality of solder electrodes is defined as 1. One imaging target area may be used. In this case, the imaging target area can be determined in units of parts.

  A substrate inspection apparatus to which the above method is applied includes an X-ray source and an X-ray detector arranged at different height positions, a substrate support unit that horizontally supports the substrate to be inspected therebetween, and an inspection A plurality of fluoroscopic imaging was performed by adjusting the relative positions of the X-ray source and the X-ray detector with respect to the board so that the electronic component mounted on the target board via the solder electrode was shot. A control processing unit that generates a tomographic image of a solder electrode from a plurality of fluoroscopic images, and an inspection unit that inspects the state of the solder electrode using the tomographic image generated by the control processing unit.

  In the control processing unit, a plurality of fluoroscopic images are generated by adjusting the relative positions so that one or more pairs of azimuths having a contradictory relationship as the azimuths of the X-ray source with respect to the target cross section are generated, A first processing unit that performs image reconstruction by tomosynthesis that generates a tomographic image of a target cross-section by integrating each X-ray fluoroscopic image based on the projection range of the target cross-section, and generates more fluoroscopic images than tomosynthesis Then, a second processing unit that executes image reconstruction by X-ray CT that generates a tomographic image by obtaining an X-ray absorption coefficient distribution of a target cross section using these images, and a predetermined imaging target region , Which determines whether image reconstruction by the first or second processing unit is to be executed, registration means for registering information indicating the determination result, and for each imaging target region based on the registered information It includes control means for the processing unit determined for the region to perform the image reconstruction processing.

  The above registration means accepts input of data indicating the configuration of the double-sided mounting board when the solder electrode of the double-sided mounting board is to be inspected, and determines the contents of the above-described image reconstruction for each imaging target area Execute the process.

  According to the present invention, when an inspection using a tomographic image by X-ray is performed on a solder electrode of an electronic component mounted on a double-sided mounting substrate, the inspection is performed due to noise caused by the component mounted on the back side surface. For a region to be imaged for which it is difficult to ensure the accuracy of X-ray CT, switching to image generation processing by X-ray CT is automatically performed to generate a high-accuracy tomographic image. Since the tomographic image is easily and quickly generated by tomosynthesis for the imaging target region that does not occur, the inspection efficiency can be improved while ensuring the accuracy of the inspection. In addition, since the image reconstruction method to be applied to each imaging target region can be automatically determined using the board configuration data, the setting can be completed in a short time without imposing a burden on the user. .

FIG. 1 shows a schematic configuration of a substrate inspection apparatus using X-rays to which the present invention is applied.
This substrate inspection apparatus generates a tomographic image by X-rays for a portion where appearance inspection is difficult, and performs inspection using the generated image. A substrate support table 2 that supports a substrate 1 to be inspected. , An X-ray tube 3, a flat panel detector 4 (hereinafter abbreviated as “FPD4”), a CCD camera 5, a displacement sensor 6, and a controller 20 shown in FIG.

  In the figure, 10 indicates a solder electrode as an example of a specific inspection object, and 11 indicates a component connected to the substrate 1 by the solder electrode 10 (hereinafter referred to as “BGA component 11”). In FIG. 1, the solder electrodes 10 are exaggerated for the sake of convenience, but the actual solder electrodes 10 are very small and many are provided on the back surface of the BGA component 11.

  Moreover, the board | substrate 1 of this Example is a double-sided mounting board | substrate, Comprising: Various components are mounted in each surface and the BGA component 11 may be mounted in both surfaces. Therefore, although a tomographic image for the solder electrode 10 of the BGA component 11 on the lower surface side may be generated, in this embodiment, the description will be given only for the case where the solder electrode 10 on the upper surface side is an inspection target. In the following, the surface on which the BGA component 11 to be inspected is mounted is referred to as the front surface, and the other is referred to as the back surface.

  The substrate support table 2 supports the substrate 1 in a horizontal posture. Note that. This support is performed on the edge portion on which the component on the back surface of the substrate is not mounted, and the component on the back surface is accommodated in an accommodation space (not shown) of the substrate support table 2. The X-ray tube 3 is of a type that emits a conical X-ray beam (cone beam), and is disposed above the substrate support table 2. The FPD 4 receives X-rays that have passed through the substrate 1 and generates a two-dimensional X-ray fluoroscopic image. Usually, the FPD 4 is held in a horizontal posture so that the detection surface 4A faces directly above by a support holder (not shown). It is supported. However, when X-ray CT imaging is performed, the FPD 4 is supported in an inclined posture so that the detection surface 4A faces the X-ray tube 3 as indicated by a one-dot chain line in the figure.

  The substrate support table 2 and the FPD 4 are respectively in the X-axis direction (the left-right direction in FIG. 1) and Y-axis direction (the direction orthogonal to the paper surface in FIG. 1) by XY stages 8 and 9 shown in FIG. Is supported so as to be movable. On the other hand, the X-ray tube 3 is fixedly disposed at a predetermined height position with the X-ray emission surface facing downward.

  The CCD camera 5 and the displacement sensor 6 are arranged at predetermined positions below the FPD 4 with their light receiving surfaces facing the substrate 1. The CCD camera 5 images a predetermined range on the lower surface of the substrate 1. The displacement sensor 6 is an optical sensor in which a laser diode or a photodiode is incorporated. The distance from the light receiving surface to the back surface of the substrate 1 (if there is a component at the measurement target position, the distance to the component is Measure). The image generated by the CCD camera 5 is used for the positioning process of the substrate 1, and the detection signal from the displacement sensor 6 is used for adjusting the height position of the target cross section.

  The CCD camera 5 and the displacement sensor 6 are arranged in a housing part (not shown) whose upper surface is opened. The opening of the housing is normally closed by the shutter 7 and is opened only when imaging by the camera 5 or distance measurement by the displacement sensor 6 is executed. Thereby, it is possible to prevent the CCD camera 5 and the displacement sensor 6 from being affected by X-rays during X-ray transmission imaging.

  In this embodiment, the FPD 4 is used as the X-ray detector. However, the present invention is not limited to this, and an image intensifier (II), a CCD, or the like may be used.

FIG. 2 is a block diagram showing the overall configuration of the substrate inspection apparatus.
The controller 20 of this board inspection apparatus is manufactured using a general-purpose personal computer, and includes a control unit 21 including a CPU, a memory 22, an arithmetic processing unit 23, a monitor 24, an operation unit 25, and the like. In addition to the X-ray tube 3 and the FPD 4 shown in FIG. 1, the controller 20 includes a CCD camera 5, an XY stage 8 for moving the substrate 1, an XY stage 9 for moving the FPD 4, and a PLC (programmable logic controller). ) 26 and the like are connected.

  The PLC 26 drives the CCD camera 5, the displacement sensor 6, and the shutter 7 in response to a command from the control unit 21 to perform imaging and distance measurement. Further, the detection signal from the displacement sensor 6 is taken in, converted into distance data, and output to the control unit 21. On the other hand, the image signal from the CCD camera 5 is input to the control unit 21 via an image interface (not shown) in the controller 20 without passing through the PLC 26.

  The memory 22 is a large-capacity memory such as a hard disk, and stores a program related to control and inspection, CAD data of a substrate to be inspected, and various setting data necessary for execution of the inspection. An image database for storing the tomographic image used for the examination is also provided in the memory 22.

  The arithmetic processing unit 23 is for executing calculations related to the reconstruction of tomographic images by X-ray CT, and is configured as a substrate on which an arithmetic circuit including a high-performance CPU is mounted.

  In the above configuration, the substrate 1 to be inspected is carried into the substrate support table 2 from an upstream transport mechanism (not shown). The control unit 21 uses the XY stage 8 to align the substrate 1 to the initial position, and then starts fluoroscopic imaging for a plurality of predetermined inspection areas on the substrate 1.

  In the process of aligning the substrate 1 to the initial position, a reference mark formed at one corner of the substrate 1 is imaged by the CCD camera 5 and X, X, and so on so that the reference mark in the image matches a predetermined reference position. The position of the substrate 1 in each Y-axis direction is adjusted.

  Prior to fluoroscopic imaging for each inspection region, the substrate 1 is once positioned so that the center of the target inspection region is aligned with the optical axis L of the X-ray tube 3, and then the substrate 1 is placed at a position corresponding to fluoroscopic imaging. Move. Even in this alignment before fluoroscopic imaging, an image of the back surface of the substrate 1 is generated by the CCD camera 5, and the substrate 1 in each of the X and Y axes is aligned so that this image matches the image of the model registered in advance. Adjust the position.

  Further, the control unit 21 corrects the height of the predetermined reference cross section based on the distance data measured by the displacement sensor 6 before starting the movement for fluoroscopic imaging after the above alignment is completed. In accordance with this correction, control parameters such as the amount of movement of the XY stages 8 and 9 during fluoroscopic imaging are adjusted. This is because the height of the cross section to be inspected is determined according to the thickness of the substrate 1, the solder electrode 10, etc., but this cross section may be displaced from the original height position due to the warp of the substrate or the like.

  Note that the height of each target cross section other than the reference cross section is represented by the distance from the reference cross section, so that the height of all target cross sections can be adjusted by correcting the height of the reference cross section. it can. Thus, by starting fluoroscopic imaging after performing the alignment of the substrate 1 and the height adjustment of the reference cross section for each inspection region, a target tomographic image can be generated with high accuracy.

  The inspection area referred to in this embodiment is a virtual three-dimensional area including the solder electrode 10 (an area sandwiched between the component 11 and the substrate 1), and is an object of fluoroscopic imaging. Since the actual solder electrode 10 is very small, a plurality of solder electrodes 10 are included in one inspection region. In this embodiment, the solder electrodes 10 are exaggerated. Assume that the inspection area is set individually.

In the substrate inspection apparatus of this embodiment, it is possible to perform image reconstruction by X-ray CT and image reconstruction by tomosynthesis, and execute either one of the image reconstruction processes for each inspection region. I have to.
In any method, a plurality of target cross sections having different heights are set, and the X-ray tube 3 with respect to the substrate 1 is adjusted so that a straight line connecting the center points of the X-ray source 3 and the FPD 4 obliquely crosses the inspection region. Further, the fluoroscopic imaging is performed a plurality of times by changing the relationship of the FPD 4. However, the number of imaging times and the arithmetic processing for image reconstruction are completely different between tomosynthesis and X-ray CT.

FIG. 3 shows a specific example of projection processing when tomosynthesis is executed.
This figure shows a change in the relationship between the three parties assuming that the X-ray source 3, the FPD 4, and the reference cross section T are seen from directly above. As the reference cross section T, an arbitrary cross section can be selected from a plurality of target cross sections.

  In this embodiment, one point O in the reference cross section T (which is a central point in this embodiment, but is not limited to this) is set as a reference point, and in each of the X and Y axial directions, respectively. The position of the substrate 1 is changed so that the reference point O sequentially moves to the four points P1, P2, P3, and P4 having the same distance from the optical axis L of the X-ray tube 3. Further, when the reference point O is aligned with these points P1, P2, P3, P4, the center point R of the FPD 4 is placed at a position (points Q1, Q2, Q3, Q4) where this point O is projected. As described above, the FPD 4 is moved in accordance with the movement of the substrate 1, and fluoroscopic imaging is performed for each movement. Specifically, when the reference point O moves to the point P1, the center point R moves to the point Q1, and when the reference point O moves to the point P2, the center point R moves to the point Q2, and the reference point O When moved to P3, the center point R moves to the point Q3, and when the reference point O moves to the point P4, the center point R moves to the point Q4.

  Since the conical beam is emitted from the X-ray tube 3 of this embodiment, the X-ray irradiation angle with respect to each direction becomes uniform. Further, since the distances of the points P1 to P4 with respect to the optical axis L are equal, the distances of the points Q1 to Q4 with respect to the optical axis L are also equal.

  Note that the positions where the substrate 1 and the FPD 4 are positioned are not limited to the four directions indicated by the positive and negative X and Y axes. For example, in addition to the above four azimuths, in a total of eight azimuths that are azimuthally inclined by 45 degrees with respect to the X and Y axes, each point having the same distance to the optical axis L is designated as the reference point O. The positioning points of the FPD 4 may be set in correspondence with these points.

  When the substrate 1 is moved so that the distance of the reference point O with respect to the optical axis L is constant as in the above example, the reference point O is along the circumference of a circle centered on the optical axis L. However, the trajectory of the reference point O is not limited to a circular trajectory. For example, the position of the substrate 1 every hour may be determined so that the position of the reference point O changes with an elliptical or rectangular orbit.

In the following, on the assumption that the substrate 1 and the FPD 4 are moved by the method shown in FIG. 3, the position of the reference cross section T is shown using points P1 to P4 at which the reference point O is aligned, and the position of the FPD 4 is indicated by the point Q1. It will be shown using ~ Q4.
In FIG. 4, the projection state when the reference cross section T is placed at the position of the point P1 and the projection state when placed at the position of the point P2 are shown in contrast. Moreover, the schematic diagram of X-ray fluoroscopic images A1 and A2 generated at each position is shown in association with each position.

  In FIG. 4, reference numeral 12 denotes a component mounted at a position facing the inspection region with the substrate 1 interposed therebetween (hereinafter referred to as “back surface component 12”). In this example, in order to simplify the description, the detection surface 4A of the FPD 4 and each X-ray of the constituent points of the solder electrode 10 in the reference cross section T and the upper surface of the back surface component 12 (surface in contact with the substrate 1), respectively. Projection ranges in the perspective images A1 and A2 are indicated by oblique line patterns having different inclination directions. In order to clarify the correspondence, the same range (S1, U1, or S2, U2) is assigned to the projection range of the detection surface 4A and the projection ranges in the images A1 and A2. In addition, although the background portions (outside portions of the projection ranges) of the images A1 and A2 are generally represented in black, in this example, the background color (white) of the paper surface is easy to confirm each projection range. And

  Hereinafter, S1 and S2 are referred to as “projection range of the solder electrode 10”, and U1 and U2 are referred to as “projection range of the back surface component 12,” but S1 and S2 are only the projection range of the solder portion in the reference cross section T. And does not represent a projected image of the entire solder electrode 10. U1 and U2 are also projection ranges of the bottom surface (in the drawing, the upper surface in the drawing) of the back surface component 12 that is in contact with the substrate 1, and do not represent the entire projected image of the back surface component 12.

  The X-ray tube 3 of this embodiment is irradiated with a conical beam that uniformly spreads in each direction, and the reference point O of the reference section T is always projected onto the center point R of the FPD 4 with the substrate 1. The position with the FPD 4 is adjusted. Therefore, even if the substrate 1 moves and the position of the reference point O changes, each point in the reference cross section T including the reference point O is always projected on the same coordinate on the detection surface 4A. On the other hand, since the constituent point of the back surface part 12 is at a different height from the reference cross section T, it is projected to a different coordinate every time. Therefore, the projection ranges S1 and S2 of the solder electrode 10 in the hourly fluoroscopic imaging are the same, but the projection ranges U1 and U2 of the back surface component 12 are different.

  The relationship between when the reference cross section T is at the position of the point P3 and when it is at the position of the point P4 is the same as in the example of FIG. 4, and X-ray fluoroscopy as indicated by reference numerals A3 and A4 in FIG. An image is obtained.

  In the tomosynthesis image reconstruction process according to this embodiment, coordinates correspond to the pixels of the four fluoroscopic images A1 to A4 generated by aligning the reference section T with the points P1, P2, P3, and P4. Each of these combinations is combined, and for each of these combinations, data of the pixel having the lowest density among the four pixels belonging to the set (that is, the X-ray transmittance is the highest among the four pixels). Data) image data. Then, the tomographic image of the reference cross section T is generated by fitting the selected image data to the coordinates common to each set.

  Further, with respect to the target cross section other than the reference cross section T, based on the distance between the cross section and the reference cross section T, the above-described images A1 to A4 are images when the target cross section is the reference cross section. The coordinates are changed (that is, the images A1 to A4 are corrected so that the coordinates of the projection points of the respective points in the target cross section match). Then, using the four images after correction, as in the case of the reference cross section T, the target image is selected by a method of selecting image data of the pixel having the lowest density for each combination of pixels having coordinates corresponding to each other. A tomographic image of a cross section is generated.

  For the reference cross section T, it is not always necessary to determine the position of the substrate 1 or the FPD 4 so that each point in the cross section T is projected on the same coordinate. However, in that case, it is necessary to correct the images A1 to A4 in the same manner as described above for the reference cross section T.

  Next, specific examples of tomographic image reconstruction and the influence of the back part 12 will be described with reference to FIGS. In this example, image reconstruction using the X-ray fluoroscopic images A1 to A4 of the reference cross section T photographed by the method of FIG. 3 is shown. However, even when these images are corrected and reconstruction processing is performed. The influence of the back surface part 12 on the image reconstruction is the same as that shown in FIGS.

  5 to 7, X-ray fluoroscopic images A1, A2, A3, and A4 generated by aligning the reference section T with the points P1, P2, P3, and P4 are the four points P1, P2, and P3 shown in FIG. , P4, and a tomographic image B reconstructed by each image is arranged at the center of these images.

  S1 to S4 in the images A1 to A4 are projection ranges of the solder electrode 10, and U1 to U4 are projection ranges of the back surface component 12. As described with reference to FIG. 4, the constituent points of the solder electrode 10 in the reference section T are projected on the same coordinates in any of the fluoroscopic images A1 to A4, but the constituent points of the back surface component 12 are projected. The coordinates to be changed vary depending on the image.

  In the example of FIG. 5, the back surface component 12 is projected on the outside (background portion) of the projection range S1 to S4 of the solder electrode 10 in any of the images A1 to A4. In addition, the projection ranges U1 to U4 of the back surface part 12 in each of the images A1 to A4 do not overlap with other ranges.

  Therefore, according to the method of selecting the data of the pixel having the lowest density among the four pixels having the corresponding relationship, any image data is selected for the projection ranges S1 to S4 of the solder electrodes 10 that match between the images. Even so, data representing the solder electrode 10 is selected. On the other hand, regarding the background portion, even if there is a portion where the density is high because the component 12 is projected in any image, the density of the same portion in the image where the other component 12 is not projected is more. Since it is low, data representing the projection state of the component 12 is not selected. Therefore, although the image S of the solder electrode 10 appears clearly in the tomographic image B, the image of the back surface component 12 does not appear.

  In FIG. 5, portions corresponding to the projection ranges U1 to U4 of the component 12 in the tomographic image B are indicated by dotted line frames. When image reconstruction is performed by a general image averaging process, a light image may appear in the range of the dotted frame.

6 shows X-ray fluoroscopic images A1, A2, A3, and A4 when the projection ranges U1 to U4 of the back surface component 12 become larger than the example of FIG. 5, and the tomogram reconstructed by these images A1 to A4. Image B is shown.
In this example, in any of the images A1 to A4, there is an overlap between the projection ranges S1 to S4 of the solder electrode 10 and the projection ranges U1 to U2 of the back surface component 12. In the part where the projection ranges overlap, the density is higher than in the other parts, so the whiteness of the image becomes stronger. However, when the appearance ranges of overlapping portions in the images A1 to A4 are compared, X-rays at the time of photographing are present although there are portions that are commonly included between images in which the orientations of the X-ray tube 3 at the time of photographing are orthogonal to each other. There is no common location between images in which the orientations of the tubes 3 are in conflict. Therefore, since there is no place common to all the projection ranges U1 to U4 of the back surface component 12 in the images A1 to A4, the projection ranges S1 to S4 of the solder electrode 10 do not always overlap the projection range of the back surface component 12. Image data is selected. For the background portion, as in the example of FIG. 5, data representing the background is always selected, so that an image of the back surface component 12 does not appear in the tomographic image B.

  FIG. 7 shows X-ray fluoroscopic images A1, A2, A3, A4 and a tomographic image B reconstructed from these images A1 to A4 when the projection ranges U1 to U4 of the back part 12 are further increased.

  In this example, the projection ranges U1 to U4 of the back surface component 12 in the X-ray fluoroscopic images A1 to A4 overlap with more than half of the projection ranges S1 to S4 of the solder electrode 10, and as a result, the central portion of the image is It is commonly included in the four projection ranges U1 to U4 of the back part 12. Since the image data in which the density of the back surface component 12 is taken into consideration is selected at the locations that are commonly included in the projection ranges U1 to U4, noise N having a higher density than the surroundings is present at the corresponding location of the reconstructed tomographic image B. appear.

  Next, in image reconstruction by X-ray CT, the position of the substrate 1 and the FPD 4 is changed variously by tilting the FPD 4 as shown by the one-dot chain line in FIG. Even if the position changes, the axis connecting the center points of the FPD 4 and the X-ray tube 3 is maintained in a state of crossing the target cross section obliquely at the position of the specific reference point. Further, the relationship between the substrate 1, the X-ray tube 3, and the FPD 4 changes in the same manner as when the X-ray source 3 and the FPD 4 are fixed and the target cross section is rotated by a predetermined angle unit around the reference point. Then, the positions of the substrate 1 and the FPD 4 are changed, and the X-ray tube 3 and the FPD 4 are driven each time the position is changed to perform fluoroscopic imaging.

  An X-ray fluoroscopic image generated by photographing every hour is input to the arithmetic processing unit 23, and the same method as disclosed in the above-mentioned Patent Document 1, that is, each input image is arranged in the normal direction of the substrate 1. A tomographic image of the target cross-section is generated by a method of converting to an image representing the state of fluoroscopy from the orthogonal direction and calculating the X-ray absorption coefficients of a plurality of points in the target cross-section R using each converted image. Is done.

  In X-ray CT, it is necessary to execute a plurality of fluoroscopic imaging and image reconstruction by the above method for each target section.

Although image reconstruction by X-ray CT takes time, processing other than the reference cross section T is removed by a complicated calculation, so that a highly accurate tomographic image can be obtained.
On the other hand, in the image reconstruction by tomosynthesis, noise N may occur. However, since the number of fluoroscopic imaging is much smaller than that of X-ray CT, imaging can be completed in a short time. The calculation for image reconstruction is also simple and can be executed only by the control unit 21. Furthermore, even when a tomographic image of a plurality of surfaces is required, if fluoroscopic imaging is performed on any one surface, the X-ray fluoroscopic image obtained by the imaging is converted into an X-ray fluoroscopic image of a surface of another height. Transform and tomographic images at that height can be reconstructed.

  As described above, if image reconstruction by tomosynthesis is performed, the processing time can be greatly shortened. Therefore, the user in the field desires to perform image reconstruction by tomosynthesis as much as possible. However, when noise N as shown in FIG. 7 occurs, there is a possibility that a misjudgment that the solder electrode 10 is defective may occur. Therefore, the inspection by tomosynthesis is given up, and a high-accuracy fault by X-ray CT is given. It is necessary to perform an inspection using an image.

  Therefore, in this embodiment, it is determined whether or not noise N caused by the back surface component 12 occurs when image reconstruction by tomosynthesis is performed on this area for each inspection area in teaching before inspection. According to the determination result, it is determined which of the two types of image reconstruction methods is to be executed. This determination process is automatically performed by the control unit 21 using the CAD data of the substrate 1. Hereinafter, the principle of this discrimination process will be described with reference to FIGS.

  8 replaces the imaging processing shown in FIGS. 3 and 4 with the substrate 1 being stationary and the X-ray tube 3 and the FPD 4 being relatively moved, and the orientation with respect to the reference point O in the reference section T is contradictory. The relationship between the X-rays from the two directions and the reference cross section T and the back surface component 12 is shown. The straight line m1 in the figure corresponds to the X-ray traveling direction with respect to the point O when fluoroscopic imaging is performed with the reference point O aligned with the point P1 (the left hand imaging state in FIG. 4), and the straight line m2 is the reference point This corresponds to the advancing direction of the X-ray with respect to the point O when fluoroscopic imaging is performed by matching O with the point P2 (right-hand imaging state in FIG. 4). Hereinafter, transmission imaging using X-rays from the direction corresponding to the straight line m1 is referred to as “first imaging”, and transmission imaging using X-rays from the direction corresponding to the straight line m2 is referred to as “second imaging”.

  X-rays from the directions indicated by the straight lines m1 and m2 pass through the back surface of the substrate 1 at the positions of points n1 and n2, respectively, and reach the center point R of the detection surface 4A of the FPD 4. Therefore, the points n1 and n2 are projected at the same position as the reference point O, that is, the point R.

  In FIG. 8, the projection state of each point below the back surface of the substrate 1 on the detection surface 4A of the FPD 4 on the basis of the vertical axis G passing through the reference point O will be described as a range from a point n1 to a point n2 ( Hereinafter, each point in “between n1 and n2” is projected at a position before the center point R when viewed from the axis G in both the first and second imaging. On the other hand, the point located on the left side of the point n1 is projected outside the center point R in the first photographing. In addition, the point located on the right side between the points n2 is projected outside the center point R in the second imaging.

  In FIG. 8, the back surface component 12 whose mounting range is included between n1 and n2 is indicated by a solid line, and the back surface component 12 whose mounting range is greater than n1 and n2 is indicated by a dashed line.

  FIG. 9 shows the projection ranges U1 and U2 in the first and second imaging for the two types of parts 12 shown in FIG. 8 as hatched patterns having different directions in the cross section of the FPD 4 respectively. In addition, a and b in the figure indicate the positions of the edges of the bottom surface of the back surface component 12 in contact with the substrate 1, a1 and b1 are the projection positions of the points a and b in the first imaging, and a2 and b2 are the second. Projection positions of points a and b in photographing are shown respectively.

  FIG. 9 (1) shows the projection state of the back surface component 12 shown by the solid line in FIG. In this component 12, since the points a and b are both between n1 and n2, the points a and b are projected to a position before the center point R when viewed from the axis G in both the first and second imaging. The Therefore, in this case, there is no overlap between the projection ranges U1 and U2.

  FIG. 9 (2) shows a projected state of the back surface component 12 shown by the one-dot chain line in FIG. In this component 12, since the points a and b are located outside between n1 and n2, the point a is projected farther from the center point R in the first photographing and the point b in the second photographing. Therefore, since any projection range U1, U2 exceeds the center point R, overlap occurs between the projection ranges U1, U2.

  That is, in the case of the above example, when the width w of the back surface component 12 is smaller than the distance between n1 and n2, the projection range U1 of the back surface component 12 in the first shooting and the projection of the back surface component 12 in the second shooting. Although the range U2 does not overlap, when the width w of the back surface component 12 exceeds the distance between n1 and n2, the projection ranges U1 and U2 overlap.

Here, returning to FIG. 8, the irradiation angle of X-rays from each direction with respect to the reference point O is represented by the angles θ1 and θ2 of the straight lines m1 and m2 with respect to the axis G, and the height of the reference section T is h. The distance df between n1 and n2 is obtained by the following equation (1).
df = h · (tan θ1 + tan θ2) (1)

  In the above, the angles θ1 and θ2 are predetermined (in this embodiment, θ1 = θ2), and the height h of the reference cross section T is also determined in setting the target cross section. Therefore, the distance df between n1 and n2 can be obtained by applying these values to the equation (1). Further, if the width w of the back surface component 12 is obtained from, for example, CAD data of the substrate 1, and the value of w is compared with df, it can be determined whether or not the projection ranges U1 and U2 of the component 12 overlap. .

  In addition, when the back surface part 12 is a rectangular parallelepiped shape, each cross section other than the bottom surface is projected in a range shifted to the near side (side closer to the X-ray tube 3) than the projection range of the bottom surface. If there is no overlap in the projection ranges U1 and U2 on the bottom surface of the part 12 that is in contact with, the projection ranges in other sections do not overlap.

  However, even when the width w of the back surface component 12 is smaller than the distance df between n1 and n2, the entire bottom surface of the component 12 is not always located between n1 and n2, as in the example of FIG. . In view of this point, one end of the bottom surface of the back surface component 12 is between n1 and n2, but the other end is between n1 and n2, and the width w and distance df of the above component 12 It is necessary to consider whether it is possible to determine by comparing the two.

FIG. 10 is a diagram showing the results of the above examination. In this figure, it is assumed that the part 12 indicated by the solid line in FIG. 8 is shifted to the right by the distance d1, and the X-ray fluoroscopic image A1 generated by the first imaging with the X-ray from the direction corresponding to the straight line m1. And an X-ray fluoroscopic image A2 generated by the second imaging with X-rays from the direction corresponding to the straight line m2. In this case, the projection ranges U1 and U2 of the back surface component 12 in each of the images A1 and A2 are respectively relative to the projection ranges U1 ′ and U2 ′ (indicated by alternate long and short dash lines) when the component 12 is in the position of FIG. It shifts to the right by a certain distance d2.
Thus, even if the position of the component 12 with respect to n1-n2 shifts, the projection ranges U1, U2 of the component 12 in the images A1, A2 shift in the same direction by a certain amount d2 according to the actual component shift. Only. Therefore, if the projection ranges U1 ′ and U2 ′ do not overlap, the projection ranges U1 and U2 do not overlap.

  Therefore, regardless of the position of the back surface part 12, by comparing the distance df between n1 and n2 obtained by the above equation (1) and the width w of the back surface part 12, the first image capturing and the second image capturing are performed. It is possible to easily determine whether or not the projection ranges U1 and U2 of the back part 12 are overlapped.

  Further, fluoroscopic imaging (hereinafter referred to as “third imaging”) in which the reference section T is aligned with the point P3 and the FPD4 is aligned with the point Q3, the reference section T is aligned with the point P4, and the FPD4 is aligned with the point Q3. In the case of fluoroscopic imaging (hereinafter referred to as “fourth imaging”) in the state of being in a state of being overlapped with the projection ranges U3 and U4 of the back surface component 12 in the same manner as that performed for the combination of the first imaging and the second imaging. Can be determined.

  As shown in FIGS. 5 to 7 above, in the image construction processing by tomosynthesis of this embodiment, each fluoroscopic imaging is combined for each of the X-ray azimuths with respect to the reference point O that are in an opposite relationship (that is, If the first image capturing and the second image capturing are combined, and the third image capturing and the fourth image capturing are combined), and at least one of the combinations does not cause an overlap in the projection range of the back surface component 12, noise N is generated. do not do.

  Therefore, as a result of comparing the width w and distance df of the above-described back surface component 12 for each combination, if w <df in at least one of the combinations, the back surface component 12 in the four X-ray fluoroscopic images A1 to A4. No overlap occurs between all the projection ranges U1 to U4, and no noise N is generated. Therefore, image reconstruction by tomosynthesis can be applied to the examination region where such a relationship is established.

  On the other hand, in any combination, when w> df, the projection ranges U1 to U4 of the back surface part 12 in photographing every hour are overlapped as shown in the example of FIG. Therefore, it is necessary to apply image reconstruction by X-ray CT to the examination region where such a relationship is established.

  Similarly, when the number of fluoroscopic imaging increases, for each of a plurality of imaging groups in which the X-ray source orientations are opposed to each other, the direction in which the X-ray source orientation conflict occurs for each combination The distance between the positions n1 and n2 where the X-ray with respect to the reference point O passes the back surface of the substrate 1 is compared with the width w of the component. In this case as well, noise N does not occur unless the projection range of the back surface component 12 overlaps in all combinations. Therefore, image reconstruction by tomosynthesis is possible by comparing the above w and df for each combination of imaging in which the X-ray orientation with respect to the target cross section at the time of imaging opposes regardless of the number of imaging. It can be determined whether there is.

8 to 10, for convenience of description, the principle for determining the presence or absence of noise due to the back surface component 12 is described using the horizontal plane passing through the center of the solder ball 10 as the reference cross section T. In the processing, it is necessary to execute the expression (1) for a cross section closest to the substrate 1 among a plurality of target cross sections (that is, a cross section having the smallest value of h). According to equation (1), the smaller the value of h, the smaller the distance df. Therefore, the minimum value of df can be obtained by taking the lowest cross section as the object of calculation. When the width w of the back part is smaller than the minimum df, no noise is generated due to the overlap of the projection range of the back part in any cross section. On the other hand, if the width w is equal to or greater than the minimum df, noise occurs due to the overlap of the projection range of the back part in any cross section.
The cross section used for calculating the distance df may be the reference cross section T in actual photographing, but no particular problem occurs even if a cross section other than the cross section used for the calculation is used as the reference cross section T.

  Next, in the example of FIGS. 8 to 10, there is one back surface component 12 corresponding to the reference cross section T. However, depending on the component arrangement on the back surface side, as shown in FIG. There may be a plurality of parts 12 in the range through which the line is transmitted. Also in such a case, if the distance D between the parts 12 is within a predetermined threshold value, these parts are regarded as one, and the value obtained by adding the width and the distance D of each part is set as w. May be executed.

  In the above determination, the back surface component 12 is assumed to be a rectangular parallelepiped. However, when the shape of the back surface component 12 is other than a rectangular parallelepiped, such as the area of the upper surface is larger than the bottom surface of the component. Assuming a rectangular parallelepiped circumscribing this part, it is desirable that the width of the rectangular parallelepiped is w. Further, when there is a reciprocal orientation of the X-ray source in directions other than X and Y, the width w may be obtained by circumscribing a rectangular parallelepiped having sides parallel to that direction.

  Further, in the above example, the case of inspecting the solder electrode 10 on the upper surface side of the substrate 1 supported by the substrate support stage 2 has been examined. Similarly, when the solder electrode 10 on the lower surface side of the substrate 1 is inspected, By comparing the distance df obtained by the above equation (1) with the width of the component on the upper surface side for each combination of photographing in which the orientation with respect to the reference point in the cross section is opposite for the cross section closest to the substrate 1 It is possible to determine whether or not noise is generated by the component. However, in the equation (1) in this case, the distance from the upper surface of the substrate to the cross section to be calculated needs to be h.

  FIG. 12 shows a teaching procedure. In the teaching of this embodiment, when a user gives a simple instruction, an inspection region to be inspected for the solder electrode 10 and a method for reconstructing a tomographic image are automatically set.

  First, in step 1, CAD data of a substrate 1 to be inspected is input. In step 2, based on the thickness of the substrate 1 and the solder electrode 10, the number of target cross sections suitable for observation of the solder electrode 10 and the height of the reference cross section are set. In addition, although the space | interval between target cross sections shall be fixed, this space | interval can be changed suitably.

In this embodiment, a constant value corresponding to the size that can be projected onto the FPD 4 is set as the area of the inspection region. In step 3, information on the BGA component 11 is extracted from the CAD data, The inspection area is allocated using a certain area. Depending on the size of the BGA component 11, a plurality of inspection areas are allocated to one component.
Thereafter, paying attention to the set inspection areas in order, Steps 4 to 11 are executed for each area.

  First, in step 4, the cross section closest to the substrate 1 (the one at the lowest position) among the target cross sections set in step 2 is set as a calculation target. In step 5, the distance df between n1 and n2 is calculated using the height h of the cross section to be calculated and the X-ray irradiation angle, and the calculated df value is used as the component width w. Is set as the criterion value for. The height h can be determined by the height of the reference cross section T, the interval between the target cross sections, and the thickness of the substrate 1.

  Also in this example, it is assumed that four times of imaging are performed based on the method shown in FIG. 3, and a combination of imaging (first imaging, second imaging, and third imaging) in which the orientations of the X-ray tube 3 are contradictory to each other. The determination reference value df is obtained every time shooting and fourth shooting).

  In step 6, a part in which the entire body or a part of the body is opposed to the inspection area being focused on the substrate 1 is identified based on CAD data, and the part corresponds to each imaging combination. The component width w is read for each direction (X direction and Y direction). Note that when a rectangular parallelepiped circumscribing the component is set due to problems such as the shape of the component, it is necessary to calculate the value of w based on the component size and mounting direction indicated by the CAD data.

  In Step 7, the judgment reference value df obtained in Step 5 is compared with the width w of the component 12 specified in Step 6, thereby overlapping the projection ranges U1 to U4 of the back surface component 12 by X-rays from each direction. Determine if it occurs. Specifically, each of the combinations of imaging in which the orientations of the X-ray tube 3 at the time of imaging are contradictory, that is, the combination of the first imaging and the second imaging, and the combination of the third imaging and the fourth imaging, respectively. The size relationship between the width w and df is checked, and if at least one of the combinations satisfies w <df, it is determined that there is no overlap between the projection ranges U1 to U4. On the other hand, in any combination, when w ≧ df, it is determined that overlap occurs between the projection ranges U1 to U4.

  In the above processing, when it is determined that there are overlapping portions in the four projection ranges U1 to U4 (when step 8 is “YES”), the X-ray CT is performed on the examination region of interest. Determine (step 9). On the other hand, when it is determined that there is no overlap between the projection ranges U1 to U4 (when step 8 is “NO”), it is determined to perform tomosynthesis on the examination region under consideration (step 10). ).

  Hereinafter, similarly, for each inspection area assigned in step 3, it is determined whether or not there is an overlap between the projection ranges U1 to U4 of the back surface component 12 in each of the images A1 to A4. Decide how to configure. The method determined here is registered in the memory 22 in association with information such as the position and size of the inspection region.

  When the processing for all the inspection areas is completed, step 11 becomes “YES” and the teaching is finished.

  Thereafter, when the inspection mode is entered, each time the substrate 1 to be inspected is carried into the substrate support table 2 and the initial positioning is performed, fluoroscopic imaging or tomography for each inspection region is performed based on the information registered by the above procedure. Image reconstruction is performed. In this case, for example, first, the tomosynthesis image reconstruction is performed, the target inspection areas are sequentially processed, and then the X-ray CT method image reconstruction is switched to perform the processing on the remaining inspection areas. By executing this, the processing efficiency can be improved.

  In the above embodiment, the width w of the back surface component 12 is derived from the CAD data. However, the method for acquiring this type of data is not limited to reading from the CAD data. For example, before the inspection, when the board 1 is imaged by the CCD camera 5 and the user manually sets the position and size of the part while referring to the generated image, the input data is used. The width w of the part may be specified. For example, if a window having a size including the part is set for each part, the width of the window frame indicated by the setting data of the window may be used as the part width w.

  As described above, in the above-described embodiment, since the data of the pixel having the lowest density among the pixels in the correspondence relationship is selected in the tomosynthesis image reconstruction, the four X-ray fluoroscopic images A1 to A4 are selected. The occurrence of duplication in the projection ranges U1 to U4 of the back surface component 12 in FIG. However, when the images A1 to A4 are averaged according to a general tomosynthesis method, in any one of the combination of the first shooting and the second shooting and the combination of the third shooting and the fourth shooting X-ray CT may be executed when it is determined that the projection range of the back part 12 overlaps (a state where w ≧ df).

  In addition, when the number of times of photographing is more than four, it is determined whether or not the projection range of the back surface part 12 overlaps for each photographing combination in which the orientation with respect to the reference section T is contradictory. X-ray CT may be executed when it is determined that the combinations overlap. Further, if the number of combinations determined that the projection ranges of the rear surface parts 12 overlap does not reach the predetermined number, the X-ray CT, You may make it make a user select which tomosynthesis is performed.

It is explanatory drawing which shows the structural example of the board | substrate inspection apparatus by X-ray tomography. It is a block diagram of said board | substrate inspection apparatus. It is explanatory drawing which shows the positional relationship of a X-ray tube, FPD, and a target cross section in the case of performing fluoroscopic imaging by tomosynthesis. It is explanatory drawing which matches and shows the projection state by the fluoroscopic imaging in which the azimuth | direction with respect to a target cross section has a contradictory relationship, and the produced | generated X-ray fluoroscope image. It is explanatory drawing which shows the relationship between the X-ray fluoroscopic image produced | generated by the imaging | photography of 4 times, and the reconstructed tomographic image. It is explanatory drawing which shows the relationship between the X-ray fluoroscopic image produced | generated by the imaging | photography of 4 times, and the reconstructed tomographic image. It is explanatory drawing which shows the relationship between the X-ray fluoroscopic image produced | generated by the imaging | photography of 4 times, and the reconstructed tomographic image. It is explanatory drawing which shows the relationship between the X-ray passage position and the edge position of components in 1st imaging | photography and 2nd imaging | photography. It is explanatory drawing which shows the relationship of the projection range in 1st imaging and 2nd imaging for 2 types of components shown in FIG. It is explanatory drawing which shows the change of the projection range of components when a back surface component shift | deviates from the state of FIG. It is explanatory drawing which shows the example where several components are located in the range which X-ray permeate | transmits. It is a flowchart which shows the procedure of teaching.

Explanation of symbols

1 Substrate 3 X-ray tube 4 Flat panel detector (FPD)
8, 9 XY stage 10 Solder electrode 12 Back part 20 Control device 21 Control unit 22 Memory T Target section O Center point of reference section T R Center point of FPD4 U1, U2, U3, U4 Projection range of back part 12

Claims (3)

For a double-sided mounting board, a tomographic image of the solder electrode connecting the electronic component mounted on this board and the board side electrode is generated using X-rays, and the state of the solder electrode is inspected using this tomographic image A way to
An X-ray source and an X-ray detector arranged at different height positions, a substrate support portion for horizontally supporting the substrate to be inspected therebetween, and an X-ray source and an X-ray detector for the substrate to be inspected And a control processing unit that generates an X-ray fluoroscopic image of the substrate by adjusting the relative position of the X-ray source so that one or more pairs of azimuths having a contradictory relationship as the azimuth of the X-ray source with respect to the target cross section are generated. , Image reconstruction by tomosynthesis that generates a plurality of fluoroscopic images by adjusting the relative position, and generates a tomographic image of the target cross section by integrating each X-ray fluoroscopic image based on the projection range of the target cross section And X-ray CT image reconstruction that generates a tomographic image by generating more X-ray fluoroscopic images than the tomosynthesis and obtaining an X-ray absorption coefficient distribution of the target cross section using these images. To prepare a test apparatus that can perform switching,
Data indicating the configuration of the double-sided mounting board to be inspected is input to the control processing unit prior to inspection,
In the control processing unit after the data input, after determining, using the input data, which of the two types of image reconstruction is to be executed for each predetermined photographing target region, and registering the determination result , For each imaging target area of the substrate to be inspected, the tomographic image is automatically generated by the registered image reconstruction, and the inspection is executed.
In the process of determining the content of image reconstruction for one shooting target area,
Assuming that image reconstruction by the tomosynthesis is performed on the imaging target region, the plurality of X-ray fluoroscopic images are combined for each of the X-ray source orientations with respect to the target section at the time of image generation that are in a contradictory relationship. For each combination, when an X-ray fluoroscopic image related to the combination is aligned based on the projection range of the target cross section, an overlapping portion is generated in the projection range of the part behind the region to be inspected in the imaging target region. Is determined based on the relationship between the size of the back side component obtained from the input data and the size of the range in which the X-ray with respect to the target cross section at the time of each image generation passes through the back side substrate surface of the imaging target area. And determining that the image reconstruction by the X-ray CT is to be executed when the determination result that the overlapping of the projection ranges occurs for a predetermined number of combinations is obtained, When combinations determination result is obtained that overlap occurs the circumference is less than the predetermined number, it determines that it will perform the image reconstruction by the tomosynthesis,
A method for inspecting a solder electrode using an X-ray tomographic image. The method of claim 1, comprising:
In the process of determining the content of image reconstruction for the one shooting target area,
Hourly fluoroscopy based on the assumed tomosynthesis is combined for each of the X-ray source orientations that are in conflict with the target cross section, and for each combination, the direction in which the X-ray source orientation conflict occurs for each combination The X-ray passing range with respect to the substrate surface on the back side of the imaging target region is determined based on the X-ray irradiation angle from each direction irradiated on a predetermined position of the target cross section and the height of the target cross section. Find, compare the obtained size with the size of the back part obtained from the input data,
In a predetermined number of combinations, if the part is larger than the X-ray passage range, it is determined that the projection range of the back part overlaps, and it is determined to perform image reconstruction by X-ray CT. A method for inspecting a solder electrode using an X-ray tomographic image. An X-ray source and an X-ray detector arranged at different height positions, a substrate support unit for horizontally supporting the substrate to be inspected therebetween, and a solder electrode on the substrate to be inspected Adjusting the relative positions of the X-ray source and the X-ray detector with respect to the substrate so that the electronic component is imaged, performing a plurality of fluoroscopic imaging, and from the generated X-ray fluoroscopic images, A control processing unit that generates a tomographic image, and an inspection unit that inspects the state of the solder electrode using the tomographic image generated by the control processing unit,
The control processing unit generates a plurality of fluoroscopic images by adjusting the relative position so that one or more pairs of azimuths having a contradictory relationship with each other as the azimuths of the X-ray source with respect to the target cross section are generated. A first processing unit that performs image reconstruction by tomosynthesis that generates a tomographic image of the target cross-section by integrating each X-ray fluoroscopic image based on the projection range of the target cross-section, and more X-rays than the tomosynthesis A second processing unit configured to generate a fluoroscopic image and perform image reconstruction by X-ray CT to generate a tomographic image by obtaining an X-ray absorption coefficient distribution of the target cross section using these images; and a predetermined processing unit A registration unit for determining whether to perform image reconstruction by the first or second processing unit for each imaging target region, and registering information indicating the determination result, and a registration unit for registering the information. Based on the information, it includes a control unit for executing image reconstruction processing to the processing unit determined for that region for each imaging target region,
When the solder electrodes of the double-sided mounting board are to be inspected, the registration unit receives input of data indicating the configuration of the double-sided mounting board, and performs image reconstruction by the tomosynthesis for each imaging target area. Assuming that the configuration is executed, the plurality of X-ray fluoroscopic images are combined for each of the X-ray source orientations with respect to the target cross section at the time of image generation which are in a contradictory relationship, and for each combination, the X-ray fluoroscopy related to the combination Whether or not there is an overlapping portion in the projection range of the back side component of the imaging target area when the image is aligned with respect to the projection range of the target cross section, the size of the back side component obtained from the input data and The X-ray with respect to the target cross section at the time of each image generation is discriminated based on the relationship with the size of the range that passes through the substrate surface on the back side of the region to be inspected. When the determination result that the projection range overlaps is obtained, it is determined that the image reconstruction by the X-ray CT is performed, and a predetermined number of combinations obtained the determination result that the projection range overlap occurs. If not, a substrate inspection apparatus that determines to perform image reconstruction by the tomosynthesis.

Images