Test method and system for simulating filling effect of adhesive on BGA (Ball Grid Array) chip

Abstract

The invention relates to the technical field of packaging tests, in particular to a test method and system for simulating the filling effect of an adhesive on a BGA chip, and the method comprises the following steps: the BGA chip filling test method comprises the steps: obtaining a chip size and an image sample, setting adhesive spraying parameters, extracting boundary pixels, collecting a fitting image, and evaluating the stability; and injecting simulation colloid, extracting a propulsion contour, and counting a filling image to generate an effect overview. According to the method, a structural boundary image set is constructed by extracting packaging design drawing parameters, glue spraying, attaching and glue injection operations in a simulation environment are realized in combination with an image recognition and path judgment mode, a glue propelling state and boundary changes are continuously recorded by using an image frame sequence, and traditional slice observation is replaced by a whole-process image recognition means. And material consumption and slicing errors are avoided, the test efficiency and controllability are improved, the precision and adaptability of filling process verification are enhanced, and the limitations of evaluation result lagging and high operation dependence in the prior art are improved.

Description

Technical Field

[0001] This invention relates to the field of packaging testing technology, and in particular to a test method and system for simulating the filling effect of adhesives on BGA chips. Background Technology

[0002] The field of packaging testing technology involves testing and evaluation methods for chips, circuit boards, and related materials during semiconductor packaging. This includes assessing the fill integrity of chip solder joint areas, observing the distribution of encapsulating adhesives, analyzing material compatibility, and verifying filling process parameters. Typically, the reliability and process effectiveness of the packaging structure are evaluated by constructing a test platform, applying observation methods, and conducting comparative experiments. Traditional BGA chip fill effect testing methods involve using a standard circuit board with BGA chips mounted on it. After soldering, actual adhesive material is applied, and after filling, cross-sectional processing or microscopic observation is performed to detect the degree of adhesive filling in the gaps below the BGA chip solder joints. The filling process typically uses a combination of materials such as actual BGA chips, circuit boards, and potting compound. Cross-sectional images of the filled state are obtained using techniques such as slicing, grinding, and microscopic imaging to determine the fill integrity and uniformity.

[0003] Current BGA chip filling effect testing techniques use actual BGA chips and standard circuit boards for filling tests. These techniques rely on actual soldering and adhesive potting processes, requiring waiting for the adhesive to cure before slicing, grinding, and microscopic observation to obtain images of the filling state. The testing process is complex and time-consuming, easily affected by batch-to-batch material variations and operational errors. Furthermore, it consumes a significant amount of testing materials, cannot flexibly control variables or achieve rapid iterative verification, and the observation results are mainly static cross-sectional images, failing to dynamically present the evolution of the filling process. There is a lack of real-time evaluation methods for the adhesive propagation path, edge filling continuity, and bonding process stability, resulting in an inability to comprehensively assess structural changes and process performance during the filling process. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a test method for simulating the filling effect of adhesives on BGA chips, comprising the following steps:

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a test method for simulating the filling effect of adhesives on BGA chips, comprising the following steps:

[0006] S1: Obtain the length, width and inter-sphere spacing of the BGA chip, set the glass cutting path, perform glass segmentation, collect the glass edge coordinates and judge the error value, filter image structure samples that meet the conditions, and generate a set of simulated packaging structure boundary images.

[0007] S2: Read the corner coordinates of the simulated packaging structure boundary image set, set the nozzle diameter and dot matrix spacing of the dispensing device, perform the spraying operation and check the stability of the spraying action, extract the boundary pixels of the spraying graphic, and construct the boundary map of the spraying positioning frame.

[0008] S3: Based on the adhesive spraying structure area in the boundary map of the adhesive spraying positioning frame, set the alignment and bonding path and collect bonding images during the bonding process, extract bonding error and response value, judge the alignment and bonding stability, and obtain the bonding stability evaluation result.

[0009] S4: Based on the bonding state corresponding to the bonding stability evaluation result, set the simulated adhesive injection path, activate the image acquisition module to record the colloid boundary image frame and extract the advancement contour pixels of each frame to generate a colloid filling advancement path map.

[0010] S5: Based on the colloidal trajectory in the colloidal filling propulsion path map, collect colloidal curing images, void images and edge continuity images, and statistically analyze image boundaries, contact relationships and void morphology to obtain an overview record of the simulated filling effect images of the BGA chip.

[0011] As a further aspect of the present invention, the image structure sample that meets the conditions specifically refers to the sample whose glass edge straight line fitting error value is less than a set error tolerance.

[0012] As a further aspect of the present invention, in the process of verifying the stability of the spraying action, it is determined whether the sequence of changes in the start and end coordinates in the dispensing displacement path is less than the nozzle return trajectory error threshold. If the threshold is met, it indicates that the action is stable.

[0013] As a further aspect of the present invention, the simulated packaging structure boundary image set includes a packaging boundary structure map, a corner coordinate distribution map, and image sample screening results; the adhesive spraying positioning frame boundary map includes a corner region colloid distribution map, an adhesive spraying path trajectory map, and an adhesive spraying boundary line pixel map; the bonding stability evaluation results include a bonding integrity index map, a bonding error distribution map, and a bonding stability judgment map; the colloid filling advancement path map includes an advancement path chain list, a leading edge advancement trajectory map, and an image frame sequence boundary map; and the BGA chip simulated filling effect image overview record includes colloid curing images, void distribution images, and edge filling continuity images.

[0014] As a further aspect of the present invention, the step of acquiring the simulated packaging structure boundary image set is as follows:

[0015] S111: Obtain BGA chip package design drawings, identify the chip length parameters, width parameters and chip ball spacing values ​​marked in the drawings, extract the pixel spacing information of the graphic area of ​​the drawings, convert the extracted size parameters into units, combine them to generate the size configuration required for package simulation, and obtain the chip package geometric parameter set;

[0016] S112: Based on the chip package geometric parameter set, set the starting coordinates of the displacement trajectory and the cutting path of the glass cutting device. After the glass cutting operation is performed, collect the corner coordinates of the four boundary cutting areas, construct the boundary contour curve data frame according to the order of the coordinates, and obtain the glass cutting corner coordinate sequence.

[0017] S113: Based on the coordinate sequence of the glass cutting edges and corners, calculate the minimum fitting residual value between each boundary fitting line and the actual edge and corner coordinate points, determine whether the residual is less than the preset edge fitting error tolerance, filter images that meet the conditions, extract the corresponding boundary contour information, and obtain a set of simulated packaging structure boundary images.

[0018] As a further aspect of the present invention, the step of obtaining the boundary map of the adhesive spraying positioning frame is as follows:

[0019] S211: Based on the corner coordinates in the simulated packaging structure boundary image set, extract the coordinate values ​​of each corner point pair in the image, calculate the arrangement relationship in the image coordinate system, obtain the glue head diameter configuration parameters and dot matrix spacing setting values ​​of the dispensing device, map the corner coordinates according to the outward expansion distance of the glue head aperture, determine the initial dispensing position and array arrangement spacing, and generate a dispensing device displacement configuration parameter set.

[0020] S212: Based on the displacement configuration parameter set of the dispensing device, collect all start and end coordinate sequences in the execution path of the dispensing device, calculate the residual difference between the displacement difference in each path segment and the nozzle return error benchmark value, determine whether the residual is less than the nozzle return trajectory error threshold, and if the set threshold requirement is met, record the current path sequence as a valid trajectory and obtain a stable dispensing trajectory path set.

[0021] S213: Based on the set of stable dispensing trajectory paths, extract the dispensing image information of the glass slide corner area, sample the gray-scale change line segments of pixels in the image boundary area, and construct a boundary curve index list according to the distribution density of pixels to establish a boundary map of the dispensing positioning frame.

[0022] As a further aspect of the present invention, the step of obtaining the attachment stability evaluation result is as follows:

[0023] S311: Based on the glue spraying graphic boundary in the glue spraying positioning frame boundary map, extract the corner coordinate point sequence of the boundary curve, set the contact path line segment and the initial displacement direction of the bonding device according to the curvature distribution of the boundary points, and perform a three-axis alignment operation in combination with the position parameters of the clamping platform plane reference surface to generate a bonding path configuration parameter set.

[0024] S312: Based on the attachment path configuration parameter set, monitor the instantaneous displacement of each attachment point during the glass sheet attachment process and record the deformation, compare the actual bonding height change of each attachment point in the adhesive spraying structure support area, determine whether the surface height deviation value of the bonding area is greater than the set bonding error threshold, and obtain the planar bonding deviation matrix of the adhesive spraying support area.

[0025] S313: Based on the planar bonding deviation matrix of the adhesive support area, collect the boundary bonding integrity information corresponding to each frame of the bonding process image sequence, calculate the contour overlap of the bonding area and the image boundary reconstruction rate, and compare the calculated values ​​with the stability judgment benchmark to establish the bonding stability evaluation result.

[0026] As a further aspect of the present invention, the step of obtaining the colloidal filling propulsion path map is as follows:

[0027] S411: Based on the adhesion stability evaluation results, set the starting coordinates of the simulated adhesive injection, and combine the stroke length of the injection system's propulsion mechanism with the adhesive application pressure output rate to calculate the injection speed parameter value and the pressurization cycle configuration set, thereby generating the adhesive injection action control parameter set.

[0028] S412: Read the set of control parameters for the colloid injection action, record the change process of the colloid boundary morphology in each frame of the image during the colloid filling process, number and sort the acquired image sequence according to the timestamp, perform contour detection operation on each frame of the image, extract the sequence of closed pixels of the boundary contour in the colloid propulsion area, and obtain the set of pixel coordinates of the colloid leading edge contour.

[0029] S413: Based on the pixel coordinate set of the colloid leading edge contour, the boundary points in each frame image are aggregated in time series to construct a contour trajectory chain of consecutive frames in the advancing direction, the centroid displacement vector of the contour coordinate between adjacent frames is identified, and all the movement trajectories of the leading edge points are connected in sequence to establish a colloid filling advancing path map.

[0030] As a further aspect of the present invention, the step of obtaining the BGA chip simulated fill effect image overview record is as follows:

[0031] S511: Based on the trajectory information in the colloid filling propulsion path map, extract the propulsion endpoint coordinates and boundary envelope distribution features, set the corresponding image acquisition area and activate the visual detection system, and sequentially acquire images of the solidified surface, void distribution, and edge filling continuity at the time point after the colloid injection is completed and solidified, generating a sequence of colloid solidification state images.

[0032] S512: Based on the image sequence of the colloid curing state, identify the boundary line between the colloid expansion area and the substrate structure boundary in the image, record the spatial position relationship between the outer edge pixel position of the colloid and the substrate reference point, and perform contour aggregation on the boundary curves of each filled area in the image to establish a dataset of the relationship between the filled boundary and the contact position.

[0033] S513: Based on the dataset of the relationship between the filling boundary and the contact position, aggregate the void distribution image, the edge continuous image and the solidified colloid boundary layer data, construct a structural image package including the location of the colloid residual voids, the contour shape of the filling area and the sealing features of the contact surface, and output an overview record of the simulated filling effect image of the BGA chip.

[0034] A testing system for simulating the filling effect of adhesives on BGA chips includes:

[0035] The package structure image generation module is used to execute S1: obtain the length, width and inter-sphere spacing of the BGA chip, set the glass cutting path, perform glass segmentation, collect the glass edge coordinates and judge the error value, filter image structure samples that meet the conditions, and generate a set of simulated package structure boundary images.

[0036] The glue spraying positioning frame construction module is used to execute S2: read the corner coordinates of the simulated packaging structure boundary image set, set the glue head aperture and dot matrix spacing of the glue dispensing device, execute the glue spraying operation and check the stability of the spraying action, extract the boundary pixels of the glue spraying graphic, and construct the boundary map of the glue spraying positioning frame.

[0037] The alignment and bonding stability judgment module is used to perform S3: based on the adhesive structure area in the boundary map of the adhesive positioning frame, set the alignment and bonding path and collect bonding images during the bonding process, extract bonding error and response value, judge the alignment and bonding stability, and obtain the bonding stability evaluation result.

[0038] The colloid filling path extraction module is used to perform S4: based on the bonding state corresponding to the bonding stability evaluation result, set the simulated adhesive injection path, activate the image acquisition module to record the colloid boundary image frame and extract the advancement contour pixels of each frame to generate a colloid filling advancement path map.

[0039] The filling effect image recording module is used to execute S5: based on the colloid trajectory in the colloid filling propagation path map, collect colloid curing images, void images and edge continuity images, and statistically analyze image boundaries, contact relationships and void morphology to obtain an overview record of the simulated filling effect image of the BGA chip.

[0040] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0041] By extracting parameters from packaging design drawings to construct a structural boundary image set, and combining image recognition and path judgment methods, the spraying, bonding, and colloid injection operations in a simulated environment are realized. The colloid propagation status and boundary changes are continuously recorded using image frame sequences. The trajectory chain list and boundary map constructed based on the images can be used to extract the colloid leading edge morphology, void distribution, and filling continuity. By replacing traditional slicing observation with full-process image recognition, material consumption and slicing errors are avoided, and experimental efficiency and controllability are improved. Dynamic observation of the filling process and quantitative evaluation of bonding stability are realized, enhancing the accuracy and adaptability of filling process verification and improving the limitations of existing technologies such as lagging evaluation results and strong operational dependence.

[0042] Using a transparent glass substrate instead of a BGA chip reduces cost.

[0043] The size, spacing, and arrangement of dots can be adjusted through the glue application process to simulate various BGA chips;

[0044] Furthermore, it allows direct observation of the adhesive filling process at the bottom of the BGA chip and the final filling effect, reducing the need for subsequent potting and dicing processes.

[0045] The use of cured adhesive instead of solder balls avoids the influence of residual flux during the soldering process on the glue filling process. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a schematic diagram of the steps of the present invention;

[0048] Figure 2 This is a flowchart illustrating the process of acquiring the boundary image set of the simulated packaging structure in this invention.

[0049] Figure 3 This is a flowchart of the process for obtaining the boundary map of the adhesive spraying positioning frame of the present invention;

[0050] Figure 4 A flowchart for obtaining stability assessment results of this invention is attached;

[0051] Figure 5 This is a flowchart of the process for obtaining the colloidal filling propulsion path map of the present invention;

[0052] Figure 6 This is a flowchart of the process for obtaining an overview record of the simulated fill effect image of the BGA chip in this invention;

[0053] Figure 7 This is a system module diagram of the present invention. Detailed Implementation

[0054] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0055] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0056] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0057] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0058] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0059] Please see Figure 1 This invention provides a test method for simulating the filling effect of adhesives on BGA chips, comprising the following steps:

[0060] S1: Obtain BGA chip packaging design drawings, identify the geometric parameters of length, width, and chip ball spacing required for BGA chip packaging simulation, set the displacement trajectory, starting coordinates, and cutting path instructions of the glass cutting device, execute the glass cutting operation, obtain the corner coordinate sequence of the four-sided cutting boundary, and filter image structure samples that meet the construction conditions by judging whether the glass edge straight line fitting error value is less than the set error tolerance, and generate a set of simulated packaging structure boundary images.

[0061] S2: Based on the corner coordinates in the simulated packaging structure boundary image set, configure the nozzle aperture and dot matrix spacing in the dispensing device. By judging whether the start and end coordinate change sequence in the dispensing displacement path is less than the nozzle return trajectory error threshold, verify the stability of the spraying action, extract the pixel set of the spraying graphic boundary line in the corner area of ​​the glass sheet, and construct the spraying positioning frame boundary map.

[0062] S3: Based on the glue spraying graphic boundary in the glue spraying positioning frame boundary map, set the contact path of the bonding device, the initial displacement direction of bonding and the contact surface position of the clamping platform, record the deformation of each bonding point in the glass sheet bonding action, identify the planar bonding error in the bonding process under the glue spraying structure support, calculate the bonding integrity index of each frame of the bonding process image, judge the alignment bonding stability, and obtain the bonding stability evaluation result.

[0063] S4: Based on the adhesion stability evaluation results, set the starting coordinates and propulsion speed of the simulated adhesive injection action, record the image frame sequence of the colloid boundary during the filling process, and extract the pixel distribution coordinates of the colloid propulsion leading edge contour in each frame image to establish a leading edge propulsion trajectory chain list for the image time series to obtain the colloid filling propulsion path map;

[0064] S5: Based on the trajectory information in the colloidal filling path map, after the colloidal injection is completed, acquire images of colloidal curing, void distribution, and edge filling continuity, construct an image package including the boundaries of different filling areas, the contact position relationship between the colloidal and the substrate, and the distribution pattern of residual voids, and generate an overview record of the simulated filling effect image of the BGA chip.

[0065] The simulated packaging structure boundary image set includes packaging boundary structure diagrams, corner coordinate distribution diagrams, and image sample selection results. The adhesive spraying positioning frame boundary map includes corner region adhesive distribution diagrams, adhesive spraying path trajectory diagrams, and adhesive spraying boundary line pixel diagrams. The bonding stability evaluation results include bonding integrity index diagrams, bonding error distribution diagrams, and bonding stability judgment diagrams. The adhesive filling advancement path map includes advancement path chain list, leading edge advancement trajectory diagrams, and image frame sequence boundary diagrams. The BGA chip simulated filling effect image overview record includes adhesive curing images, void distribution images, and edge filling continuity images.

[0066] Please see Figure 2 The specific steps of S1 are as follows:

[0067] S111: Obtain BGA chip package design drawings, identify the chip length parameters, width parameters and chip ball spacing values ​​marked in the drawings, extract the pixel spacing information of the graphic area of ​​the drawings, convert the extracted size parameters into units, combine them to generate the size configuration required for package simulation, and obtain the chip package geometric parameter set;

[0068] To obtain BGA chip packaging design drawings, the graphic information contained in the drawings must first undergo vector conversion of the image format. The lines and text information marked in the drawings are extracted into a computable data format. Then, by calling the primitive hierarchy structure in the drawings, the structural dimension markings around the chip boundary and ball array are identified sequentially to further determine the specific physical meaning represented by the chip's boundary contour. Primitives with actual size units are parsed, such as "length = 5.0mm", "width = 4.0mm", "pitch = 0.5mm", etc., and extracted and categorized through text content matching. Next, the pixel accuracy of the image in the drawings needs to be evaluated. Common design drawings have an image resolution of 600dpi. Converting the pixel distance between any two points in the image to physical distance requires multiplying by a unit pixel conversion factor of 0.0423mm / pixel, thus obtaining the actual geometric dimensions represented by the chip contour. For example, if the pixel distance between the start and end points of the identified primitive is 118 pixels, the actual length is calculated to be... Simultaneously, vector line segments are extracted from the spacing between each solder ball in the chip ball array in the drawing, the pixel distance between adjacent ball centers is measured, and then unit conversion is performed. If the spacing between adjacent balls is 12 pixels, then the actual distance is... Therefore, the chip sphere spacing is 0.5076mm. All the structural dimensions after analysis need to be uniformly packaged into structured data format and summarized in the form of a parameter set for use by the subsequent path planning module. The specific data is listed in the table below:

[0069] Table 1. Parameter Analysis Table for Packaging Design Drawings

[0070]

[0071] As shown in Table 1, the package geometry can be calculated from the relationship between primitives and pixels in the drawing, and all data must conform to the layer annotation conventions in the design drawing. For example, the commonly used standard for package ball pitch in BGA design is 0.5mm, and the corresponding extracted value is 0.5076mm. The relative deviation is no more than 2%, which is within the acceptable range. Through the above drawing information recognition, dimension extraction and conversion operations, the chip package geometry parameter set can be obtained.

[0072] S112: Based on the chip packaging geometric parameter set, set the starting coordinates of the displacement trajectory and the cutting path of the glass cutting device. After the glass cutting operation is performed, collect the corner coordinates of the four boundary cutting areas, construct the boundary contour curve data frame according to the order of the coordinates, and obtain the glass cutting corner coordinate sequence.

[0073] Based on the length, width, and ball spacing information in the chip package geometric parameter set, the initial starting coordinates of the glass cutting device must first be set. The principle is to ensure that the cutting trajectory covers all boundaries of the chip package area. The starting coordinates can be set to the lower left corner of the glass substrate (0, 0). Then, based on the chip package structure width of 4.0 mm and length of 5.0 mm, and combined with the cutting head displacement step setting of 0.01 mm, a cutting path coordinate sequence is constructed. Straight line cutting is performed in the horizontal and vertical directions respectively. The cutting path point sequences are set as [(0, 0) - (5.0, 0)], [(5.0, 0) - (5.0, 4.0)], [(5.0, 4.0) - (0, 4.0)], and [(0, 4.0)]. [0, 0], forming a rectangular closed loop path. Then, the cutting device operation command is executed, and the set path is sent to the device control module. After the glass cutting operation is completed, the corner area of ​​the glass sheet after cutting needs to be image acquired. After the image is acquired by the industrial camera, the corner extraction algorithm is used to identify the coordinates of each corner point of the glass sheet boundary. In the corner point identification operation, the abrupt change of image grayscale needs to be extracted as corner candidate points. The corner points on each boundary are processed by curve fitting according to the x / y axis change trend to extract the boundary inflection points that meet the conditions of continuity and angle change. Finally, eight key coordinate points on the glass boundary are obtained to form the boundary structure of the packaging area. The entire set of corner point coordinate data constitutes the glass cutting corner coordinate sequence.

[0074] S113: Based on the coordinate sequence of glass cutting edges and corners, calculate the minimum fitting residual value between each boundary fitting line and the actual edge and corner coordinate points, determine whether the residual is less than the preset edge fitting error tolerance, filter images that meet the conditions, extract the corresponding boundary contour information, and obtain a set of simulated packaging structure boundary images.

[0075] Based on the set of boundary points in the coordinate sequence of the glass cutting edges, a least-squares linear fitting operation is performed on each boundary line segment. Specifically, a fitting linear model is defined. And calculate each boundary point Vertical distance difference to the fitted line The least square residual formula is used. The residual values ​​of all points are accumulated to calculate the total fitting error value for each boundary. All residual values ​​are compared with the set edge fitting error tolerance, which is derived from glass cutting error experimental data and set to 0.12 mm. If the total fitting residual of a boundary line segment does not exceed the error tolerance, the image sample is determined to meet the structural recognition condition in terms of boundary extraction. For example, if the residual value of a boundary corner point after fitting is 0.089 mm, which is lower than the 0.12 mm threshold, it is considered a valid image sample. The image number corresponding to the image sample is recorded, and the corresponding boundary point sequence is format-encoded to form a combination of image label and structural data. The structural data of all image samples that meet the fitting error condition are merged into a dataset, and finally, a simulated encapsulation structure boundary image set is established.

[0076] Please see Figure 3 The specific steps of S2 are as follows:

[0077] S211: Based on the corner coordinates in the simulated packaging structure boundary image set, extract the coordinate values ​​of each corner point pair in the image, calculate the arrangement relationship in the image coordinate system, obtain the glue head diameter configuration parameters and dot matrix spacing setting values ​​of the dispensing device, map the corner coordinates according to the outward expansion distance of the glue head aperture, determine the initial dispensing position and array arrangement spacing, and generate the dispensing device displacement configuration parameter set.

[0078] Based on the corner coordinates of the simulated packaging structure boundary image set, each corner coordinate point in the image needs to be read and decoded point by point. A pixel coordinate transformation formula is then used to map its position to the working coordinate system of the dispensing device. Typically, the conversion factor between image pixel units and device coordinate units is 0.01 mm / pixel. In a standard 300 dpi image, corner coordinates such as (x=230, y=410) correspond to device coordinates of (2.3 mm, 4.1 mm). Next, the nozzle orifice diameter parameters in the dispensing device are acquired. For example, if the nozzle orifice diameter is 0.3 mm, the dispensing coverage is set by radially expanding outwards from this orifice. The radius of the dispensing area is 0.15mm. Based on the extracted corner coordinates, the initial dispensing position is constructed. Simultaneously, the dispensing array spacing is configured. The dot matrix spacing needs to be weighted according to the nozzle orifice diameter. In this embodiment, a spacing of 0.6mm is used to ensure that the dispensing points are evenly distributed in the structural corner area. If a 4×4 dot matrix is ​​formed at one corner, the array coverage width is 1.8mm, which can cover a typical BGA corner area. Then, the corner coordinates need to be mapped to the dot matrix center coordinates, and the offset of each point in the array is calculated sequentially to form a coordinate compensation list. Combined with the nozzle motion trajectory generation device configuration table, as shown in Table 2:

[0079] Table 2 Displacement Configuration Parameters for Dispensing Device

[0080]

[0081] As shown in Table 2, the relationship between the equipment coordinates and the nozzle orifice diameter is clearly defined for each point, and the motion start parameters are established accordingly. All points are derived and calculated from the corner coordinates. After configuration, the system generates a set of displacement configuration parameters for the dispensing device.

[0082] S212: Based on the displacement configuration parameter set of the dispensing device, collect all start and end coordinate sequences in the execution path of the dispensing device, calculate the residual difference between the displacement difference in each path segment and the nozzle return error benchmark value, determine whether the residual difference is less than the nozzle return trajectory error threshold, and if the set threshold requirement is met, record the current path sequence as a valid trajectory and obtain a set of stable dispensing trajectory paths.

[0083] Based on the displacement configuration parameter set of the dispensing device, the start and end coordinates of each set of points during the dispensing process need to be extracted along the entire path. The displacement variation amplitude is then calculated based on the actual movement distance of each segment within the path. The displacement is calculated using the Euclidean distance between two adjacent points. For example, the displacement from point P1 to P2 is... After recording all segment displacement changes in the sequence, a reference value for the nozzle return trajectory error needs to be introduced for comparison and judgment. This reference value is set to 0.05mm based on experimental data. This value comes from the ±3σ range of the average nozzle return error measurement value, that is, the average nozzle return offset measured in 50 no-load runs is 0.032mm, and the standard deviation is 0.006mm. Therefore, the error threshold is calculated as follows: If the displacement error of any segment in the actual path exceeds the threshold, it is considered to be unstable. Through actual judgment, if the displacement of a certain trajectory is 0.058mm, its deviation exceeds the limit, it is marked as an abnormal trajectory and removed. In the end, the displacement error of each segment in all the remaining segments meets the requirement of less than 0.05mm. The trajectory segments are retained by indexing and marking, and each valid segment in the path is encoded to form a set of stable dispensing trajectory paths.

[0084] S213: Based on the set of stable dispensing trajectory paths, extract the dispensing image information of the glass slide corner area, sample the gray-scale change line segments of pixels in the image boundary area, and construct a boundary curve index list according to the distribution density of pixels to establish a boundary map of the dispensing positioning frame.

[0085] Based on the set of stable dispensing trajectory paths, the image regions corresponding to the execution areas are scanned and processed. The dispensing image content of the corner regions is extracted. The image regions are cropped into square regions of 2mm×2mm using corner positioning points. Gray-scale detection is performed on the pixels in the regions. Gradient transformation is used to identify edge lines with abrupt gray-scale changes in the image. Under the condition of an image resolution of 600dpi, the pixel density per unit length is 23.62 pixels / mm. The detection of boundary lines requires a gray-scale difference greater than 30. If the gray-scale of a pixel in the detection region changes abruptly from 230 to 150, the line segment is marked as a boundary contour. Adjacent boundary pixel segments are sorted by length and direction as clustering criteria and merged into a continuous set of boundary curves. For example, if 86 collinear pixels are detected in a certain boundary segment, they are grouped into one category according to that direction. Finally, all boundary curve sequences are identified by number to form an index matrix. The boundary curve information of all corner regions is output in a structured manner, and each curve is sorted by direction, thereby establishing the boundary map of the dispensing positioning frame.

[0086] Please see Figure 4 The specific steps of S3 are as follows:

[0087] S311: Based on the glue spraying graphic boundary in the glue spraying positioning frame boundary map, extract the corner coordinate point sequence of the boundary curve, set the contact path line segment and the initial displacement direction of the bonding device according to the curvature distribution of the boundary points, and perform three-axis alignment operation in combination with the position parameters of the clamping platform plane reference surface to generate a bonding path configuration parameter set.

[0088] Based on the adhesive spraying graphic boundary in the boundary map of the adhesive spraying positioning frame, the first step is to extract the set of boundary coordinate points of the adhesive spraying structure outline in the map. The starting and ending points of each boundary curve are identified using a contour scanning method. Then, curve turning points are identified based on the direction of the boundary line segments and the curvature values ​​at corner changes. Corner points with curvature changes greater than 45 degrees are set as turning control points. The extracted boundary line segments are divided into multiple directional segments, each defined as an attachment path unit. During the contact path configuration process, the attachment device needs to determine its initial displacement direction based on this curve direction. Typically, an initial directional path is formed by arranging the outer contour of the adhesive spraying boundary counterclockwise. For example, the first displacement path segment is the starting coordinate (3.2). From the starting point (5.2, 2.5) to the endpoint (5.2, 2.5), the direction is positive X-axis. Then, the device control system sets the motion direction control signal according to the unit direction of the path, and records the slope of the straight line fitting between the starting point and the endpoint as the reference angle of the displacement direction of the clamping platform. Subsequently, the position of the clamping platform needs to be adjusted. The platform contact surface is set according to the reference plane height initially set by the device. For example, if the platform reference surface is set to the Z-axis 0 position, and the adhesive thickness is 0.1mm, the platform needs to be adjusted to the Z-axis -0.1mm position to contact the lower surface of the encapsulation glass. The entire set of path parameters is finally collected into the contact path coordinate set, the initial direction set and the platform displacement height set, and combined to generate the attachment path configuration parameter set.

[0089] S312: Based on the attachment path configuration parameter set, monitor the instantaneous displacement of each attachment point during the glass sheet attachment process and record the deformation. Compare the actual bonding height changes of each attachment point in the adhesive spraying structure support area, determine whether the surface height deviation value of the bonding area is greater than the set bonding error threshold, and obtain the planar bonding deviation matrix of the adhesive spraying support area.

[0090] Based on the parameter set configured for the bonding path, it is necessary to collect data on the deformation changes at each bonding point during the actual bonding process in real time. A laser displacement sensor is used to record the instantaneous displacement of the bonding points, and the compression deformation of each bonding point in the Z-axis direction is stored. If the initial bonding height is 0mm and the measured value after bonding is -0.08mm, then the deformation is 0.08mm. Ten points are recorded at each boundary, for a total of 40 bonding points. Then, the deformation data of all points in the glue-spraying area and the non-glue-spraying area during the bonding process are extracted and classified. The average deformation values ​​of the two types of areas are compared to determine whether the difference exceeds the bonding error threshold. The bonding error threshold is set to 0.03mm, which is derived from the statistical results of the mean residual of three times the glue-spraying elastic rebound. In the experiment, the mean residual was 0.01mm and the standard deviation was 0.006mm. Therefore… If the deformation measured in the glue-sprayed area is 0.08 mm and in the non-glue-sprayed area is 0.11 mm, with a difference of 0.03 mm, it is determined to be a critical state. This type of sample is marked as a deviation boundary sample, and its position index is recorded. The height difference of all points is uniformly stored in the deviation matrix, and finally the plane bonding deviation matrix of the glue-sprayed support area is obtained.

[0091] S313: Based on the planar bonding deviation matrix of the adhesive support area, collect the boundary bonding integrity information corresponding to each frame of the bonding process image sequence, calculate the contour overlap of the bonding area and the image boundary reconstruction rate, and compare the calculated values ​​with the stability judgment benchmark to establish the bonding stability evaluation result.

[0092] Based on the planar bonding deviation matrix of the adhesive support area, frame-by-frame acquisition and boundary bonding judgment of the image sequence during the bonding process are required. An industrial camera is used to acquire one frame every 0.2 seconds. The overlap between the bonding area and the adhesive boundary in the image is analyzed. First, the visible contour edges of the bonding area in the image are identified, and a set of boundary pixels is established. Then, pixel-level overlap matching is performed with the boundary curve in the adhesive graphic boundary map. The percentage of overlapping pixels in a unit area is calculated. For example, if the adhesive boundary has 650 pixels and the bonding area has 605 overlapping pixels, the overlap rate is... This value is used as the contour overlap index. Next, the image boundary reconstruction rate is calculated, defined as the ratio of the boundary closure degree in the attached image to the theoretical contour length. For example, if the theoretical boundary is 720 pixels and the closed contour in the image is 690 pixels, then the reconstruction rate is... If any indicator is lower than the evaluation benchmark threshold, it is judged as unstable attachment. The evaluation benchmark values ​​are set as contour overlap of not less than 0.90 and reconstruction rate of not less than 0.92. Based on the above calculation results, a stability scoring matrix is ​​constructed. After summarizing the stability scores of all frame images, they are sorted by time and output to finally establish the attachment stability evaluation results.

[0093] Please see Figure 5 The specific steps of S4 are as follows:

[0094] S411: Based on the adhesion stability assessment results, set the starting coordinates of the simulated adhesive injection, and combine the stroke length of the injection system's propulsion mechanism and the adhesive application pressure output rate to calculate the injection speed parameter value and the pressurization cycle configuration set, thereby generating the adhesive injection action control parameter set.

[0095] Based on the adhesion stability assessment results, it is necessary to first extract the set of coordinate points of regions in the assessment matrix whose stability scores are higher than the preset stability threshold. The stability threshold is set to 0.88, which is determined based on the average adhesion integrity score of 0.85 and the standard deviation of 0.01 in multiple adhesion samples. All coordinate regions with scores greater than 0.88 are marked as high-stability regions. The center point of this region is selected as the starting point coordinate for simulating the glue injection action. For example, the center point position in the image coordinate system is (165, 240). Combining the propulsion device stroke setting and glue injection pressure output characteristics, the propulsion speed setting value is calculated. The device stroke is 20mm, the step time is 2.5s, and the glue injection speed is derived as follows: Simultaneously, the injection pressure was set to 0.4 MPa, which is derived from 70% of the upper limit of the device's design pressure of 0.6 MPa. The injection parameters are as follows:

[0096] Table 3. Control Parameters for Injection Action

[0097]

[0098] As shown in Table 3, the propulsion speed, starting coordinates and injection pressure are all derived from the selected area in the stability assessment results. All parameters are combined to generate a set of control parameters for the colloid injection action.

[0099] S412: Read the set of control parameters for colloid injection action, record the change process of colloid boundary morphology in each frame of the image during colloid filling, number and sort the acquired image sequence according to the timestamp, perform contour detection operation on each frame of the image, extract the sequence of closed pixels of the boundary contour in the colloid advancement area, and obtain the set of pixel coordinates of the colloid leading edge contour.

[0100] The system reads the control parameter set for the colloid injection action, initiates the injection action, and simultaneously acquires image sequences. An industrial camera records the colloid filling process at a fixed frame rate of 0.1 seconds per frame. Images are named using a timestamp + sequence number format, such as Frame_001, Frame_002, etc. The total recording time is set to 2.5 seconds, and a total of 25 frames are acquired. The contour information of the colloid region is extracted from each frame. The contour recognition operation uses gray-level difference to locate the boundary line. The threshold for gray-level abrupt changes is set to ΔG≥35. If the pixel gray level drops sharply from 210 to 175, then ΔG=35 satisfies the boundary recognition condition. The coordinate set of boundary pixel points of the identified closed region is extracted, and the image frame number is recorded to form a one-to-one correspondence between image frames and boundary contours. For example, there are a total of 684 boundary pixels in the Frame_005 image. After extraction, they are stored in the index dictionary as entries. The contour point data of all frame images are aggregated according to the time series and uniformly formatted to finally obtain the pixel coordinate set of the colloid leading edge contour.

[0101] S413: Based on the pixel coordinate set of the colloid leading edge contour, the boundary points in each frame image are aggregated in time series to construct the contour trajectory chain of consecutive frames in the advancing direction, the centroid displacement vector of the contour coordinate between adjacent frames is identified, and all the motion trajectories of the leading edge points are connected in sequence to establish the colloid filling advancing path map.

[0102] Based on the pixel coordinate set of the colloid's leading edge contour, temporal registration is required for the boundary contour coordinate points extracted from each frame. First, the centroid position of the boundary points in each frame is calculated. The centroid calculation formula is the arithmetic mean of the coordinate points. For example, the centroid in Frame_001 is (164.3, 239.7), and in Frame_002 it is (165.1, 240.5). The displacement vector between the centroids of adjacent frames is used as the instantaneous path increment of the colloid's propulsion trajectory. The path increment is... The above operation is repeated for all 25 frames of images. All path increment vectors are connected according to the time series to form a complete propulsion trajectory linked list. Each node in the path linked list corresponds to an image frame number and its propulsion vector. After concatenating all nodes, a continuous propulsion trajectory index list is established, and the propulsion speed change trend is marked. If the average displacement of the first 10 frames in the trajectory is 0.9 pixels / frame and that of the last 15 frames is 1.4 pixels / frame, it indicates that the propulsion trend is in an accelerating state. After the trajectory linked list is paired with the time series, the output is a two-dimensional path map structure data, and finally a colloid filling propulsion path map is established.

[0103] Please see Figure 6 The specific steps of S5 are as follows:

[0104] S511: Based on the trajectory information in the colloid filling propulsion path map, extract the propulsion endpoint coordinates and boundary envelope distribution features, set the corresponding image acquisition area and activate the vision detection system. At the time point after the colloid injection is completed and solidified, sequentially acquire solidified surface images, void distribution images and edge filling continuity images to generate a colloid solidification state image sequence.

[0105] Based on the trajectory information in the colloid filling path map, the envelope coordinates of the colloid leading edge trajectory in the final frame image of the dispensing process are first extracted, and its circumscribed rectangle is defined as the curing image acquisition area. The acquisition delay is set to 15 minutes using the dispensing endpoint timestamp to ensure the colloid completes the physical curing process. An optical imaging system is used to acquire cured image data within the defined area. The image resolution is set to 800×600 pixels, and the frame rate is 1 frame / 0.5 seconds. Three image types are used: colloid curing surface image, internal void detection image, and edge continuity analysis image. The void image uses X-ray... Image acquisition was performed using X-ray penetration. Edge images were acquired using a side-view laser triangulation device to obtain three-dimensional cross-sectional data. The three types of images were named Image_Fix, Image_Cavity, and Image_Edge. Each image was assigned a separate sequence folder according to a fixed naming rule, and the capture time and image number were recorded. The image acquisition duration was controlled within 2 seconds, and 4 frames of data were acquired for each type of image. The image data were uniformly saved in grayscale BMP format with an 8-bit image matrix. After image quality inspection of the acquisition results, 2 blurry images were removed, and 10 valid images were retained to finally generate a sequence of images of the colloid solidification state.

[0106] S512: Based on the image sequence of the colloid curing state, identify the boundary line between the colloid expansion area and the substrate structure in the image, record the spatial position relationship between the outer edge pixel position of the colloid and the substrate reference point, and perform contour aggregation on the boundary curves of each filled area in the image to establish a dataset of the relationship between the filled boundary and the contact position.

[0107] Based on the image sequence of the colloid's cured state, boundary recognition is performed on the Image_Edge image group. Gradient grayscale edge detection is used to extract the outer contour of the colloid-filled region in the image. The longest side segment among the contour points is identified as the main axis direction of the filled region. After aligning the main axis with the image centerline, the shortest distance vector between the colloid and the substrate edge is detected to construct a spatial contact correspondence. If the colloid edge coordinates are (425, 310) and the substrate reference boundary point is (400, 300), then the contact vector length is... Next, the void regions appearing in the Image_Cavity image are marked with region connectivity, and the length and width information of their boundary rectangles are calculated. The distance relationship between the boundary expansion and the substrate connection point is identified. If the void is found to be less than 20 pixels from the boundary, it is marked as a weak point of the boundary. All boundary curves, boundary-substrate contact point pairs, and void boundary attributes identified in the above operations are combined in the form of triples {boundary number, contact position, void belonging surface}. The pixel range information in the image is added to each triple to form a complete structured index list. Finally, a dataset of the relationship between filling boundary and contact position is established.

[0108] S513: Based on the dataset of the relationship between the filling boundary and the contact position, aggregate the void distribution image, the edge continuous image and the solidified colloid boundary layer data to construct a structural image package including the location of colloid residual voids, the contour morphology of the filling area and the sealing features of the contact surface, and output an overview record of the simulated filling effect image of the BGA chip.

[0109] Based on the dataset of relationships between filled boundaries and contact positions, the Image_Fix, Image_Edge, and Image_Cavity images are layered and overlapped. First, the Image_Fix image is set as the main layer reference. The boundary layer and the hole layer are visually blended using transparency overlay with 40% and 60% transparency respectively. Image alignment is performed using a homogeneous matrix transformation based on the pixel coordinate reference point (0,0) to unify the matrix format of the three images to 800×600 pixels. Then, edge feature enhancement processing is applied to the images, increasing the contrast of edge lines to greater than 40 gray levels. A visual image is constructed after image fusion. Legendary labels are added to the image: red indicates void distribution, green indicates edge filling, and blue indicates substrate contact points, forming the main view of the image package. Then, each fused image is divided into regional image slicing operations. Each image is divided into 9 regions according to the outline of the filling area. Each region is exported as an independent image file and labeled with the region number and layer category. Finally, a complete image overview structure is generated. The output file names are Sim_BGA_Overview_0001 to Sim_BGA_Overview_0009, and they are all stored in the folder / BGA_Sim_Summary. Finally, an image overview record of the simulated filling effect of the BGA chip is generated.

[0110] Please see Figure 7 A testing system for simulating the filling effect of adhesives on BGA chips, comprising:

[0111] The package structure image generation module is used to execute S1: obtain the length, width and inter-sphere spacing of the BGA chip, set the glass cutting path, perform glass segmentation, collect the glass edge coordinates and judge the error value, filter image structure samples that meet the conditions, and generate a set of simulated package structure boundary images.

[0112] The glue spraying positioning frame construction module is used to execute S2: read the corner coordinates of the simulated packaging structure boundary image set, set the glue head aperture and dot matrix spacing of the glue dispensing device, execute the glue spraying operation and check the stability of the spraying action, extract the boundary pixels of the glue spraying graphic, and construct the boundary map of the glue spraying positioning frame.

[0113] The alignment and bonding stability judgment module is used to execute S3: based on the adhesive structure area in the adhesive positioning frame boundary map, set the alignment and bonding path and collect bonding images during the bonding process, extract bonding error and response value, judge the alignment and bonding stability, and obtain the bonding stability evaluation result.

[0114] The colloid filling path extraction module is used to execute S4: Based on the bonding state corresponding to the bonding stability evaluation result, set the simulated adhesive injection path, activate the image acquisition module to record the colloid boundary image frame and extract the advancement contour pixels of each frame to generate the colloid filling advancement path map.

[0115] The filling effect image recording module is used to execute S5: based on the colloidal trajectory in the colloidal filling propagation path map, it collects colloidal curing images, void images and edge continuity images, and statistically analyzes image boundaries, contact relationships and void morphology to obtain an overview record of the simulated filling effect image of the BGA chip.

[0116] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A test method for simulating the filling effect of adhesives on BGA chips, characterized in that, Includes the following steps: S1: Obtain the length, width and inter-sphere spacing of the BGA chip, set the glass cutting path, perform glass segmentation, collect the glass edge coordinates and judge the error value, filter image structure samples that meet the conditions, and generate a set of simulated packaging structure boundary images. S2: Read the corner coordinates of the simulated packaging structure boundary image set, set the nozzle diameter and dot matrix spacing of the dispensing device, perform the spraying operation and check the stability of the spraying action, extract the boundary pixels of the spraying graphic, and construct the boundary map of the spraying positioning frame. S3: Based on the adhesive spraying structure area in the boundary map of the adhesive spraying positioning frame, set the alignment and bonding path and collect bonding images during the bonding process, extract bonding error and response value, judge the alignment and bonding stability, and obtain the bonding stability evaluation result. S4: Based on the bonding state corresponding to the bonding stability evaluation result, set the simulated adhesive injection path, activate the image acquisition module to record the colloid boundary image frame and extract the advancement contour pixels of each frame to generate a colloid filling advancement path map. S5: Based on the colloidal trajectory in the colloidal filling propulsion path map, collect colloidal curing images, void images and edge continuity images, and statistically analyze image boundaries, contact relationships and void morphology to obtain an overview record of the simulated filling effect images of the BGA chip.

2. The experimental method for testing the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that: The image structure samples that meet the conditions specifically refer to samples whose glass edge straight line fitting error value is less than a set error tolerance.

3. The test method for the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that: During the process of verifying the stability of the spraying action, it is determined whether the sequence of changes in the start and end coordinates in the dispensing displacement path is less than the nozzle return trajectory error threshold. If the threshold is met, it indicates that the action is stable.

4. The test method for the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that: The simulated packaging structure boundary image set includes a packaging boundary structure map, a corner coordinate distribution map, and image sample screening results. The adhesive spraying positioning frame boundary map includes a corner region adhesive distribution map, an adhesive spraying path trajectory map, and an adhesive spraying boundary line pixel map. The bonding stability evaluation results include a bonding integrity index map, a bonding error distribution map, and a bonding stability judgment map. The adhesive filling advancement path map includes an advancement path list, a leading edge advancement trajectory map, and an image frame sequence boundary map. The BGA chip simulated filling effect image overview record includes an adhesive curing image, a void distribution image, and an edge filling continuity image.

5. The test method for the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that, The steps for obtaining the simulated packaging structure boundary image set are as follows: S111: Obtain BGA chip package design drawings, identify the chip length parameters, width parameters and chip ball spacing values ​​marked in the drawings, extract the pixel spacing information of the graphic area of ​​the drawings, convert the extracted size parameters into units, combine them to generate the size configuration required for package simulation, and obtain the chip package geometric parameter set; S112: Based on the chip package geometric parameter set, set the starting coordinates of the displacement trajectory and the cutting path of the glass cutting device. After the glass cutting operation is performed, collect the corner coordinates of the four boundary cutting areas, construct the boundary contour curve data frame according to the order of the coordinates, and obtain the glass cutting corner coordinate sequence. S113: Based on the coordinate sequence of the glass cutting edges and corners, calculate the minimum fitting residual value between each boundary fitting line and the actual edge and corner coordinate points, determine whether the residual is less than the preset edge fitting error tolerance, filter images that meet the conditions, extract the corresponding boundary contour information, and obtain a set of simulated packaging structure boundary images.

6. The test method for the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that, The steps for obtaining the boundary map of the adhesive spray positioning frame are as follows: S211: Based on the corner coordinates in the simulated packaging structure boundary image set, extract the coordinate values ​​of each corner point pair in the image, calculate the arrangement relationship in the image coordinate system, obtain the glue head diameter configuration parameters and dot matrix spacing setting values ​​of the dispensing device, map the corner coordinates according to the outward expansion distance of the glue head aperture, determine the initial dispensing position and array arrangement spacing, and generate a dispensing device displacement configuration parameter set. S212: Based on the displacement configuration parameter set of the dispensing device, collect all start and end coordinate sequences in the execution path of the dispensing device, calculate the residual difference between the displacement difference in each path segment and the nozzle return error benchmark value, determine whether the residual is less than the nozzle return trajectory error threshold, and if the set threshold requirement is met, record the current path sequence as a valid trajectory and obtain a stable dispensing trajectory path set. S213: Based on the set of stable dispensing trajectory paths, extract the dispensing image information of the glass slide corner area, sample the gray-scale change line segments of pixels in the image boundary area, and construct a boundary curve index list according to the distribution density of pixels to establish a boundary map of the dispensing positioning frame.

7. The experimental method for testing the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that, The steps for obtaining the attachment stability assessment results are as follows: S311: Based on the glue spraying graphic boundary in the glue spraying positioning frame boundary map, extract the corner coordinate point sequence of the boundary curve, set the contact path line segment and the initial displacement direction of the bonding device according to the curvature distribution of the boundary points, and perform a three-axis alignment operation in combination with the position parameters of the clamping platform plane reference surface to generate a bonding path configuration parameter set. S312: Based on the attachment path configuration parameter set, monitor the instantaneous displacement of each attachment point during the glass sheet attachment process and record the deformation, compare the actual bonding height change of each attachment point in the adhesive spraying structure support area, determine whether the surface height deviation value of the bonding area is greater than the set bonding error threshold, and obtain the planar bonding deviation matrix of the adhesive spraying support area. S313: Based on the planar bonding deviation matrix of the adhesive support area, collect the boundary bonding integrity information corresponding to each frame of the bonding process image sequence, calculate the contour overlap of the bonding area and the image boundary reconstruction rate, and compare the calculated values ​​with the stability judgment benchmark to establish the bonding stability evaluation result.

8. The test method for the filling effect of simulated adhesive on BGA chips according to claim 1, characterized in that, The steps for obtaining the colloidal filling propulsion path map are as follows: S411: Based on the adhesion stability evaluation results, set the starting coordinates of the simulated adhesive injection, and combine the stroke length of the injection system's propulsion mechanism with the adhesive application pressure output rate to calculate the injection speed parameter value and the pressurization cycle configuration set, thereby generating the adhesive injection action control parameter set. S412: Read the set of control parameters for the colloid injection action, record the change process of the colloid boundary morphology in each frame of the image during the colloid filling process, number and sort the acquired image sequence according to the timestamp, perform contour detection operation on each frame of the image, extract the sequence of closed pixels of the boundary contour in the colloid propulsion area, and obtain the set of pixel coordinates of the colloid leading edge contour. S413: Based on the pixel coordinate set of the colloid leading edge contour, the boundary points in each frame image are aggregated in time series to construct a contour trajectory chain of consecutive frames in the advancing direction, the centroid displacement vector of the contour coordinate between adjacent frames is identified, and all the movement trajectories of the leading edge points are connected in sequence to establish a colloid filling advancing path map.

9. The test method for the filling effect of simulated adhesive on BGA chips according to claim 8, characterized in that, The steps for obtaining the overall image record of the simulated fill effect of the BGA chip are as follows: S511: Based on the trajectory information in the colloid filling propulsion path map, extract the propulsion endpoint coordinates and boundary envelope distribution features, set the corresponding image acquisition area and activate the visual detection system, and sequentially acquire images of the solidified surface, void distribution, and edge filling continuity at the time point after the colloid injection is completed and solidified, generating a sequence of colloid solidification state images. S512: Based on the image sequence of the colloid curing state, identify the boundary line between the colloid expansion area and the substrate structure boundary in the image, record the spatial position relationship between the outer edge pixel position of the colloid and the substrate reference point, and perform contour aggregation on the boundary curves of each filled area in the image to establish a dataset of the relationship between the filled boundary and the contact position. S513: Based on the dataset of the relationship between the filling boundary and the contact position, aggregate the void distribution image, the edge continuous image and the solidified colloid boundary layer data, construct a structural image package including the location of the colloid residual voids, the contour shape of the filling area and the sealing features of the contact surface, and output an overview record of the simulated filling effect image of the BGA chip.

10. A testing system for simulating the filling effect of adhesives on BGA chips, characterized in that, The system is used to implement the experimental method for filling BGA chips using simulated adhesives as described in any one of claims 1-9, the system comprising: The package structure image generation module is used to execute S1: obtain the length, width and inter-sphere spacing of the BGA chip, set the glass cutting path, perform glass segmentation, collect the glass edge coordinates and judge the error value, filter image structure samples that meet the conditions, and generate a set of simulated package structure boundary images. The glue spraying positioning frame construction module is used to execute S2: read the corner coordinates of the simulated packaging structure boundary image set, set the glue head aperture and dot matrix spacing of the glue dispensing device, execute the glue spraying operation and check the stability of the spraying action, extract the boundary pixels of the glue spraying graphic, and construct the boundary map of the glue spraying positioning frame. The alignment and bonding stability judgment module is used to perform S3: based on the adhesive structure area in the boundary map of the adhesive positioning frame, set the alignment and bonding path and collect bonding images during the bonding process, extract bonding error and response value, judge the alignment and bonding stability, and obtain the bonding stability evaluation result. The colloid filling path extraction module is used to perform S4: based on the bonding state corresponding to the bonding stability evaluation result, set the simulated adhesive injection path, activate the image acquisition module to record the colloid boundary image frame and extract the advancement contour pixels of each frame to generate a colloid filling advancement path map. The filling effect image recording module is used to execute S5: based on the colloid trajectory in the colloid filling propagation path map, collect colloid curing images, void images and edge continuity images, and statistically analyze image boundaries, contact relationships and void morphology to obtain an overview record of the simulated filling effect image of the BGA chip.

Images