A PCB precision forming method based on composite compensation, a PCB, an apparatus and a device

Abstract

The application relates to a PCB precision forming method based on composite compensation, a PCB, a device and equipment, the method comprising the following steps: according to the comparison result of the actual coordinates and the theoretical coordinates of a target point, solving global transformation parameters of the whole plate, and applying the global transformation parameters to the whole forming equation; collecting the regional image of a functional area, extracting the reference features of the functional area after pretreatment of the collected image; converting the pixel coordinates of the extracted reference features of the functional area into actual poses under a workpiece coordinate system, and comparing the actual poses with the theoretical poses of the functional area after the preliminary compensation; calculating secondary compensation according to the local deviation and a preset tolerance band; performing preliminary forming path correction according to the preliminary compensation step, and performing secondary forming path correction according to the secondary compensation step. The application can eliminate the cumulative deviation in the machining process, strictly control the size error within the required tolerance range, and improve the machining precision of high-end PCB products.

Description

Technical Field

[0001] This invention relates to the field of PCB processing technology, and in particular to a PCB precision forming method, PCB, apparatus and equipment based on composite compensation. Background Technology

[0002] Currently, PCB manufacturing requires cutting and shaping equipment, such as milling machines, to obtain finished PCBs that meet dimensional requirements. During PCB production, dimensions such as the distance between functional areas and the cut edges affect the performance of the processing area. If dimensional deviations exceed tolerances, the product will be defective.

[0003] For example, in high-end electronic products such as 5G communications and high-speed optical modules, the distance from the chip bonding area (gold fingers) to the PCB molding edge is a critical dimension affecting signal integrity and soldering reliability, typically requiring control within a micrometer-level tolerance band (e.g., 0–25 μm). Currently, most common compensation methods only calculate and compensate for factors such as PCB expansion and contraction, and rotation, before cutting. The gold fingers themselves are not directly measured, while the truly critical dimension to be guaranteed is the "gold finger-molding edge." This results in existing compensation methods being unable to match the cumulative errors in actual processing, ultimately leading to yield rates that fail to meet the production requirements of high-end products. Summary of the Invention

[0004] This invention aims to solve the problem that existing PCB molding processes only perform one global compensation, which cannot guarantee the critical dimensional accuracy between functional areas and molding edges. It provides a PCB precision molding method, PCB, device, and equipment based on composite compensation. By combining global preliminary compensation with local secondary compensation of functional areas, the cumulative deviation during the processing is eliminated, and the dimensional error between functional areas and molding edges is strictly controlled within the required tolerance range, thereby improving the processing accuracy of high-end PCB products.

[0005] To achieve the objectives of this invention, the embodiments of this invention adopt the following technical solutions:

[0006] A PCB precision forming method based on composite compensation includes:

[0007] Preliminary compensation: Scan multiple target points on the PCB, calculate the global transformation parameters of the entire board based on the comparison between the actual coordinates and theoretical coordinates of the target points, and apply the global transformation parameters to the entire processing program to complete the preliminary compensation of the entire PCB board.

[0008] Functional area feature extraction involves acquiring regional images of the functional areas and preprocessing the acquired images to extract the baseline features of the functional areas.

[0009] Actual pose calculation involves converting the pixel coordinates of the extracted functional area reference features into the actual pose in the workpiece coordinate system, and comparing the actual pose with the theoretical pose of the functional area after the preliminary compensation to calculate the local deviation.

[0010] Secondary compensation is performed by calculating the secondary compensation amount based on the local deviation and the preset tolerance zone.

[0011] Molding path correction: a preliminary molding path correction is performed according to the preliminary compensation step, and a secondary molding path correction is performed according to the secondary compensation step.

[0012] The PCB is processed and shaped according to the secondary forming path.

[0013] In some embodiments, the local deviation includes the offset and rotation of the functional area reference feature. The offset is the component of the processing area's position after preliminary compensation relative to the theoretical position in the direction perpendicular to the forming edge; the rotation is the rotation angle of the processing area's reference line after preliminary compensation relative to the theoretical direction.

[0014] In some embodiments, in the functional area feature extraction step, the functional area is illuminated by a forward light source opposite to the functional area to obtain a clear image of the area's brightness and darkness contrast, and then the functional area's baseline features are extracted by an edge detection algorithm.

[0015] In some embodiments, the functional area reference features include the functional area edge, the geometric center or axis of symmetry of the functional area array, and the designated reference point of the functional area.

[0016] In some embodiments, when calculating the secondary compensation amount, a predicted compensation amount is added. The predicted compensation amount is the cumulative machining amount of the tool, and the predicted compensation amount is obtained based on the tool wear drift curve.

[0017] In some embodiments, a re-inspection step is performed after processing and forming. The re-inspection is set as no inspection, full inspection or sampling inspection according to quality requirements. During the re-inspection, the finished product image is collected, the measured data of the functional area reference features are measured, and it is determined whether the measured data falls within the tolerance zone.

[0018] In some embodiments, the collected measured data and predicted compensation amount are stored together in a historical database for statistical model maintenance; after the part number is changed, the tool wear drift curve is refitted and updated in batches using all the accumulated processing data of that part number.

[0019] A PCB is manufactured using the aforementioned PCB precision forming method based on composite compensation.

[0020] A PCB precision forming apparatus based on composite compensation includes:

[0021] The coordinate acquisition module is used to scan multiple target points on the PCB to obtain the actual coordinates of the target points, and then calculate the global transformation parameters of the entire board after comparing them with the theoretical coordinates of the target points.

[0022] The image acquisition module is used to acquire regional images of the PCB functional areas and extract the reference features of the functional areas after image preprocessing.

[0023] The deviation calculation module is used to convert the pixel coordinates of the extracted functional area reference features into the actual pose in the workpiece coordinate system, and compare them with the theoretical pose of the functional area after preliminary compensation to obtain the local deviation.

[0024] The compensation calculation module is used to calculate the secondary compensation amount based on the local deviation and the preset tolerance zone.

[0025] The path correction module is used to complete the initial forming path correction based on the global transformation parameters, and then complete the secondary forming path correction based on the secondary compensation amount.

[0026] The processing execution module is used to process and shape the PCB according to the modified secondary forming path.

[0027] A PCB processing device includes a slide table, a spindle, an image acquisition unit, an illumination unit, and a control unit. The image acquisition unit, illumination unit, and spindle are each electrically connected to the control unit. The slide table is used to support and fix the PCB to be processed. The spindle is used to mount forming tools and drive the tools to complete PCB cutting. The image acquisition unit is used to acquire the coordinates of PCB target points and images of functional areas. The illumination unit is used to provide suitable lighting conditions for image acquisition. The control unit is equipped with a PCB precision forming device based on composite compensation, which is used to perform compensation calculations and processing path correction, and output processing commands to control the spindle and slide table to complete the processing actions.

[0028] The present invention has the following main advantages:

[0029] 1. A two-stage compensation system is adopted, with coarse and fine division of labor and direct correlation with the gold fingers: The first stage, global compensation of the four target holes, solves the macroscopic alignment of the entire plate; the second stage, gold finger benchmark compensation, specifically eliminates the residual local deviations of the gold fingers after global compensation, and includes minor rotation correction. Existing technologies stop at the global / regional level; this solution superimposes precise gold finger compensation on top of this, eliminating the residuals of the gold fingers.

[0030] 2. Closed-loop critical dimension: The gold fingers are directly measured and can be re-inspected as needed after molding (no inspection / full inspection / sampling inspection), realizing a closed-loop mode in which the target, gold fingers and molding results are all inspected.

[0031] 3. More stable and higher precision: Visual feedback solves individual residual deviations of PCBs, and the statistical prediction curve feedforward solves the systemic drift caused by tool wear. Both roughing and finishing tool compensation ensure accurate machining measurements and early compensation, reducing per-piece fluctuations. The statistical prediction curve provides feedforward compensation based on the cumulative machining volume and is updated in batches with measured data after changing part numbers. The tool continuously self-calibrates throughout its entire life cycle.

[0032] 4. Compatible with existing conditions and requires minimal modification: The first stage uses the existing four-target hole global compensation process, and this solution only adds gold finger precision compensation afterward; it does not rely on special light sources, and the forming process uses conventional multiple processes of roughing and finishing, resulting in minimal modification for implementation.

[0033] Other advantages of the technical solution of the present invention will be described in specific embodiments. Attached Figure Description

[0034] 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.

[0035] Figure 1 This is a flowchart of a PCB precision forming method based on composite compensation provided in an embodiment of the present invention.

[0036] Figure 2 This is a schematic diagram of the PCB board provided in an embodiment of the present invention.

[0037] Figure 3 This is a partially enlarged schematic diagram of the gold finger provided in an embodiment of the present invention.

[0038] Figure 4 This is a schematic diagram of the molding edge correction provided in an embodiment of the present invention.

[0039] Figure 5 This is a schematic diagram of the tool wear drift curve provided in an embodiment of the present invention.

[0040] Figure 6 This is a flowchart illustrating a specific example in an embodiment of the present invention.

[0041] Figure 7 This is a schematic diagram of the PCB processing equipment provided in an embodiment of the present invention.

[0042] In the attached diagram: 1. Target point; 2. Functional area; 3. Slide stage; 4. Spindle; 5. Image acquisition unit; 6. Illumination unit; 7. Control unit. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0044] In this embodiment, "several" and "more than" refer to two or more. In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0045] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0046] This embodiment describes a PCB precision forming method, PCB, apparatus, and equipment based on composite compensation, using the processing of gold fingers on a PCB as an example. Currently, there are two main compensation methods in the industry for high-precision forming processes based on machine vision:

[0047] Method 1: Combining peripheral target hole scanning and preliminary compensation of the entire board program. Positioning target holes / markers are set on the four peripheral edges of the PCB (usually at the corners). Before processing, these four target holes are scanned to obtain the actual coordinates, compared with the theoretical coordinates, and the global transformation parameters for expansion, rotation, and offset of the entire board are calculated. This global transformation is uniformly applied to the entire processing program before cutting and shaping in one step. In other words, the four corner target holes represent the entire board state, and a single global affine transformation is used to approximately compensate for the alignment error of the entire board.

[0048] Method 2: Regional Compensation Based on Layout. For panel products, instead of using a single global transformation, the entire panel is divided into several sub-regions according to the layout structure, and each sub-region has its own local target. Before processing, the targets of each region are scanned, and the expansion / rotation / offset of each region is calculated. Local transformation compensation is then applied to the corresponding program segment of each region.

[0049] The commonality between the two methods described above is that compensation is based on the geometric alignment of the target / hole (whether global or regional), and is calculated once before cutting. The disadvantage of these methods is that:

[0050] 1. The reference is always the target; the gold finger itself is never directly measured. Whether it is preliminary compensation or zone compensation, the reference is the target hole / target, while the key dimension that really needs to be guaranteed is the "gold finger - forming edge". The actual position of the gold finger relative to the target (affected by inner layer alignment deviation and local deformation) is always a quantity that is not directly observed. For the key dimension, both existing forms are open-loop.

[0051] 2. Zonal compensation is still a "regional" approximation and cannot compensate for the microscopic local area where the gold finger is located. Although Method 2 refines to sub-regions, each region still uses a unified transformation approximation, assuming uniform deformation within the region. The microscopic position of the gold finger relative to its regional target may still have residual deviations (inner layer alignment, uneven local expansion and contraction, stress release at the cross-section), which cannot be eliminated by the regional transformation. The residuals often reach or even exceed the micrometer tolerance band. The farther the target is from the gold finger and the more uneven the deformation within the region, the larger the residuals.

[0052] 3. Only geometric alignment is performed, lacking feedforward for systematic and predictable errors. Both methods only compensate for the geometric alignment in the current cycle, without utilizing historical machining data. They cannot feed forward predictable and regular systematic errors such as the regular drift of tool wear with cumulative machining volume and batch-specific material springback, leading to accuracy drift at different stages of tool life and across different material batches, resulting in poor first-piece hit rate and batch stability.

[0053] <Example 1>

[0054] like Figures 1 to 5 As shown, this embodiment provides a PCB precision forming method based on composite compensation, including preliminary compensation, functional area feature extraction, actual pose calculation, secondary compensation, forming path correction, and processing steps. Specifically:

[0055] S100: Preliminary compensation step, i.e., global compensation, involves scanning multiple target points on the PCB. Based on the comparison between the actual and theoretical coordinates of the target points, the global transformation parameters of the entire PCB are calculated, and these global transformation parameters are applied to the entire machining program to complete the preliminary compensation of the entire PCB. The target points are the pre-set targets or markers on the PCB and have corresponding theoretical coordinates. Specifically, four positioning target points 1 at the corners of the PCB are scanned to obtain the actual coordinates of each target point 1. These actual coordinates are compared with the theoretical coordinates in the design file to calculate the global expansion, rotation, and offset transformation parameters of the entire PCB. These parameters are then applied to the entire PCB machining program to complete the preliminary compensation of the entire PCB. This global transformation is applied to the entire machining program (including the nominal coordinates of the forming contour and the gold finger index) to complete the coarse alignment of the entire program. The motion control program then drives the spindle with the camera to move above the nominal position (after global compensation) within the gold finger's field of view. This level of compensation solves the macroscopic alignment of the entire plate, but it is insufficient to guarantee the micron-level precision of the gold fingers in the local area, so a second level of fine compensation is required.

[0056] S200: Functional Area Feature Extraction. The system acquires an image of functional area 2 and preprocesses the acquired image to extract the functional area's baseline features. In this embodiment, functional area 2 is the gold finger area of ​​the PCB. A forward-facing light source illuminates the gold finger area, eliminating the need for special light sources or angle limitations. Existing processing equipment's camera light source can be used to obtain an image with clear contrast between light and dark areas. After preprocessing (denoising and ROI cropping) in the image processing program, a sub-pixel edge detection algorithm is used to extract the gold finger baseline features. These baseline features are flexibly selected according to actual detection standards. For example, the upper edge of the gold finger function, the geometric center / symmetry axis of the gold finger array, or the center of a specified baseline gold finger can be used. This embodiment uses the upper edge of the gold finger function as an example, aiming to control the distance from the upper edge of the gold finger function to the molding edge within the tolerance zone.

[0057] S300: Actual pose calculation. The pixel coordinates of the extracted functional area reference features are converted into the actual pose in the workpiece coordinate system. The actual pose is then compared with the theoretical pose of the functional area after preliminary compensation to calculate local deviations. In this embodiment, local deviations include the offset and rotation of the gold finger. The offset is the component of the gold finger position relative to the theoretical position (theoretical forming edge) after preliminary compensation, in the direction perpendicular to the forming edge. Δd = normal distance deviation = the component of the actual gold finger position relative to the theoretical position (after global compensation) in the direction perpendicular to the forming edge. The rotation is the rotation angle of the gold finger reference line relative to the theoretical direction after preliminary compensation. Δθ = angle deviation = the rotation angle of the gold finger reference line relative to the theoretical direction (after global compensation). Figure 4 The minute rotation angle of the pre-formed edge is measured by the second-level fine compensation, which measures the residual local deviation of the gold finger after global coarse compensation.

[0058] S400: Secondary compensation. Based on the calculated local deviation and the preset tolerance zone, the secondary compensation amount is calculated. The final fine compensation is synthesized for the forming edge corresponding to the gold finger, divided into distance and angle components.

[0059] Visual feedback component: C_vision = actual position − target position (target position is the median of the tolerance zone)

[0060] Distance compensation = C_vision

[0061] Final angle compensation = Δθ

[0062] As an improvement, a predicted compensation amount is incorporated into the calculation process. This predicted compensation amount is obtained from the current tool's cumulative machining amount through a pre-stored tool wear drift curve. The predicted compensation amount C_stat corresponding to the current tool is retrieved from the statistical model library. This model uses the tool's cumulative machining amount as the dominant variable, presenting an error drift curve (a fitted continuous function) that changes with the cumulative machining amount, with the value taken on the curve according to the current tool's cumulative machining amount. It can be overlaid with corrections for secondary factors such as material batch and temperature. C_stat = g(tool cumulative machining amount).

[0063] Final distance compensation = C_vision + C_stat.

[0064] S500: Forming path correction. First, the initial forming path correction is completed through global transformation parameters, and then the secondary forming path correction is completed based on the secondary compensation amount. Based on the toolpath after step S100, only the forming edge corresponding to the gold finger is subjected to the second-level fine correction: translated along the normal direction according to "final distance compensation" and rotated slightly according to "final angle compensation Δθ" to keep the forming edge at the target distance and parallel to the gold finger reference. The rest of the contour remains unchanged after global compensation.

[0065] S600: Controls the machining equipment to cut and shape the PCB according to the corrected secondary forming path. Forming consists of multiple processes, including roughing and finishing. The motion control program drives the machining equipment's spindle to cut according to the finely corrected toolpath. A composite compensation is applied in both the roughing and finishing processes, ensuring that roughing approaches the target and finishing falls within the tolerance zone. After machining, quality requirements can be set for non-inspection, full inspection, or sampling inspection of the finished product. During re-inspection, images of the finished product are collected, and the actual dimensions from the gold finger edge to the forming edge are measured to determine if the measured data falls within the tolerance zone. The collected measured data, along with the predicted compensation amount used, are stored in the historical database for maintaining the tool wear drift statistical model. The model is not updated in real-time for each piece; instead, after changing the machining part number, the accumulated machining data for that part number is used to batch update the statistical prediction curve for subsequent production self-calibration. In other words, all accumulated machining data for that part number is used to batch refit and update the tool wear drift curve, ensuring the accuracy of the predicted compensation amount.

[0066] This embodiment uses the gold finger itself as the measurement benchmark, without cutting boundaries, binding secondary forming, or relying on special side lighting. Moreover, the compensation amount is a composite compensation of visual feedback and statistical prediction, which can completely solve the open-loop error caused by the traditional method of using target holes to replace functional areas. It not only eliminates the alignment and deformation residuals between the target and the functional area, but also offsets the systematic drift caused by tool wear and material batches in advance through statistical feedforward. Finally, the key dimensions from the gold finger to the forming edge are stably controlled within the micron-level tolerance zone, which effectively improves the yield and batch processing stability of high-end PCB forming.

[0067] like Figure 6 As shown, the following description will further illustrate this embodiment in conjunction with specific application scenarios.

[0068] This specific example takes the gold finger molding of a high-speed optical module PCB as an example. The product requires that the "distance from the upper edge of the gold finger function to the molding edge" be controlled within the tolerance zone of 0 to 25 μm, and the target value is taken as the median of the tolerance zone, 12.5 μm.

[0069] The equipment used in this specific example includes: one precision molding machine equipped with vision system, with an XY slide repeatability of ±2μm; a high-resolution industrial camera (e.g., 5 megapixels, with a telecentric lens) fixedly mounted on one side of the spindle, with an imaging field of view of approximately 6mm × 4mm and a calibrated pixel equivalent of approximately 4μm / px; conventional ring-shaped white light illumination; and a control unit running the target hole global compensation program, the gold finger recognition program, the composite compensation calculation program, and the motion control program. Hand-eye calibration of the camera and workpiece coordinate systems was completed during system deployment. Hand-eye calibration refers to establishing the transformation relationship between camera pixel coordinates and PCB workpiece world coordinates, enabling the gold finger reference feature pixel coordinates identified by the camera to be converted into actual physical coordinates in the workpiece coordinate system, providing a basis for subsequent deviation calculations.

[0070] Preparations before processing in this specific example: Positioning target points are set at the four corners of the PCB of this model in advance. The theoretical coordinates of each target point, the theoretical coordinates of the upper edge of the gold finger function and the theoretical path of the forming edge are recorded in the design file. The initial tool wear drift curve under this processing condition is pre-stored in the statistical model library.

[0071] S1: Global coarse compensation and coarse positioning of the four target holes. After the PCB substrate is loaded and clamped by the vacuum adsorption stage of the processing equipment, the motion control program drives the spindle of the processing equipment to move the camera sequentially to the theoretical coordinates of the four peripheral target holes of the PCB to collect images. The center positioning algorithm in the gold finger recognition program calculates the actual pixel coordinates of the four target holes, which are converted into actual coordinates (x_i, y_i) in the machine tool coordinate system through hand-eye calibration, i=1~4. The target hole global compensation program performs least squares fitting between the four actual coordinates and the theoretical coordinates of the drawing to calculate the global transformation parameters of the entire board: the expansion / contraction coefficient (s_x, s_y), the rotation angle φ, and the offset (Δx, Δy). For example, the solutions are s_x=1.0008, s_y=0.9994, φ=0.015°, Δx=8μm, Δy=−5μm, indicating that the entire board has slight expansion / contraction, slight rotation, and translation relative to the theory. This global transformation is applied to the entire machining process—including the forming contour path and the nominal coordinates of the gold finger index—to achieve macroscopic alignment of the entire pattern. Subsequently, the motion control program, based on the compensated coordinates, drives the camera to move above the nominal position of the gold finger's field of view, completing coarse positioning. This level of compensation eliminates the macroscopic alignment error of the entire pattern, but because the target hole is far from the gold finger, and there are local expansion and contraction imbalances and inner layer alignment deviations within the pattern, micron-level local deviations still remain at the gold finger, requiring a second level of fine compensation.

[0072] S2: Image Acquisition and Subpixel Feature Extraction for the Gold Finger. Under ring-shaped white light illumination, the camera acquires images of the gold finger region. The image processing program first performs ROI cropping (locking the gold finger array region) and noise reduction filtering, and then uses a subpixel edge detection algorithm (such as subpixel localization based on gray-level centroid or polynomial fitting) to extract baseline features. In this embodiment, the "functional upper edge of the gold finger" is taken as the baseline feature according to the customer's detection standard, that is, a straight line is fitted to the edge points of all finger tips to obtain the gold finger baseline line.

[0073] S3: Actual pose calculation. The pixel coordinates of feature points on the extracted gold finger baseline are converted into the actual pose (position and orientation of the baseline) in the workpiece coordinate system using hand-eye calibration. This is compared with the "theoretical pose of the gold finger after global compensation" to calculate the two remaining deviations after global compensation:

[0074] Normal distance deviation Δd: The component of the actual reference line of the gold finger relative to its theoretical position in the direction perpendicular to the molding edge. In specific calculations, the distances from several theoretical points to the gold finger reference line can be selected, and the average value of these distances can be calculated. As those skilled in the art know, different values ​​can be used for Δd depending on the specific application scenario and requirements. In this PCB, the current actual position of the gold finger is measured to be 7μm off from its theoretical position in the normal direction, i.e., Δd = 7μm.

[0075] Angular deviation Δθ: The small rotation angle of the gold finger baseline relative to the theoretical direction. The fitting result for this PCB is Δθ = 0.02°.

[0076] S4: Statistical prediction compensation amount retrieval. The composite compensation calculation program reads the cumulative machining amount of the current tool (e.g., the finishing tool has accumulated 50mm of machining), and takes the value from the statistical prediction drift curve g corresponding to the tool to obtain the predicted compensation amount caused by systematic factors such as tool wear, obtaining C_stat=4μm (reflecting that the tool has a certain amount of wear and the cutting edge is systematically deviated inward by 4μm relative to the command position, which needs to be fed forward to compensate). This curve is fitted by historical measured errors and can be superimposed with material batch and temperature correction terms.

[0077] S5: Composite compensation amount synthesis, synthesizing the final fine compensation for the forming edge corresponding to the gold finger, divided into two components: distance and angle.

[0078] Visual feedback component C_vision = Actual position − Target position = Δd − Target value = 7μm − 12.5μm = −5.5μm (requires cutting back 5.5μm towards the gold finger direction)

[0079] Final distance compensation = C_vision + C_stat = −5.5μm + 4μm = −1.5μm

[0080] Final angle compensation = Δθ = 0.02°

[0081] Here, "actual distance - target distance" is used as the visual feedback component, with the positive and negative signs indicating the cutting direction. The statistical prediction C_stat is superimposed in the same direction to obtain the actual distance compensation of -1.5μm and angle compensation of 0.02° required for this edge of the PCB. The actual software can further embed tool radius compensation and material springback coefficient to form a more accurate process model.

[0082] S6: Single-sided translation and micro-rotation fine-tuning of the toolpath. Based on the toolpath after the first-level global compensation, the motion control program only performs a second-level fine-tuning on the forming edge corresponding to the gold finger: the edge is translated along the normal direction by a final distance compensation of −1.5μm, and rotated slightly by a final angle compensation of 0.02°, so that the forming edge maintains a target distance of 12.5μm from the gold finger baseline and remains parallel. The remaining forming contours remain unchanged after global compensation, thus accurately applying compensation to the local area where the critical dimensions are located.

[0083] S7: Multi-stage forming process with roughing and finishing (both roughing and finishing compensation). The forming of this product consists of two stages: roughing and finishing. The motion control program first uses a roughing cutter to cut along a modified toolpath, removing most of the excess material and bringing the edge close to the target. Then, it switches to a finishing cutter and uses the same composite compensation toolpath to complete the final forming. The composite compensation is applied in both the roughing and finishing stages, ensuring that the roughing process approaches the target and the finishing process ultimately brings the "distance from the upper edge of the gold finger function to the forming edge" into the 0-25μm tolerance zone, approaching the 12.5μm median value.

[0084] S8: Optional re-inspection and data feedback. This PCB is set to random sampling inspection (re-inspection can be set to no inspection, full inspection, or random sampling inspection according to quality requirements). During re-inspection, the camera re-captures the finished product image, measures the actual distance and angle from the upper edge of the gold finger function to the final formed edge, and confirms that the actual distance falls within 0-25μm (this PCB re-inspection yielded 12.0μm, which is qualified). The collected measured data, along with the predicted compensation amount of this PCB, are stored in the historical database for statistical model maintenance; the model is not updated in real time for each piece, but after changing the part number, all the accumulated processing data of that part number are used to batch refit and update the statistical prediction drift curve g for subsequent production self-calibration, realizing the feedforward accuracy maintenance throughout the tool's entire life cycle.

[0085] <Example 2>

[0086] In this embodiment, the parts that are the same as in Embodiment 1 are given the same reference numerals, and the same text descriptions are omitted.

[0087] Compared to Embodiment 1, this embodiment provides a PCB manufactured using the aforementioned composite compensation-based precision PCB forming method. The PCB has gold fingers and target points. After forming, the distance from the edge of the gold fingers to the corresponding forming edge stably falls within the required tolerance range, and the dimensional accuracy and consistency meet the assembly and use requirements of high-end precision electronic components.

[0088] <Example 3>

[0089] In this embodiment, the parts that are the same as in Embodiments 1 and 2 are given the same reference numerals, and the same text descriptions are omitted.

[0090] This embodiment provides a PCB precision forming device based on composite compensation, including: a coordinate acquisition module for scanning multiple target points on the PCB to obtain the actual coordinates of the target points, and calculating the global transformation parameters of the entire board after comparing them with the theoretical coordinates of the target points; an image acquisition module for acquiring regional images of the functional areas of the PCB, extracting the reference features of the functional areas after image preprocessing; a deviation calculation module for converting the pixel coordinates of the extracted reference features of the functional areas into the actual pose in the workpiece coordinate system, and comparing them with the theoretical pose of the functional areas after preliminary compensation to obtain the local deviation; a compensation calculation module for calculating the secondary compensation amount based on the local deviation and the preset tolerance zone; a path correction module for completing the preliminary forming path correction based on the global transformation parameters, and then completing the secondary forming path correction based on the secondary compensation amount; and a processing execution module for processing and forming the PCB according to the corrected secondary forming path.

[0091] The PCB precision forming device based on composite compensation in this embodiment is used in the PCB precision forming method based on composite compensation described in Embodiment 1. Each module works together to complete two-stage compensation processing, which can stably output PCB finished products with dimensional accuracy that meet the requirements. It is compatible with the transformation and upgrading of existing PCB precision forming equipment and can improve processing accuracy without large-scale hardware replacement.

[0092] <Example 4>

[0093] In this embodiment, the parts that are the same as in Embodiments 1 to 3 are given the same reference numerals, and the same text descriptions are omitted.

[0094] like Figure 7 As shown, this embodiment provides a PCB processing equipment including a slide table 3, a spindle 4, an image acquisition unit 5, an illumination unit 6, and a control unit 7. The image acquisition unit 5 is fixedly installed on one side of the spindle and moves with it. The image acquisition unit 5 can be a camera. The control unit 7 stores a computer program that can execute the PCB precision forming method based on composite compensation, including a target hole global compensation program, a gold finger recognition program, a composite compensation calculation program, and a motion control program. It can complete target hole positioning, image acquisition, deviation calculation, composite compensation calculation, and path correction according to the method flow of Embodiment 1, and finally drive the spindle and slide table to complete the cutting process. The equipment as a whole adopts the main hardware architecture of existing precision forming machines (such as PCB routers). Only the corresponding image acquisition and control programs need to be added to complete the modification. The modification cost is low, and it is suitable for high-end precision gold finger PCB processing production from small to large batches. It can stably meet the stringent dimensional tolerance requirements, reduce the product defect rate, and adapt to the PCB precision forming processing needs of fields such as high-speed optical modules and high-end communication equipment.

[0095] In the above embodiments one to four, during the working process, depending on the different working environments, some of the technical implementation methods of embodiments one to four can be combined or replaced.

[0096] The technical principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that these descriptions are merely for explaining the principles of the present invention and should not be construed as limiting the scope of protection of the present invention in any way. Based on this explanation, those skilled in the art can conceive of other specific embodiments or equivalent substitutions of the present invention without creative effort, and all such embodiments will fall within the scope of protection of the present invention.

Claims

1. A PCB precision forming method based on composite compensation, characterized in that, include: Preliminary compensation: Scan multiple target points on the PCB, calculate the global transformation parameters of the entire board based on the comparison between the actual coordinates and theoretical coordinates of the target points, and apply the global transformation parameters to the entire processing program to complete the preliminary compensation of the entire PCB board. Functional area feature extraction involves acquiring regional images of the functional areas, preprocessing the acquired images, and then extracting the baseline features of the functional areas. Actual pose calculation involves converting the pixel coordinates of the extracted functional area reference features into the actual pose in the workpiece coordinate system, and comparing the actual pose with the theoretical pose of the functional area after the preliminary compensation to calculate the local deviation. Secondary compensation is performed by calculating the secondary compensation amount based on the local deviation and the preset tolerance zone. Molding path correction: a preliminary molding path correction is performed according to the preliminary compensation step, and a secondary molding path correction is performed according to the secondary compensation step. The PCB is processed and shaped according to the secondary forming path.

2. The PCB precision forming method based on composite compensation according to claim 1, characterized in that, The local deviation includes the offset and rotation of the functional area reference features. The offset is the component of the processing area's position after preliminary compensation relative to the theoretical position in the direction perpendicular to the forming edge; the rotation is the rotation angle of the processing area's reference line after preliminary compensation relative to the theoretical direction.

3. The PCB precision forming method based on composite compensation according to claim 1, characterized in that, In the functional area feature extraction step, the functional area is illuminated by a positive light source opposite to the functional area to obtain a clear image of the area's light and dark contrast, and then the baseline features of the functional area are extracted by an edge detection algorithm.

4. The PCB precision forming method based on composite compensation according to claim 3, characterized in that, The functional area reference features include the functional area edge, the geometric center or axis of symmetry of the functional area array, and the designated reference point of the functional area.

5. The PCB precision forming method based on composite compensation according to claim 1, characterized in that, When calculating the secondary compensation amount, a predicted compensation amount is added. The predicted compensation amount is the cumulative machining amount of the tool, and the predicted compensation amount is obtained based on the tool wear drift curve.

6. The PCB precision forming method based on composite compensation according to claim 5, characterized in that, After processing and shaping, a re-inspection step is carried out. The re-inspection is set as no inspection, full inspection or sampling inspection according to the quality requirements. During the re-inspection, the finished product image is collected, the measured data of the functional area reference characteristics are measured, and it is determined whether the measured data falls within the tolerance zone.

7. The PCB precision forming method based on composite compensation according to claim 6, characterized in that, The collected measured data and predicted compensation amounts are stored together in the historical database for statistical model maintenance; after the part number is changed, the tool wear drift curve is refitted and updated in batches using all the accumulated machining data of that part number.

8. A PCB, characterized in that, It is manufactured using the PCB precision forming method based on composite compensation as described in any one of claims 1 to 7.

9. A PCB precision forming device based on composite compensation, characterized in that, include: The coordinate acquisition module is used to scan multiple target points on the PCB to obtain the actual coordinates of the target points, and then calculate the global transformation parameters of the entire board after comparing them with the theoretical coordinates of the target points. The image acquisition module is used to acquire regional images of the PCB functional areas and extract the reference features of the functional areas after image preprocessing. The deviation calculation module is used to convert the pixel coordinates of the extracted functional area reference features into the actual pose in the workpiece coordinate system, and compare them with the theoretical pose of the functional area after preliminary compensation to obtain the local deviation. The compensation calculation module is used to calculate the secondary compensation amount based on local deviations and preset tolerance zones. The path correction module is used to complete the initial forming path correction based on the global transformation parameters, and then complete the secondary forming path correction based on the secondary compensation amount. The processing execution module is used to process and shape the PCB according to the modified secondary forming path.

10. A PCB processing equipment, characterized in that, The device includes a slide table, a spindle, an image acquisition unit, an illumination unit, and a control unit. The image acquisition unit, illumination unit, and spindle are electrically connected to the control unit. The slide table is used to support and fix the PCB to be processed. The spindle is used to install forming tools and drive the tools to complete PCB cutting. The image acquisition unit is used to acquire the coordinates of PCB target points and images of functional areas. The illumination unit is used to provide suitable lighting conditions for image acquisition. The control unit is equipped with the PCB precision forming device based on composite compensation as described in claim 9, which is used to perform compensation calculations and processing path corrections, and output processing commands to control the spindle and slide table to complete processing actions.

Images