PCB solder mask screen printing automatic pinning method based on machine vision and three-axis robot

Abstract

The present application relates to the field of printed circuit board manufacturing, and discloses a PCB solder mask screen printing automatic pin arrangement method based on machine vision and three-axis manipulator, comprising the following steps: placing a reference PCB on a uniform hole array substrate for teaching; a machine vision module collects images and extracts feature points and hole position data to be avoided; an affine transformation model is used to calculate the rotation angle and the translation amount, and the theoretical coordinates are mapped to the substrate coordinate system; based on the projection area envelope and the Euclidean distance algorithm, the reference holes that meet the safety threshold are selected as the layout points; a three-axis manipulator automatically grabs the support pin according to the planned path and accurately inserts it into the target hole position. The present application replaces manual experience with visual perception and mathematical modeling, realizes flexible automatic pin arrangement under the premise of avoiding through holes, significantly improves the operation efficiency and support stability, and solves the technical problems of time-consuming pin arrangement and easy surface damage in traditional processes.

Description

Technical Field

[0001] This invention relates to the field of printed circuit board manufacturing technology, specifically to an automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot. Background Technology

[0002] In the solder mask screen printing process of printed circuit boards (PCBs), to ensure printing quality and prevent the PCB from sinking or deforming under the pressure of the squeegee, a bed of nails consisting of numerous support nails is usually built under the screen printing machine's worktable. The layout of these support nails has strict requirements; they must precisely avoid key features on the PCB such as vias and through-holes to prevent quality defects such as solder mask ink clogging or board surface damage.

[0003] Currently, the industry mainly constructs the aforementioned support platforms through manual pin placement or the use of integrated array-type automated pin systems. However, existing pin placement methods have significant limitations in practical applications. Traditional manual pin placement relies heavily on operators manually placing and adjusting pins according to processing data such as Gerber. This work mode is not only labor-intensive and inefficient, making it difficult to adapt to the demands of modern fast-paced production, but it is also highly susceptible to pin placement deviations due to human error or fatigue, leading to serious quality problems.

[0004] Meanwhile, although some automated ejector pin systems integrating cylinder or motor arrays have emerged on the market, capable of controlling specific positions to eject pins and form support surfaces based on PCB data, these systems are typically extremely complex, leading to high manufacturing costs and difficult maintenance. More importantly, due to the physical size limitations of the cylinder or motor drive units, the minimum spacing of the ejector pin array in these systems is usually large (often greater than or equal to 30mm), making it difficult to meet the spatial resolution requirements for high-precision support. Especially when dealing with thin PCBs less than 1.4mm thick or small-diameter PCBs with densely distributed holes, the sparse support points cannot provide sufficient support density, easily causing localized collapse or physical damage to the PCB during printing, significantly limiting its applicability. Therefore, there is an urgent need to develop an automated pin placement solution that is cost-effective, adaptable, highly precise, and can effectively solve the support challenges of PCBs of various thicknesses. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an automated pin placement method for PCB solder mask screen printing based on machine vision and a three-axis robot, overcoming the inefficiencies and poor precision of existing manual pin placement methods, as well as the high cost and poor adaptability to thin boards and high-density perforated boards of automated pin systems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of the present invention provides an automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robotic arm. This method employs an automatic pin placement system for PCB solder mask silkscreen printing, the system comprising a uniform aperture array substrate, a three-axis robotic arm, a pin feeding module, a pin placement end effector, a machine vision module, and a control system; the uniform aperture array substrate is fabricated with matrix-distributed reference holes. The method includes the following steps: Step S1: Place the reference PCB on the uniform aperture array substrate, obtain the processing data of the reference PCB, and use the machine vision module to collect image information of the reference PCB on the uniform aperture array substrate; Step S2: The control system processes the image information, identifies the actual hole position features of the reference PCB, maps the coordinate system where the image information is located to the coordinate system of the uniform hole array substrate, and generates an avoidance hole position map containing the coordinate information of the avoidable area by combining the processing data. Step S3: Based on the avoidance hole location map and the outline information of the reference PCB, the control system performs pin placement planning in the coordinate system of the uniform hole array substrate, selects target reference holes from the reference holes of the uniform hole array substrate, and forms a support pin placement matrix. Step S4: Remove the reference PCB from the uniform aperture array substrate to expose the reference aperture; Step S5: The control system generates control commands based on the support nail array, controls the three-axis robot to drive the nail-laying end effector to grab the support nails from the nail supply module, and inserts the support nails into each target reference hole of the uniform hole array substrate in sequence.

[0007] Furthermore, this invention employs image registration and coordinate compensation techniques to achieve high-precision obstacle avoidance. In step S2, the control system parses the processing data to extract the theoretical outline and theoretical hole position information of the reference PCB; simultaneously, it performs image recognition on the image information to obtain the actual position information of the vias and plug-in holes on the reference PCB. The positional deviation of the reference PCB relative to the uniform hole array substrate is calculated using an image registration algorithm.

[0008] Specifically, the principle of coordinate compensation using positional deviation is as follows: Let the theoretical coordinate point in the reference PCB manufacturing data be... The actual image coordinates of the corresponding feature points acquired by machine vision are: The system establishes a configuration that includes rotational deviations. Translational deviation The coordinate transformation model. The coordinate compensation relationship is expressed as: The control system uses the position deviation to perform coordinate compensation on the theoretical hole position information, corrects the theoretical data to the current physical placement state, and fuses the compensated information with the actual position information to generate an accurate avoidance hole position map.

[0009] Furthermore, the present invention specifies in detail the pin placement planning rules in step S3. The system pre-establishes a coordinate database of all reference holes on the uniform hole array substrate, traverses this database, and selects reference holes that simultaneously meet the following conditions as the target reference holes: Contour projection condition: The coordinates of the reference hole are located within the projection range of the reference PCB contour on the substrate; Avoidance conditions: The coordinates of the reference hole are not located within the coordinate range of the avoidance area in the avoidance hole location diagram, ensuring that the support pin avoids the hole on the PCB.

[0010] As a preferred embodiment, in order to quantify the above-mentioned avoidance conditions, the system calculates the coordinates of the reference hole center. With any edge of the hole to be avoided in the hole location diagram Euclidean distance between : The system determines when When the value exceeds the preset safety threshold, the reference hole meets the avoidance condition.

[0011] Furthermore, the pin placement plan also includes an optimization strategy. From the selected set of reference holes that meet the above conditions, a secondary selection is performed according to the principle of uniform distribution or the principle of maximizing the number of support points to determine the final support pin placement array, ensuring uniform stress on the PCB.

[0012] Furthermore, the present invention clarifies the image acquisition method. The three-axis robotic arm is controlled to move the machine vision module mounted thereon to above the reference PCB for scanning and imaging, or an independently fixed industrial camera is controlled to perform a global image capture of the reference PCB on the uniform aperture array substrate.

[0013] Furthermore, the present invention also covers subsequent production and application steps. After step S5 is completed, the PCB to be processed, which is the same model as the reference PCB, is placed on the pre-laid support pin array, and the placement position of the PCB to be processed on the uniform aperture array substrate is kept consistent with the placement position of the reference PCB in step S1, and subsequent solder mask printing is performed.

[0014] A second aspect of the present invention provides an automatic pin-laying system for PCB solder mask silkscreen printing, the system comprising: A uniform aperture array substrate is fixedly set and has a matrix of uniformly distributed reference holes on its surface, preferably through holes or blind holes arranged longitudinally and transversely at a preset spacing of 5 mm to 10 mm. A three-axis robotic arm is positioned above the uniform aperture array substrate and has X-axis, Y-axis and Z-axis degrees of freedom of motion. The nail supply module is used to store and output support nails one by one; The nail-holding end effector, installed at the Z-axis end of the three-axis manipulator, is used to grasp the support nail and is selected from any one of pneumatic grippers, electromagnetic chucks, or mechanical chucks. The machine vision module is used to acquire image information of the PCB. The control system is electrically connected to the three-axis robot, the nail feeding module, the nail-laying end effector, and the machine vision module.

[0015] The control system is equipped with a processing unit, which is configured to perform the data processing and instruction generation steps of the control system in the above method, specifically including: parsing the processing data of the reference PCB; processing the image of the reference PCB acquired by the machine vision module to generate an avoidance hole location map; planning and generating a support pin placement matrix in the substrate coordinate system; and generating control instructions based on the support pin placement matrix to drive the three-axis robot to perform pin placement actions.

[0016] This invention provides an automated pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robotic arm. It offers the following advantages: 1. Breaking through traditional thickness limitations, this invention achieves zero-deformation support for thin, high-density boards. Utilizing a high-density, uniform aperture array substrate, combined with the sub-pixel-level compensation capabilities of machine vision, the system can safely place pins in narrow areas extremely close to the holes to be avoided. This "high-frequency, densely distributed" support characteristic effectively solves the problem of sparse support points in traditional processes causing "trampoline effect" or concave deformation of thin boards under screen printing squeegee pressure, providing unprecedented flatness assurance for high-density interconnects and flexible circuit boards.

[0017] 2. A simplified hardware architecture replaces a complex, fully-driven system, significantly reducing marginal costs. Unlike existing technologies that rely on thousands of independent cylinders or motors for drive in a "fully arrayed active ejector pin system," this invention innovatively employs a simplified architecture of a "passive uniform aperture array substrate + a single three-axis robotic arm." This design eliminates the expensive and difficult-to-maintain complex drive matrix, achieving equal or even better support performance through software planning alone. This not only significantly reduces the manufacturing cost and maintenance difficulty of the equipment but also eliminates the potential risk of PCB damage due to individual cylinder failures, greatly improving the long-term operational reliability of the system.

[0018] 3. Visual closed-loop and dynamic compensation mechanism to establish micron-level pin placement accuracy. The system does not mechanically execute preset coordinates, but introduces a "perception-correction" closed loop through machine vision. For minute offsets and rotations that occur during PCB manufacturing tolerances or placement, the control algorithm can calculate the deviation matrix in real time and compensate for the coordinates of the theoretical hole positions. This mechanism ensures that even with minor errors in PCB placement, the support pins can accurately fall into the predetermined safety "islands," fundamentally solving the problems of collisions and scrap caused by blind pin placement.

[0019] 4. Software-defined manufacturing process enables second-level changeover and flexible production. This invention transforms the rigid constraints of physical fixtures into flexible planning using software algorithms. By using a "reference PCB" for physical teaching or directly importing Gerber data, the system can reconstruct a completely new support matrix within minutes. When faced with PCB order changes of different sizes and hole layouts, the production line does not need to stop to replace bulky mechanical molds or rearrange cylinder arrays, achieving rapid response to multi-variety, small-batch orders.

[0020] 5. Fully automated workflow, reshaping the efficiency of screen printing processes. By integrating image recognition, obstacle avoidance calculation, and pin picking into a fully automated closed loop, this invention completely replaces the inefficient work mode that relies on manual hand-eye coordination. The tireless, high-speed, and precise operation of the robotic arm significantly improves pin placement speed, eliminates subjective errors caused by human fatigue, and perfectly matches the production cycle of the solder resist screen printing process with the front-end and back-end automated processes. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the system structure of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention; Figure 3 This is a schematic diagram illustrating the coordinate mapping and compensation principle of the present invention; Figure 4 This is a schematic diagram of the support hole selection and avoidance logic of the present invention; Figure 5 This is a schematic diagram showing the mating relationship between the end effector of the fabric nail and the substrate of the present invention; Figure 6 This is a comparison diagram of the spatial distribution of the teaching state and the pinning effect of the present invention. Detailed Implementation

[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see the appendix Figure 1 - Figure 6 This invention provides an automated pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot. This method employs an automated pin placement system for PCB solder mask silkscreen printing based on machine vision and a three-axis robot. This system aims to replace traditional manual pin placement operations with digital means, achieving automated and intelligent control of the support process during PCB solder mask silkscreen printing. Physically, the system mainly consists of a uniform aperture array substrate, a three-axis robot, a pin feeding module, a pin placement end effector, a machine vision module, and a control system.

[0024] The uniform aperture array substrate, serving as the working platform for the entire system, is fixedly positioned within the working range of the three-axis robot. The surface of this substrate is machined with high-precision, matrix-distributed reference holes. In this embodiment, all reference holes are spaced at a predetermined fixed interval. The reference holes are arranged in a strict grid pattern, forming a physical grid covering the entire working area. Each reference hole is a through hole or blind hole perpendicular to the substrate surface, and its diameter is precisely matched with the diameter of the insertion end of the support pin to ensure the perpendicularity and stability of the support pin after insertion.

[0025] To facilitate subsequent algorithm description and position calculation, this embodiment constructs a global physical coordinate system on the uniform aperture array substrate, denoted as the substrate coordinate system. Set the lower left corner (or center point) of the substrate plane as the origin of the coordinate system. The row direction of the reference hole matrix is ​​parallel to The axis and column direction are parallel to Axis. For the first axis in the matrix. line, number The theoretical center coordinates of any reference hole in the column in the substrate coordinate system. It can be defined by the following formula: in, The offset of the first reference hole relative to the origin. The height coordinates of the substrate surface are given. This structured hole design discretizes the physical space, providing a physical basis for subsequently mapping the continuous image coordinates of the PCB to a discrete support matrix.

[0026] A three-axis robotic arm is mounted above a uniformly perforated substrate, possessing three orthogonal degrees of freedom in space. Driven by servo motors or stepper motors, and combined with a precision lead screw or synchronous belt transmission mechanism, the robotic arm can drive an end effector in [various directions]. and Perform planar positioning in the direction, and in The system performs lifting and lowering movements in the direction of movement. The control system controls the rotation angle of each axis motor by sending pulse signals, thereby achieving closed-loop control of the end position.

[0027] The pin-holding end effector is mounted on the Z-axis end flange of the three-axis robot. This actuator is designed with various mechanisms for gripping or adsorbing the support pin, such as pneumatic fingers, electromagnetic chucks, or mechanical grippers. The actuator's central axis is coaxial with the Z-axis direction of movement. An integrated positioning sensor, either internally or on the side of the actuator, provides feedback on whether the support pin has been successfully gripped and inserted into the reference hole.

[0028] The nail feeding module is positioned within the travel range of the three-axis robot, but on one side of the uniform hole array substrate, so as not to interfere with the main nail placement area. The nail feeding module includes a vibratory feeder or linear feeding mechanism to automatically sort and transport the scattered support nails to a fixed pick-up position. The position coordinates are taught and stored in the control system during system initialization to ensure that the robot arm can accurately grasp a single support pin each time.

[0029] The machine vision module is the core of this system's perception, used to acquire image information from a reference PCB. In a preferred implementation architecture, the machine vision module adopts an "eye-on-hand" mounting method, where the industrial camera and light source assembly are directly fixed to the crossbeam or Z-axis slider of a three-axis robot, moving synchronously with the robot. This method allows the camera to perform localized high-resolution imaging or multi-image stitching scanning of a large-format PCB from different positions. The industrial camera's optical axis is perpendicular to the surface of the uniformly perforated substrate, and, in conjunction with a high-brightness coaxial light source or ring light source, clearly presents the contrast of the metal pads, vias, and substrate contours on the PCB surface.

[0030] The machine vision module establishes an independent image pixel coordinate system. To achieve visual guidance, the system needs to establish a mapping relationship between the image coordinate system and the substrate coordinate system through calibration. The control system, as the computation and scheduling center of the entire system, is electrically connected to the three-axis robot actuator, the nail feeding module controller, the end effector solenoid / pneumatic valve, and the industrial camera via an industrial bus or I / O interface. The control system is internally equipped with a high-performance processing unit, running software programs that include image processing algorithms, path planning algorithms, and motion control logic.

[0031] The system's workflow begins with the placement of the reference PCB. The operator or loading mechanism places a reference PCB, serving as a standard sample, at any position on the uniform aperture array substrate. At this point, the actual physical position of the reference PCB relative to the substrate coordinate system... There exists a certain positional relationship between them, which is determined by the translation coordinates of the reference PCB on the substrate. and rotation angle These parameters are uniquely determined. Due to the randomness of the placement operation, these parameters are unknown before each job and need to be obtained through subsequent visual scanning and algorithmic calculation.

[0032] The control system first drives a three-axis robotic arm to move the machine vision module above the reference PCB, performing a single shot or a serpentine path scan based on the preset field of view (FOV). The acquired image data is transmitted to the control system's memory for subsequent feature extraction and coordinate system transformation calculations. This hardware architecture and motion logic ensure that the system can perceive the true pose of the PCB by "seeing" without the need for dedicated fixture positioning, providing a complete physical and data foundation for subsequently generating high-precision obstacle avoidance pin placement schemes.

[0033] In this embodiment, after acquiring image information of the reference PCB on the uniform aperture array substrate, the control system initiates an image processing and feature recognition program. Since the original image may contain ambient light noise or interference from the substrate background texture, the system first preprocesses the acquired image data. The preprocessing includes converting the color image to a grayscale image and applying Gaussian filtering or median filtering algorithms to remove random noise and enhance the sharpness of image edges. Subsequently, the system uses an adaptive threshold segmentation algorithm (such as the Otsu algorithm) to binarize the filtered image, separating the feature region of the reference PCB from the background region, thereby extracting clear connected component contours.

[0034] After acquiring the binarized image, the control system performs geometric feature analysis on the connected components. The system identifies circular features in the image representing vias, plug-in holes, or marker points, and calculates the position of each circular feature in the image coordinate system using the gray-scale centroid method or circle fitting algorithm. The pixel center coordinates below These pixel coordinates precisely characterize the positions of each hole on the reference PCB within the field of view of the vision sensor. To convert the visual information into spatial coordinates that the robotic arm can execute, the control system calls pre-calibrated hand-eye calibration parameters. These parameters define the transition from the image pixel coordinate system to the physical coordinate system of the uniform aperture array substrate. The mapping relationship.

[0035] In this invention, while parsing image data, the control system simultaneously reads and parses the digital processing data of a reference PCB. The processing data is typically a Gerber file or an ODB++ file, containing the theoretical outline data of the PCB and the theoretical position information of all drilled holes. The system extracts the theoretical coordinates of key feature points (such as positioning holes or specific pads) used for positioning from the processing data, denoted as a set. Let the first... The theoretical coordinates of the feature points are: At this point, the system's task is to establish a mathematical transformation model between the theoretical coordinate system (PCB design coordinate system) and the current actual physical coordinate system (uniform aperture array substrate coordinate system).

[0036] Since the reference PCB is randomly placed on the uniform aperture array substrate, either manually or mechanically, it inevitably exhibits translational and rotational deviations relative to the theoretical state within the substrate plane. To eliminate these deviations, the control system executes an image registration algorithm. The system sets the coordinates of the actual feature points identified visually. The set of coordinates of theoretical feature points in the processing data Perform matching. Calculate the transformation matrix parameters, including the rotation angle, using the least squares method or the RANSAC (Random Sample Consensus) algorithm. And the translation in the X and Y axes. and .

[0037] Based on the calculated deviation parameters, a coordinate compensation model is established for the control system. For any theoretical hole position coordinate recorded in the machining data... Its actual projected coordinates on the uniform aperture array substrate Full-field coordinate compensation is performed using the affine transformation formula. This coordinate compensation formula is expressed as follows: In the formula, This refers to the theoretical coordinates of a specific hole location in the machining data. This refers to the actual coordinates of the hole location in the coordinate system of the uniform aperture array substrate after compensation. The rotation angle of the PCB relative to the substrate coordinate system is used as a reference, with the counterclockwise direction being positive; and These are the translation components of the reference PCB origin relative to the origin of the substrate coordinate system.

[0038] Through the above calculations, the control system not only locates the feature points used for registration, but more importantly, maps the theoretical coordinates of hundreds or thousands of vias, through-holes, and areas of components that need to be avoided from the processing data to the current physical substrate coordinate system. The system fuses these compensated and corrected coordinate data to generate a "avoidance hole map" that strictly corresponds to the physical dimensions of the uniform aperture array substrate. This avoidance hole map precisely marks which areas on the uniform aperture array substrate, in the current placement state, have PCB holes or sensitive components directly above them, thus defining the no-entry zones for support pin placement. This process ensures that subsequent pin placement planning is entirely based on the actual physical orientation of the PCB, rather than an idealized theoretical state.

[0039] In this embodiment, after generating the avoidance hole location map containing precise coordinate information, the control system enters the core pin placement planning stage. The goal of this stage is to select a set of target reference holes from all available reference holes on the uniform hole array substrate that can effectively support the reference PCB without interfering with any avoidance areas. The system first retrieves the reference hole coordinate database of the uniform hole array substrate, which stores the center coordinates of all physical hole positions distributed in a matrix on the substrate, denoted as the set. ,in The total number of reference holes on the substrate, any number of reference holes The coordinates are .

[0040] The control system determines the vertical projection area of ​​the reference PCB on the uniform aperture array substrate based on the actual contour information after coordinate compensation, and records it as the effective support area. To ensure that the support pins effectively provide support rather than being exposed on the PCB board, the system first performs a first round of filtering: traversing the set. For each reference hole in the array, determine its coordinates. Is it located in the effective support area? Inside the geometric boundary. Commonly used algorithms for judgment include the ray casting method or the rotation number method. After this round of screening, the system obtains a subset of candidate reference holes located under the PCB board. .

[0041] Subsequently, the control system targeted the candidate benchmark hole set. Perform the crucial second round of obstacle avoidance screening. The system reads the coordinate data of all points to be avoided (such as through holes and plug-in holes) in the obstacle avoidance hole map and records them as the obstacle avoidance point set. ,in The total number of points to be avoided, and any number of points to be avoided. The coordinates are To quantify obstacle avoidance safety, the system calculates the safety of each candidate reference hole. With all avoidance points Determine the Euclidean distance between them and determine whether the distance satisfies the safety conditions.

[0042] The specific avoidance judgment logic is as follows: For any candidate reference hole Calculate its relationship with the set The closest avoidance point at medium distance minimum distance between The calculation formula is: In the formula, For the first The center coordinates of the candidate reference holes For the first The system specifies the center coordinates of the points that need to be avoided. A safety threshold is preset in the system. This threshold is typically set by comprehensively considering the radius of the support pin tip, the PCB hole diameter tolerance, and the positioning error of the robotic arm. If the calculated... Then determine the reference hole Keep away from all danger zones and meet safety anchoring requirements; conversely, if If the reference hole is too close to or directly opposite the PCB hole, it indicates that there is a risk of damage, and the system will reject it.

[0043] After two rounds of rigorous screening, the system obtained a set of qualified reference holes that were located under the PCB board and avoided all other hole positions. To further optimize the support effect and prevent uneven distribution of support points from causing PCB deformation during silkscreen printing, this invention introduces a density-based secondary optimization strategy. The control system analyzes the spatial distribution density of the qualified reference holes, retaining all qualified holes according to the principle of maximizing the number of support points, or thinning out locally high-density areas according to the principle of uniform distribution, ultimately determining the optimal support pin layout array, denoted as […]. .

[0044] The array of dots Each point in the data not only represents the position of a physical reference hole, but also contains a set of corresponding robot arm motion command data. Thus, the control system completes the entire process from image perception, data fusion, geometric calculation to path planning, transforming complex physical avoidance requirements into clear digital dot matrix coordinates, providing precise guidance data for the subsequent automated execution of the robot arm.

[0045] In this embodiment, the control system generates the final support nail placement matrix. Afterwards, the system prompts the operator or the automated transfer mechanism to remove the reference PCB from the uniform aperture array substrate. This step is crucial, as it fully exposes the reference holes, previously covered by the reference PCB, to the workspace of the three-axis robot, removing physical obstacles for subsequent pin insertion. At this point, the uniform aperture array substrate returns to an unloaded state, ready to receive the placement of support pins.

[0046] Control system then analyzed The coordinate data is used to generate a series of continuous motion control commands, driving the three-axis robot to perform automated nail placement. This process is a cyclical grasping and placing operation. First, the control system moves the three-axis robot to the fixed pick-up position of the nail supply module. Once above this position, the Z-axis drives the pin-holding end effector to descend to a predetermined height, activating the gripping mechanism (such as closing the pneumatic fingers or opening the electromagnetic chuck) to securely grasp a support pin. After the end effector's built-in sensor confirms successful gripping, the Z-axis rises to a safe movement height to avoid any potential obstacles on the substrate.

[0047] Subsequently, the control system uses the coordinates of the first target reference hole in the array of dots. The optimal path from the current position to the target position is planned. A gantry-shaped trajectory is typically used for point-to-point movement, i.e., first vertically lifting, then horizontally moving, and finally vertically descending. The robot arm moves at high speed in the XY plane to directly above the target reference hole and performs precise positioning. After positioning, the Z-axis descends vertically at a controlled speed, accurately aligning and inserting the insertion end of the support pin into the target reference hole of the uniform aperture array substrate.

[0048] During insertion, to prevent jamming due to minute errors, the control system can employ a compliant control strategy or introduce a small number of floating degrees of freedom into the mechanical structure. When the Z-axis reaches the preset insertion depth... At this time, the end effector releases the support pin (e.g., by releasing the pneumatic finger or disconnecting the electromagnetic chuck), and then the Z-axis rises back to a safe height, completing a single pin-laying action. The system monitors the execution status of each step in real time. If any abnormality occurs, such as gripping failure or insertion obstruction, an alarm is immediately triggered and the operation is suspended.

[0049] The above sequence of actions—grabbing, moving, inserting, and releasing—is executed cyclically until the support pins are deployed in a dot matrix. All coordinate points have been pinned. Once the last support pin is in place, the robotic arm automatically returns to its origin or standby position. At this point, a support array tailored to a specific PCB model has been formed on the uniform aperture array substrate. The operator or automatic feeder places the PCB to be processed, which matches the reference PCB model, onto this support array.

[0050] Because this support array is planned and generated based on the actual image and hole position data of the reference PCB, and the placement of the PCB to be processed is consistent with the original teaching position of the reference PCB (ensuring this through positioning pins or visual secondary alignment), all support pins will be precisely lifted onto the PCB board plane, perfectly avoiding all via holes, through-holes, and component areas. This process ensures that the PCB board surface receives uniform and stable support during subsequent solder mask silkscreen printing, effectively resisting deformation caused by squeegee pressure, thereby significantly improving silkscreen printing quality and yield.

[0051] In this embodiment, after the automated pin placement process is completed, the system enters the actual production application stage. At this time, a set of high-precision, customized support pin arrays has been formed on the uniform aperture array substrate according to the predetermined plan. The distribution pattern of this array perfectly matches the bottom topology of the PCB to be processed, providing both sufficient mechanical support density and creating perfect physical clearance space. To verify the pin placement effect and start formal production, the operator or automated conveyor line transports the first PCB of the same model to be processed (production board) to the worktable.

[0052] When placing the PCB to be processed, it is essential to ensure that its position and orientation on the uniform aperture array substrate are highly consistent with the placement of the reference PCB (teaching board) in step S1. To achieve this critical repeatability, the present invention employs two implementation strategies: first, mechanical positioning is achieved using pre-set positioning pins or positioning stops on the substrate to ensure complete physical overlap between the two; second, a machine vision module is used again to perform "secondary visual alignment" of the PCB to be processed. This involves the system acquiring real-time images of the PCB to be processed, calculating its positional deviation from the reference PCB, and compensating for this minute error by controlling the moving platform of the screen printing machine or adjusting the gripping and placing position of the robotic arm, thereby achieving sub-millimeter level precision alignment.

[0053] After the PCB to be processed is positioned and stabilized, the solder mask screen printing machine begins the printing operation. The squeegee moves across the screen under constant pressure and speed, printing solder mask ink through the mesh onto the PCB surface. During this process, the array of support pins distributed beneath the PCB plays a crucial role. Because the support points are positioned based on Euclidean distance... With rigorous avoidance calculations, no support pin will accidentally enter the PCB's vias or through-holes, thus completely eliminating common quality hazards such as "damaging the copper in the hole", "piercing the solder mask ink inside the hole", or "support pin getting stuck in the hole".

[0054] Meanwhile, thanks to the uniform distribution or maximized density strategy adopted in the pin layout, the support pin array can effectively counteract the vertical load generated when the squeegee presses down. Especially when the PCB board is thin (e.g., below 0.8mm) or the board surface is large, the dense support points act like countless miniature pillars, firmly supporting the PCB board surface and preventing the board from "denting" or "vibrating" due to stress. This stable support environment ensures the uniformity of the solder mask ink thickness, avoids ink seepage, exposed copper, or pattern misalignment caused by board surface deformation, and significantly improves the yield and appearance quality of PCB solder mask silkscreen printing.

[0055] Furthermore, the system described in this invention possesses extremely high flexibility in production. When the production line needs to switch between different PCB models, operators do not need to replace the entire support fixture; they only need to reset the system and replace it with a new reference PCB for a new round of visual scanning and teaching. The control system will automatically clear the old pin data and instantly generate a new pin dot matrix. The three-axis robot then performs "pin removal" or "rearrangement" actions (or the substrate can be quickly reset manually), completing a new round of support environment setup in a very short time. This "scan-and-use" rapid changeover mode greatly reduces reliance on specialized fixtures and warehousing costs, providing strong technical support for agile manufacturing of multi-variety, small-batch PCBs.

[0056] This embodiment also provides an overall workflow control logic based on the above system to further illustrate the execution details of the method claims of this invention. Although the foregoing embodiments mainly describe a single preferred architecture, those skilled in the art should understand that, without departing from the core concept of this invention, the specific configuration of the system can be adjusted and modified in various ways according to the actual production environment.

[0057] In this embodiment, regarding the deployment of the machine vision module, in addition to the aforementioned "eye on hand" follow-up scanning mode, the present invention also covers a "eye outside hand" global shooting mode. In this variant embodiment, the industrial camera is fixedly mounted directly above the uniform aperture array substrate, equipped with a high-resolution large target area sensor or telecentric lens. After the reference PCB is placed, the camera performs a single-exposure image of the entire substrate area. The control system directly performs coordinate recognition and distortion correction based on this global image, without the need for a robotic arm to perform scanning motion. Although this configuration has higher requirements for camera hardware costs, it can further shorten the image acquisition time and is suitable for application scenarios with extremely high production cycle requirements.

[0058] Furthermore, in the avoidance judgment logic described in step S3, although the calculation method based on Euclidean distance has been primarily described above, in other embodiments of the present invention, Manhattan distance or Chebyshev distance can also be used as the judgment criterion, or a geometric intersection test method based on polygon Boolean operations can be used. That is, the circular or square outline of the top of the support pin is directly constructed, and collision detection is performed with the outline of the PCB avoidance hole. As long as the logic is based on coordinate information of visual recognition to determine physical interference, it falls within the protection scope of the present invention.

[0059] Regarding the optimization strategy for the support pin placement matrix, in addition to the aforementioned principle of uniform distribution, this invention also supports intelligent planning based on "mechanical simulation." The control system can incorporate a simplified finite element analysis (FEA) model to simulate the board surface deformation cloud map under screen printing pressure, based on the size, thickness, and material properties of the reference PCB (such as the elastic modulus of FR4 substrate). The system prioritizes increasing the support point density in areas with the largest predicted deformation values ​​(usually the board center or large areas without holes). This physical model-based planning method can achieve optimal flatness control with the minimum number of support pins, further improving pin placement efficiency.

[0060] At the hardware execution level, the three-axis manipulator is not limited to a gantry structure; it can also employ a SCARA horizontal articulated manipulator or a six-axis industrial robot. As long as it possesses the ability to precisely position itself within the substrate plane and perform vertical insertion and removal along the Z-axis, it can serve as the actuator for this system. Similarly, the end effector for nail placement can be selected from magnetic, vacuum-adhesive, or mechanical clamping types depending on the material of the support nails. It can also be designed as a multi-head parallel actuator, capable of grasping multiple support nails at once for batch placement, thereby significantly increasing operational efficiency.

[0061] In summary, this invention, through a technical approach of "physical demonstration + visual compensation + automatic planning," effectively solves industry pain points such as low efficiency, poor precision, and susceptibility to board damage in the PCB solder mask screen printing support process. The system not only achieves fully automated closed-loop control but also possesses strong flexible production adaptability through decoupled hardware and software design.

[0062] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot, characterized in that, An automatic pin-laying system for PCB solder mask silkscreen printing is applied. The system includes a uniform aperture array substrate, a three-axis robot, a pin feeding module, a pin-laying end effector, a machine vision module, and a control system. The uniform aperture array substrate is fabricated with a matrix-distributed reference aperture. The method Includes the following steps: Includes the following steps: Step S1: Place the reference PCB on the uniform aperture array substrate, obtain the processing data of the reference PCB, and use the machine vision module to collect image information of the reference PCB on the uniform aperture array substrate; Step S2: The control system processes the image information, identifies the actual hole position features of the reference PCB, maps the coordinate system where the image information is located to the coordinate system of the uniform hole array substrate, and generates an avoidance hole position map containing the coordinate information of the avoidable area by combining the processing data. Step S3: Based on the avoidance hole location map and the outline information of the reference PCB, the control system performs pin placement planning in the coordinate system of the uniform hole array substrate, selects target reference holes from the reference holes of the uniform hole array substrate, and forms a support pin placement matrix. Step S4: Remove the reference PCB from the uniform aperture array substrate to expose the reference aperture; Step S5: The control system generates control commands based on the support nail array, controls the three-axis robot to drive the nail-laying end effector to grab the support nails from the nail supply module, and inserts the support nails into each target reference hole of the uniform hole array substrate in sequence.

2. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, Step S2 specifically includes: The control system analyzes the processing data and extracts the theoretical outline and theoretical hole position information of the reference PCB; The control system performs image recognition on the image information to obtain the actual position information of the vias and plug-in holes on the reference PCB. The positional deviation of the reference PCB relative to the uniform aperture array substrate is calculated using an image registration algorithm. The positional deviation is then used to perform coordinate compensation on the theoretical aperture position information. The compensated information is then fused with the actual position information to generate the avoidance aperture position map.

3. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, In step S3, the specific rules for pin placement planning in the coordinate system of the uniform aperture array substrate are as follows: A coordinate database of all reference holes on the uniform hole array substrate is pre-established; Traverse the coordinate database and select the reference holes that simultaneously meet the following conditions as the target reference holes: Condition 1: The coordinates of the reference hole are located within the projection range of the reference PCB outline on the substrate; Condition 2: The coordinates of the reference hole are not located within the coordinate range of the area to be avoided in the avoidance hole location diagram, ensuring that the support pin avoids the hole on the PCB.

4. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 3, characterized in that, The rules for pin placement planning also include optimization strategies: Among the selected set of reference holes that meet conditions one and two, a second screening is conducted according to the principle of uniform distribution or the principle of maximizing the number of support points to determine the final support pin layout array, so as to ensure that the PCB is subjected to uniform force.

5. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, The method for acquiring image information in step S1 is as follows: The three-axis robot arm is controlled to move the machine vision module mounted on it to the top of the reference PCB for scanning and imaging, or an independently fixed industrial camera is controlled to perform a global image of the reference PCB on the uniform aperture array substrate.

6. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, Step S5 specifically includes: The three-axis robot is controlled to move to the fixed outlet position of the nail feeding module, and a single support nail is grasped by the nail-feeding end effector; Based on the coordinates of the current target reference hole in the support pin array, control the three-axis robot to move in the horizontal plane to directly above the target reference hole; Control the Z-axis of the three-axis robot to descend, vertically insert the support pin into the target reference hole, and then release it; Repeat the above steps until all positions in the support nail array have been nailed.

7. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, The reference holes of the uniform aperture array substrate are through holes or blind holes arranged in a longitudinal and transverse pattern at a preset spacing of 5 mm to 10 mm.

8. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, The end effector for the nailing is selected from any one of pneumatic grippers, electromagnetic chucks, or mechanical chucks.

9. The automatic pin placement method for PCB solder mask silkscreen printing based on machine vision and a three-axis robot as described in claim 1, characterized in that, The method further includes step S6: After step S5 is completed, the PCB to be processed, which is the same model as the reference PCB, is placed on the pre-laid support pin array, and the placement position of the PCB to be processed on the uniform hole array substrate is consistent with the placement position of the reference PCB in step S1, and the subsequent solder mask silkscreen printing operation is performed.

10. An automatic pin-laying system for PCB solder mask silkscreen printing, used to implement the method described in any one of claims 1 to 9, characterized in that, include: A uniform aperture array substrate is fixedly mounted, and its surface is machined with a matrix of uniformly distributed reference holes. A three-axis robotic arm is positioned above the uniform aperture array substrate and has X-axis, Y-axis and Z-axis degrees of freedom of motion. The nail supply module is used to store and output support nails one by one; The nail end effector is installed at the Z-axis end of the three-axis manipulator and is used to grasp the support nail; The machine vision module is used to acquire image information of the PCB. The control system is electrically connected to the three-axis robot, the nail feeding module, the nail-laying end effector, and the machine vision module. The control system is equipped with a processing unit configured to perform the data processing and instruction generation steps belonging to the control system in any one of claims 1 to 9. The steps include: parsing the processing data of the reference PCB; processing the image of the reference PCB acquired by the machine vision module to generate an avoidance hole location map; planning and generating a support nail placement matrix in the substrate coordinate system; and generating control instructions based on the support nail placement matrix to drive the three-axis robot to perform the nail-laying action.

Images