A method and system for protecting peelable gold fingers during electric milling operations

Abstract

This disclosure relates to the field of electronic manufacturing technology and discloses a method and system for protecting peelable adhesive gold fingers during electric milling operations. It aims to solve problems such as burrs, metal wire drawing, and high post-processing costs associated with electric milling of gold fingers. The invention obtains the gold finger position and the electric milling path, proactively analyzes and calculates dynamic protection parameters, drives a protective adhesive application module to precisely apply liquid protective adhesive to the edge of the milling path according to the parameters, and a dynamic curing module rapidly cures the adhesive to form a high-strength protective layer. The protective layer is then removed by an electric milling machine and a peeling mechanism. By adopting the above technical solution, this application can effectively suppress milling burrs and metal wire drawing, improve the processing quality, efficiency, and yield of gold fingers, and reduce costs.

Description

Technical Field

[0001] This invention belongs to the field of electronic manufacturing technology, and more specifically, relates to a method and system for protecting peelable gold fingers during electric milling operations. Background Technology

[0002] The printed circuit board (PCB) manufacturing technology field encompasses several core stages, including pattern transfer, etching, surface treatment, and molding. The core of this technology lies in forming conductive patterns and interconnect structures on an insulating substrate to provide electrical connections and mechanical support for electronic components. The gold fingers, as the key interface connecting the PCB to external systems, directly impact the reliability and performance of the product due to their processing quality. With the development of electronic products towards higher density and integration, PCB manufacturing technology has been widely applied, covering multiple application scenarios such as communication equipment, computers, consumer electronics, and industrial control. Technological advancements in this field are the foundation and guarantee for the development of the entire electronics and information industry.

[0003] Among them, the gold finger electro-milling forming process refers to cutting the edges of the circuit board with a high-speed rotating milling cutter to precisely form the outline and size of the gold fingers. This process aims to mechanically remove excess board material and copper foil, enabling mass production and low-cost processing of densely packed gold fingers. In this process, by setting parameters such as the milling path, rotation speed, and feed rate, the milling cutter is controlled to process the circuit board to meet the design tolerances and appearance standards. This method is widely used in the final forming stage of circuit boards due to its high efficiency and relatively low cost.

[0004] Current technologies typically involve directly machining the gold fingers during electric milling operations, lacking effective process protection mechanisms. This leads to the easy formation of metal burrs on the edges of the gold fingers under cutting forces, affecting product appearance and shortening the lifespan of the milling cutter. The ductility of the metal can also cause wire drawing, resulting in short circuits between adjacent gold fingers in densely packed wiring areas, directly leading to product scrap. Post-processing steps introduced to address these defects, such as chemical etching or mechanical polishing, not only increase process complexity and production costs but also pose a risk of damaging the gold finger plating, making yield control difficult. Faced with increasingly stringent requirements for appearance and electrical performance, simply adjusting milling parameters is no longer a fundamental solution, while alternatives such as laser cutting significantly increase manufacturing costs and extend production cycles. This contradiction between direct exposure of the gold finger body during processing and the lack of effective protection has become a key bottleneck restricting the application of electric milling technology in high-performance circuit board manufacturing. Summary of the Invention

[0005] The purpose of this invention is to overcome the defects of existing technologies in the processing of gold fingers by electric milling, such as edge burrs, metal wire drawing short circuits, high post-processing costs, and poor adaptability to high-precision processing scenarios. In this way, a method and system for protecting peelable gold fingers during electric milling operations is proposed.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a peelable adhesive gold finger protection system for electric milling operations, the operation method of which includes the following steps:

[0007] S1: Obtain the reference point coordinates and layout design data of the printed circuit board to be processed. The layout design data includes the geometric topology information of the gold finger area. Based on the reference point coordinates, the printed circuit board is accurately positioned and its posture is corrected by the machine vision system to generate a gold finger position mapping map in the global coordinate system.

[0008] S2: Retrieve the preset milling tool path planning file. The path planning file defines the motion trajectory function P of the milling head in the global coordinate system. m (t); The system analyzes the motion trajectory function and, combined with the preset protection lead time Δt, calculates and generates a protective adhesive application path function P that is synchronized with the electric milling tool path but leads in time. p (t), where P p (t)=P m (t+Δt).

[0009] S3: Based on the gold finger position mapping map and the electric milling cutter path planning file, a forward-looking analysis window is established. This window moves forward in real time along the electric milling cutter path. Within this window, the system performs quantitative analysis on the geometric features of the gold fingers that are about to enter the machining area and calculates and generates a set of dynamic protection parameters. These dynamic protection parameters are used to adjust the application status and curing strength of the subsequent protective adhesive in real time.

[0010] S4: Drive the controllable phase change protective adhesive application module, causing its nozzles to follow the protective adhesive application path function P. p (t) moves and adjusts the amount of protective adhesive applied in real time according to the colloid flow parameter Q(t) in the dynamic protection parameter set. Before the electric milling cutter arrives, a layer of liquid protective adhesive is precisely applied to the edge of the milling path of the gold finger.

[0011] S5: Drive the dynamic curing module to follow immediately after the protective adhesive application module, and based on the curing energy parameter E in the dynamic protective parameter set. c (t) Instantaneous and localized energy input is applied to the liquid protective adhesive that has been applied, causing it to undergo rapid phase change and solidify, forming a solid protective layer with high mechanical strength and high modulus. This protective layer provides rigid support for the edge of the gold finger.

[0012] S6: Drive the electric milling execution module so that its tool strictly follows the electric milling tool path function P_m(t) to mill the gold finger area covered and supported by the solid protective layer; after the machining is completed, the solidified protective layer is removed from the circuit board surface by a special peeling mechanism.

[0013] As a further aspect of the present invention, the step S3 of generating dynamic protection parameters specifically includes:

[0014] S311: The system follows the milling tool path P m (t) Set a rectangular look-ahead analysis window W with dimensions L×W, and its center point is located at the future position P of the milling cutter. m (t+Δt lookahead At point W, the system extracts the area of ​​all gold finger regions falling within the window W from the gold finger position mapping map and calculates the ratio of the area to the total area of ​​the window, thereby obtaining the gold finger density index FDI(t), which characterizes the local gold finger density.

[0015] S312: System's milling tool path function P m (t) Perform differentiation operation at the current analysis point to obtain the curvature κ(t) of the path; based on the curvature κ(t), calculate the geometric stress factor GSF(t) that characterizes the complexity of the milling path and the risk of stress concentration. The calculation formula is: GSF(t)=1+ɑ·κ(t), where α is the preset curvature weighting coefficient.

[0016] S313: The system integrates the gold finger density index FDI(t) and geometric stress factor GSF(t), and calculates and generates real-time colloidal flow rate parameter Q(t) and curing energy parameter E through a preset control model. c (t). The formula for calculating the colloidal flow parameter Q(t) is: Q(t) = Q base ·(1+β·FDI(t))·GSF(t), where Q base The reference colloid flow rate is given by β, which is the density influence coefficient; the curing energy parameter E is given by β. c The formula for calculating E(t) is: c (t)=E c_base +γ·FDI(t)+δ·GSF(t), where E c_base γ is the base curing energy, γ is the gain coefficient of density on curing energy, and δ is the gain coefficient of geometric stress on curing energy.

[0017] As a further aspect of the present invention, the system also includes a closed-loop feedback control step, which is executed after step S5 and before step S6:

[0018] S5_feedback: After the dynamic curing module, a laser displacement sensor is set up. This sensor is used to measure the surface contour height h of the cured protective layer in real time. measured (t); The system will measure the height h. measured (t) and the target height h calculated based on dynamic protection parameters target By comparing (t), we obtain the height error e(t) = h. target (t)-h measured (t); The system adopts a proportional-integral-derivative (PID) control algorithm, which generates a flow regulation quantity ΔQ based on the height error e(t). feedback And the subsequent colloidal flow rate parameter Q(t+Δt) feedback The compensation and correction are performed, and the correction relationship is: Q corrected (t+Δt feedback )=Q(t+Δt feedback )+ΔQ feedback ,in

[0019]

[0020] K p K i K d These are the proportional, integral, and derivative coefficients of the PID controller, respectively.

[0021] As a further aspect of the present invention, the system further includes an acoustic emission signal monitoring sub-step in step S6:

[0022] S6_monitor: A high-frequency acoustic emission sensor is installed on the spindle of the electric milling execution module. This sensor is used to collect acoustic emission signals during the cutting process of the tool cutting the gold fingers and solidifying the protective layer. The system performs a Fast Fourier Transform (FFT) on the collected acoustic emission signals to analyze their energy amplitude in a specific high-frequency band (e.g., 300kHz-500kHz). When the energy amplitude shows a sudden spike exceeding a preset threshold, the system determines it as a sign of cutting instability or impending burr formation, and records the timestamp, coordinates, and corresponding FDI(t) and GSF(t) values ​​of the event into a self-optimization database for offline iterative optimization of the coefficients α, β, γ, and δ in the control model.

[0023] As a further aspect of the present invention, the controllable phase change protective adhesive is a polymer colloid with a dual-curing system. Its components include: a main resin, which is bisphenol A epoxy resin; a thermal initiator, used to trigger initial gelation at a lower temperature (e.g., 60°C-80°C) to increase the colloid viscosity and prevent flow; and a photoinitiator, a triarylsulfonium salt cationic photoinitiator, used to achieve rapid and deep volumetric curing under ultraviolet light irradiation of a specific wavelength. In its uncured state, the colloid has a viscosity of 45,000 to 55,000 centipoise at 25°C; after complete curing, its Shore D hardness is not less than 80, and its peel strength to a gold surface is less than 5 Newtons per centimeter.

[0024] As a further aspect of the invention, the dynamic curing module comprises a micro ceramic resistance heater and an ultraviolet light-emitting diode array connected in series, both coaxially mounted and immediately following the nozzle of the protective adhesive application module. The micro ceramic resistance heater is used to achieve initial thermal initiation; its power is adjustable, enabling the applied adhesive to reach the target gel temperature within 0.5 seconds. The ultraviolet light-emitting diode array has a center wavelength of 365 nanometers, and its optical power density is linearly adjustable within the range of 0 to 25 watts per square centimeter. Its output power is controlled in real time by the curing energy parameter E_c(t) to achieve final curing of the protective layer.

[0025] The present invention also provides a peelable gold finger protection system for electric milling operations, the system comprising:

[0026] The substrate positioning and data input module is used to load and use machine vision technology to perform high-precision alignment of the printed circuit board, obtain its precise position in the device coordinate system, and load layout design data containing gold finger geometry information and motion path files of the electric milling cutter.

[0027] The predictive path and parameter generation module integrates a path calculation unit and a parameter model calculation unit. The path calculation unit receives the motion path of the electric milling cutter and calculates and generates the synchronous motion trajectory of the protective adhesive application module based on the set time advance Δt. The parameter model calculation unit establishes a look-ahead analysis window along the motion trajectory, analyzes the gold finger pattern features within the window in real time, and continuously calculates and generates a dynamic protective parameter sequence for controlling the amount of protective adhesive applied and the curing energy based on the predefined gold finger density index (FDI) and geometric stress factor (GSF) models.

[0028] The controllable phase change protective adhesive application module includes a high-precision volumetric screw valve and a sub-millimeter nozzle. Driven by a servo motor, the module strictly follows the motion trajectory and adhesive flow parameter Q(t) specified by the predictive path and parameter generation module to precisely apply protective adhesive to the area to be milled on the gold finger.

[0029] The dynamic curing module, physically following the protective adhesive application module, integrates a programmable miniature heat source and a focused ultraviolet light source. The energy output of the miniature heat source and ultraviolet light source is determined by the curing energy parameter E generated by the predictive path and parameter generation module. c (t) Perform independent and real-time control to ensure rapid and controllable phase change curing of the applied protective adhesive.

[0030] The sensing and closed-loop control module includes a laser displacement sensor and a high-frequency acoustic emission sensor. The laser displacement sensor is deployed after the dynamic curing module to measure the actual morphology of the cured protective layer and feeds the measurement data back to a PID controller, which corrects the colloid flow parameters to eliminate morphology errors. The acoustic emission sensor is installed on the electric milling spindle to monitor stress wave signals during the cutting process and provides data input for the offline self-optimization of the system model.

[0031] The milling execution and stripping module includes a standard high-speed milling spindle and an automated protective layer stripping mechanism; the movement of the milling spindle is strictly synchronized with the aforementioned modules to ensure that it cuts when the protective layer provides maximum support; the stripping mechanism is activated after all milling operations are completed to completely remove the cured protective layer from the circuit board through non-contact or low-adhesion contact methods.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0033] This invention fundamentally solves the problems of burrs and metal streaking caused by the lack of support at the edges of gold fingers in electric milling by constructing a predictive and adaptive protection system. Instead of passively addressing machining defects, the system proactively analyzes the upcoming machining area and dynamically calculates and generates optimal protection parameters based on the actual density of the gold fingers and the geometric complexity of the milling path. This method makes the application of the protective layer highly targeted and resource-efficient, providing support only where necessary and with the required strength.

[0034] Furthermore, the coordinated operation of the controllable phase change protective adhesive application module and the dynamic curing module enables "instant application and curing" at the microscale. The resulting solid protective layer has extremely high mechanical strength, providing solid lateral and vertical support for the high-speed rotating milling cutter. It effectively suppresses the plastic deformation of metal materials under cutting forces, thereby ensuring the cleanliness and steepness of the milled edges. The finished product's appearance quality is comparable to that of laser cutting.

[0035] Furthermore, the integrated sensing and closed-loop control modules endow this invention with adaptive and self-optimizing capabilities. Real-time height feedback control based on a laser displacement sensor ensures the accuracy and consistency of the protective layer's morphology, unaffected by slight changes in equipment status or environment. The introduction of an acoustic emission sensor establishes a reverse optimization link from the processing result to the control model, enabling the system to continuously iterate and improve its internal predictive model by learning from microscopic failure events in actual processing, thereby continuously enhancing the protective effect and processing yield.

[0036] In summary, this invention deeply integrates materials science, precision mechanics, real-time sensing, and predictive control algorithms to break through the technical bottlenecks of traditional electric milling processes with a low-cost, high-efficiency automated integrated solution. This enables electric milling to handle the high-density, high-appearance-requirement gold finger processing tasks that were previously only achievable with high-cost processes, significantly improving production efficiency, reducing manufacturing costs, and enhancing process stability and reliability. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the overall process of a peelable adhesive gold finger protection method provided in an embodiment of the present invention;

[0038] Figure 2 This is a functional module diagram of a peelable adhesive gold finger protection system provided in an embodiment of the present invention;

[0039] Figure 3 for Figure 1 A flowchart of the steps for generating dynamic protection parameters in China;

[0040] Figure 4 This is a schematic diagram illustrating the application scenario of the coordinated operation of protective adhesive application, dynamic curing, and electric milling in an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram of the closed-loop feedback control process for the protective layer height based on a laser sensor in an embodiment of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In the description of this invention, it should be understood that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and 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, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. (See also...) Figure 1 This illustrates the overall process of a method for protecting peelable adhesive gold fingers during electric milling operations. In one specific embodiment, the method can be... Figure 2 The system is executed as shown. This method specifically includes the following steps:

[0043] Step S1: Obtain the reference point coordinates and layout design data of the printed circuit board to be processed. The layout design data includes the geometric topology information of the gold finger area. Based on the reference point coordinates, the printed circuit board is accurately positioned and its posture is corrected by the machine vision system to generate a gold finger position mapping map in the global coordinate system.

[0044] Specifically, this step can be broken down into multiple sub-steps.

[0045] Step S11: The system loads the layout design data and reference point coordinate file of the printed circuit board to be processed. The layout design data is usually in the industry-standard GerberRS-274X format or ODB++ database format. The system's internal parser module reads the file and specifically extracts all geometric primitives related to the gold finger layer (e.g., top or bottom gold plating layer). Each gold finger primitive is parsed into a data structure containing the primitive type (e.g., polygon, circular pad), vertex coordinate sequence, layer identifier, and a unique internal ID. For example, a rectangular gold finger is stored as {ID:"GF_007",Type:"Polygon",Layer:"TopCopper",Vertices:[(x1,y1),(x2,y2),(x3,y3),(x4,y4)]}, where all coordinates are within the original coordinate system of the layout design software. Simultaneously, the system loads a reference point coordinate file in XML or JSON format. This file predefines the precise coordinates of at least two non-collinear reference points (FiducialMarks) in the layout design coordinate system. For example, the file content could be {Fiducials:[{ID:"FM1",X:5.000,Y:5.000},{ID:"FM2",X:195.000,Y:5.000}]}, with units in millimeters.

[0046] Step S12: Capture actual reference point images of the printed circuit board using a machine vision system. The machine vision system includes a high-resolution industrial CCD camera, such as a monochrome camera with a 2048x2048 pixel resolution, whose lens is calibrated to eliminate radial and tangential distortion. The camera is mounted on a programmable XY motion platform along with a coaxial LED ring light source. After the printed circuit board is fixed to the worktable, the system drives the XY platform, causing the camera to move sequentially to approximately the position above each predetermined reference point and capture high-resolution images. In one embodiment, to address potential slight board bending or warping, the system can adjust the Z-axis height via an autofocus module during image capture to ensure the sharpness of the reference point images.

[0047] Step S13: Process the captured image, calculate the precise position of the actual reference point in the device coordinate system (i.e., camera coordinate system), and generate a coordinate transformation matrix. The system uses a template matching algorithm, such as the Normalized Cross-Correlation (NCC) algorithm, to match the pre-stored reference point standard template image with the real-time captured image to locate the pixel coordinates of the reference point center with sub-pixel accuracy.

[0048] In a specific computational example:

[0049] The design reference point coordinates are: FM1 design = (5.000, 5.000) mm, FM2 design = (195.000, 5.000) mm.

[0050] The pixel-to-millimeter conversion after camera calibration is: 0.01mm / pixel.

[0051] The pixel coordinates captured by the camera are: FM1 cam = (502, 498) pixels, FM2 cam = (19508, 503) pixels.

[0052] Convert to millimeter coordinates in the device coordinate system:

[0053] FM1 actual =(502*0.01,498*0.01)=(5.02,4.98)mm.

[0054] FM2 actual =(19508*0.01,503*0.01)=(195.08,5.03)mm.

[0055] Based on the design coordinate point set {FM1 design FM2 actual} and the actual coordinate point set {FM1 design FM2 actual The system calculates a two-dimensional affine transformation matrix T. This matrix T describes the translation, rotation, and scaling transformations from the layout design coordinate system to the device global coordinate system. The transformation relationship is represented by P. actual =T·P design The matrix T effectively compensates for translational and rotational deviations that occur during the clamping process of the circuit board. For example, the calculated translational vector is (0.02, -0.02) mm, and the rotation angle is arctan((5.03-4.98) / (195.08-5.02))≈0.0147 degrees.

[0056] Step S14: Generate a gold finger position mapping map in the global coordinate system. The system applies the affine transformation matrix T obtained in step S13 to the coordinates of all vertices of all gold finger primitives parsed in step S11. After the transformation, the geometric information of each gold finger is converted from the ideal design coordinates to the actual physical coordinates of the device workbench. These transformed gold finger geometric data are organized into an efficient data structure, namely the gold finger position mapping map. In a preferred embodiment, the mapping map is stored using a spatial index structure such as a quadtree or an R-tree. This structure divides the entire working area into a grid, enabling rapid retrieval of all gold finger primitives within a given coordinate range, greatly improving the query efficiency of the prospective analysis window in subsequent steps. As an alternative, a simple hash table can be used, where the key is a discrete grid cell ID and the value is a list of gold finger objects falling into that cell, which provides acceptable performance even when the gold fingers are relatively evenly distributed.

[0057] Step S2: Retrieve the preset electric milling tool path planning file and parse the motion trajectory function P. m (t), and combined with the preset protection lead time Δt, calculate and generate a protective adhesive application path function P that is synchronized with the electric milling tool path but ahead in time. p (t). Specifically, this step can be broken down into the following sub-steps.

[0058] Step S21: The system retrieves and parses the electric milling toolpath planning file. This file is typically in standard G-code (ISO6983 / EIARS-274-D) format, generated by CAM software based on milling requirements. The file contains a series of instructions, such as G00 (rapid positioning), G01 (linear interpolation), G02 / G03 (circular interpolation), and F (feed rate). The system's G-code parser reads the file line by line, converting the discrete instruction sequence into a continuous, time-parameterized motion trajectory function P. m (t)=(x m (t),y m (t)).

[0059] For example, a G01 command G01X50.0Y20.0F300 indicates that the tool moves linearly from the current point to coordinates (50.0, 20.0) at a speed of 300 mm / min (i.e., 5 mm / s). If the current point is (10.0, 20.0), then the length of this path is 40 mm, and the time taken is 40 mm / 5 mm / s = 8 seconds. The corresponding motion trajectory function is P. m(t) = (10.0 + 5*t, 20.0), where 0 ≤ t ≤ 8. The system converts the entire G-code file into a piecewise function sequence, which fully describes the spatiotemporal position of the milling cutter throughout the machining process. At this stage, the system performs a syntax check. If it finds unparseable G-code or kinematically unreachable instructions (such as exceeding travel limits), it will interrupt the process and issue an alarm.

[0060] Step S22, calculate the protective adhesive application path function P p (t). The core of this step is to introduce a protection lead time Δt. This lead time is a composite parameter whose source is precisely calibrated, Δt = Δt response +Δt dispense +Δt gel .

[0061] Where, Δt response It refers to the response and communication latency of the protective adhesive application module servo system, representing the time from receiving the instruction to the nozzle actually starting to move. It is measured by a laser interferometer, with a typical value of 35 milliseconds.

[0062] Δt dispense It is the time required for the colloid to form a stable adhesive strip on the substrate surface after it is extruded from the nozzle. It is related to the viscosity of the colloid, the nozzle diameter, and the application speed. It is calibrated by observation with a high-speed camera, and the typical value is 45 milliseconds.

[0063] Δt gel This refers to the time required for the micro-heat source in the subsequent dynamic curing module to initially heat the colloid, bringing it to a gel state to prevent it from spreading. This depends on the thermosensitive properties of the colloid and the power of the heater, and is experimentally determined; in this embodiment, this value is 500 milliseconds.

[0064] Therefore, the total lead time for protection is Δt = 35 + 45 + 500 = 580 milliseconds, or 0.58 seconds.

[0065] Protective adhesive application path function P p (t) is defined as P p (t)=P m (t+Δt). This means that at any time t, the position of the protective adhesive application nozzle is equal to the position of the milling cutter t+0.58 seconds later.

[0066] Continuing with the example in step S21, the milling toolpath is P. m (t) = (10.0 + 5*t, 20.0).

[0067] The protective adhesive application path function is:

[0068] P p (t)=P m(t+0.58)=(10.0+5*(t+0.58),20.0)=(10.0+5t+2.9,20.0)=(12.9+5t,20.0).

[0069] This indicates that when the milling cutter is located at (10.0, 20.0) (at t = 0), the protective adhesive nozzle is already located at (12.9, 20.0), physically ahead by 2.9 mm. This lead distance d lead =v·Δt, which changes dynamically with the milling speed v.

[0070] Step S3: Based on the gold finger position mapping map and the electric milling tool path planning file, a forward-looking analysis window is established. This window moves forward in real time along the electric milling tool path. Within this window, the system performs quantitative analysis on the geometric features of the gold fingers about to enter the machining area and calculates and generates a set of dynamic protection parameters. (Refer to...) Figure 3 The detailed process for this step is as follows.

[0071] Step S311: The system sets a rectangular forward-looking analysis window W along the milling cutter path P_m(t). The geometric center of this window is defined at the future position P of the milling cutter. m (t+Δt lookahead At point ), where Δt lookahead This is a configurable look-ahead analysis time, greater than the protection lead time Δt, typically set between 1.0 and 2.5 seconds. This parameter is set based on the delay in system parameter calculation and the buffer time required to ensure smooth parameter output. In one embodiment, Δt... lookahead The time is set to 1.2 seconds. The window size is L×W, where the length L is parallel to the tangent direction of the path at that point, and the width W is perpendicular to the tangent direction. The length L is related to the milling speed v(t), L=v(t)Δt window , where Δt window This is the time span covered by the analysis window, for example, 0.8 seconds. The width W is set based on the widest possible array of gold fingers plus a safety margin, for example, 6.0 mm.

[0072] The system utilizes the gold finger position mapping map (spatial index structure) generated in step S14 to efficiently extract all gold finger polygons that intersect with or are contained within the rectangular region of the current analysis window W. Subsequently, the system employs a polygon clipping algorithm (such as the Weiler-Atherton algorithm) to accurately calculate the area of ​​each gold finger within the window, and sums all these areas to obtain the total area of ​​the gold fingers within the window. fingers Finally, the gold finger density index is calculated.

[0073]

[0074] The following calculations are illustrated using two sets of virtual data examples:

[0075] Example 1 (Sparse Region): The milling cutter is approaching a region containing several 0.2mm wide gold finger pins along a straight line. The current speed v(t) = 10mm / s.

[0076] Window length L = 10mm / s * 0.8s = 8mm. Window width W = 6mm. Total window area Area fingers =8*6=48mm 2 .

[0077] After spatial index lookup and cropping calculation, the total area of ​​the gold fingers within the window is obtained. fingers =4.5mm 2 .

[0078] The gold finger density index FDI(t) = 4.5 / 48 ≈ 0.094.

[0079] Example 2 (Dense Area): The milling cutter is about to enter the edge of a dense BGA pad array. Current speed v(t) = 10 mm / s.

[0080] Window size remains unchanged, Area window =48mm 2 .

[0081] After spatial index lookup and cropping calculation, the total area of ​​the gold fingers within the window is obtained. fingers =32.0mm 2 .

[0082] The gold finger density index FDI(t) = 32.0 / 48 ≈ 0.667.

[0083] As an alternative, the gold finger density index can be calculated not based on area, but on the count of path intersections. This alternative defines FDI. alt (t) represents the milling toolpath P within the analysis window. m (t) is the number of intersections with the geometric boundary of the gold finger. This method may be more sensitive to assessing the processing complexity of regions composed of a large number of independent, subtle features.

[0084] Step S312: The system performs a differential operation on the electric milling tool path function P_m(t) at the current analysis point to obtain the curvature κ(t) of the path. For the piecewise parameterized path P parsed from the G-code... m (t)=(x(t),y(t)), curvature κ(t) is obtained through the formula

[0085]

[0086] The calculation involves the first derivative (x', y') representing the velocity vector and the second derivative (x”, y”) representing the acceleration vector. In practical calculations, the system numerically approximates these derivatives by performing differences on discrete points along the path.

[0087] The geometric stress factor GSF(t) is calculated based on the curvature κ(t), and its calculation formula is GSF(t)=1+α*κ(t).

[0088] Here, α is the curvature weighting coefficient, derived from extensive cutting experiments and finite element simulation analysis. During the calibration phase, standard specimens were milled using circular arc paths of varying radii, while the cutting state was monitored using high-speed cameras and force sensors. It was observed that as curvature increased (radius decreased), the cutting force, particularly the lateral force, increased non-linearly, leading to a sharp increase in the risk of burrs and delamination. The coefficient α was calibrated to a value that strongly positively correlates GSF(t) with the measured peak cutting force or the observed defect probability. In a specific embodiment, α was set to 80 mm. This unit makes α*κ(t) a dimensionless quantity.

[0089] The following calculations are illustrated using two sets of virtual data examples:

[0090] Example 1 (Straight Path): The path is the straight line segment G01. x””t)=0, y””t)=0.

[0091] Curvature κ(t) = 0.

[0092] The geometric stress factor GSF(t) = 1 + 80 * 0 = 1.0. This means that during straight-line cutting, there is no additional stress caused by the path geometry.

[0093] Example 2 (sharp turn): The path is a G02 circular arc segment with a radius R = 1.0 mm.

[0094] Curvature κ(t) = 1 / R = 1 / 1.0mm = 1.0mm -1 .

[0095] Geometric stress factor GSF(t) = 1 + 80mm * 1.0mm -1 =1 + 80 = 81.0. This huge value reflects how sharply the lateral force load on the tool and the tearing effect on the gold finger edge will worsen at sharp turns.

[0096] As an alternative, the geometric stress factor can incorporate a jerk term into the path to address the impact effects when entering and leaving curves. The alternative formula is:

[0097]

[0098] Where α_jerk is the jerk weighting coefficient.

[0099]

[0100] It is the rate of change of curvature. This solution can improve the protection level earlier, preparing for the upcoming curve.

[0101] Step S313: The system integrates the gold finger density index FDI(t) and geometric stress factor GSF(t), and calculates and generates the real-time colloidal flow rate parameter Q(t) and curing energy parameter E through a preset nonlinear control model. c (t).

[0102] The formula for calculating the colloidal flow rate parameter Q(t) is: Q(t) = Q base ·(1+β·FDI(t))·GSF(t).

[0103] Parameter source explanation:

[0104] Q base : Reference colloid flow rate. This is the minimum flow rate required to form a standard protective adhesive strip when milling along a straight line (GSF=1) on a substrate without gold fingers (FDI=0). It is determined by the nozzle inner diameter, colloid viscosity, and target adhesive strip cross-sectional area, and is calibrated experimentally. For example, Q base =2.0μL / s.

[0105] β: Density Influence Coefficient. A dimensionless gain coefficient reflecting the additional amount of adhesive required to fully cover and fill the surface microstructure of the gold finger as the density increases. This coefficient is determined by performing application experiments on test plates with different FDI values ​​and checking the coverage integrity. For example, β = 1.8.

[0106] The formula for calculating the curing energy parameter E_c(t) is: The formula here has been optimized to directly correlate the energy increment with the flow increment and the stress factor increment.

[0107] Parameter source explanation:

[0108] Reference curing energy. This is the reference curing flow rate Q. base The energy input required for a colloid is expressed in watts per square centimeter (W / cm²). 2 It is related to the photo / thermochemical properties of the colloid, the curing time, and the target mechanical properties (such as hardness). For example,

[0109] γ: Flow-energy gain coefficient. It represents the additional curing energy required to achieve the same curing depth and crosslinking degree within a specified time for each unit increase in colloidal flow rate. This parameter is determined by measuring the curing kinetics of the colloid at different energies using digital subtraction calorimetry (DSC) and a rheometer. For example, γ = 0.5 (W / cm²). 2 ) / (μL / s).

[0110] δ: Geometric stress-energy gain coefficient. This indicates how much additional curing energy is required per unit increase in geometric stress factor. This is used to achieve higher modulus and hardness of the protective layer by increasing the degree of curing in areas facing high cutting stress. This coefficient is calibrated by establishing a curve showing the relationship between curing energy and mechanical properties (storage modulus, hardness) through dynamic mechanical analysis (DMA) and nanoindentation testing on samples cured at different energies. For example, δ = 0.05 (W / cm²) / GSF_unit.

[0111] Integrated calculus example:

[0112] Scenario 1 (sparse region with straight lines): FDI = 0.094, GSF = 1.0.

[0113] Q(t)=2.0*(1+1.8*0.094)*1.0=2.0*1.1692≈2.34μL / s.

[0114] E c (t)=8.5+0.5*(2.34-2.0)+0.05*(1.0-1)=8.5+0.17+0=8.67W / cm 2 .

[0115] At this point, the flow rate and energy were only slightly higher than the baseline values.

[0116] Scenario 2 (sharp turn in dense area): FDI = 0.667, GSF = 81.0.

[0117] Q(t)=2.0*(1+1.8*0.667)*81.0=2.0*(1+1.2006)*81.0=2.0*2.2006*81.0≈356.5μL / s.

[0118] E c (t)=8.5+0.5*(356.5-2.0)+0.05*(81.0-1)=8.5+177.25+4.0=189.75W / cm 2 .

[0119] At this point, the calculated flow rate and solidification energy both show a significant increase to cope with the extremely complex processing conditions. The system sends this series of time-varying (Q(t), Ec(t)) parameter pairs to the subsequent execution modules in the form of a data stream.

[0120] During processing, the system must consider exceptional cases. For example, if the computational load of the prospective analysis window is too large, causing the parameter generation delay to exceed Δt... lookahead -Δt, the system will trigger an "insufficient computing power" alarm and may selectively reduce the milling speed F to buy more time for parameter calculation, or temporarily adopt a set of conservative (higher) protection parameters based on historical data to ensure machining safety.

[0121] Step S4: Drive the controllable phase change protective adhesive application module so that its nozzle moves according to the protective adhesive application path function Pp(t) and adjusts the application amount of protective adhesive in real time according to the adhesive flow parameter Q(t) in the dynamic protective parameter set. Before the electric milling cutter arrives, a layer of liquid protective adhesive is precisely applied to the edge of the milling path of the gold finger.

[0122] Specifically, this step is achieved through the coordinated action of a high-precision volumetric screw valve and a sub-millimeter nozzle. (Refer to...) Figure 4 The nozzle of the controllable phase change protective adhesive application module is driven by a high-rigidity three-axis servo motor system. Its motion controller receives and executes the protective adhesive application path function Pp(t) generated in step S22. The motion controller interpolates the position command at an extremely high refresh rate (e.g., 2 kHz) to ensure the smoothness and accuracy of the nozzle motion trajectory, and its trajectory tracking error is controlled within 5 micrometers.

[0123] Simultaneously, the data stream of the colloid flow rate parameter Q(t) is sent to the screw valve controller in real time. The screw valve contains a precision-machined Archimedes screw, driven by a brushless DC servo motor. The screw's rotation angle has a precise linear relationship with the discharged colloid volume, a relationship calibrated beforehand using a weighing method. For example, the calibration result is that 0.5 microliters of colloid are discharged for every 360 degrees of rotation. The controller converts the received flow rate parameter Q(t) (unit: μL / s) into the target rotational speed of the screw motor (unit: RPM) in real time. The conversion formula is RPM(t) = Q(t)·C valve C valve It is the valve body conversion factor, for example, C. valve =(60s / min) / (0.5μL / rev)=120rev / (μL*min).

[0124] Continue using the calculation example from step S313:

[0125] In scenario one (sparse region straight line), Q(t)≈2.34μL / s.

[0126] The target speed of the screw motor is RPM(t) = 2.34 * 120 = 280.8 RPM.

[0127] In scenario two (sharp turn in dense area), Q(t)≈356.5μL / s.

[0128] The target speed of the screw motor is RPM(t) = 356.5 * 120 = 42780 RPM.

[0129] The closed-loop control system inside the screw valve controller (e.g., based on Hall sensor or encoder feedback) ensures that the actual motor speed accurately tracks the target speed, thereby enabling dynamic and instantaneous control of the amount of colloid applied.

[0130] The nozzle's inner diameter is optimized, for example, to 0.25 mm, to form a moderately wide, shape-stable strip at typical application speeds. The applied liquid protective adhesive precisely covers the edge of the gold finger to be milled, and its application position relative to the milling path is precisely calibrated to ensure that the cured protective layer provides the most effective lateral support to the tool's cutting edge.

[0131] Step S5: Drive the dynamic curing module so that it follows the protective adhesive application module immediately. Based on the curing energy parameter E_c(t) in the dynamic protection parameter set, it provides instantaneous and local energy input to the applied liquid protective adhesive, causing it to rapidly undergo phase change curing and form a solid protective layer with high mechanical strength and high modulus.

[0132] Specifically, the dynamic curing module is physically integrated with the protective adhesive application module, maintaining a fixed, extremely small gap (e.g., 5 mm) to ensure immediate curing before the adhesive spreads. (See reference...) Figure 4 The module consists of a miniature ceramic resistance heater and an array of ultraviolet light-emitting diodes (UV-LEDs) connected in series, and the two are mounted coaxially.

[0133] Controlled phase change protective adhesive is a carefully designed dual-curing polymer colloid. Its main component is bisphenol A epoxy resin, which serves as the host resin, providing high mechanical strength and excellent adhesion to the substrate after curing. A block-type thermal initiator is incorporated into the system. This initiator is inactive at room temperature, but when the temperature reaches a specific threshold (e.g., 60℃-80℃), its block structure breaks down, releasing active groups and triggering the initial ring-opening polymerization reaction of the epoxy resin. This reaction is slow, but it can increase the viscosity of the colloid from 45,000 centipoise in the liquid state to over 1,000,000 centipoise in the gel state within a very short time (e.g., within 0.5 seconds), effectively fixing the shape of the adhesive strip and preventing it from flowing or spreading due to surface tension or slight tilting of the substrate before subsequent UV curing. The system also contains a highly active triarylsulfonium salt cationic photoinitiator. This photoinitiator is highly sensitive to ultraviolet light of a specific wavelength (e.g., 365 nm). After absorbing photon energy, it rapidly decomposes to produce a strong protic acid. This acid acts as a catalyst for cationic polymerization, initiating an extremely rapid chain polymerization reaction of epoxy groups, achieving deep and volumetric curing of the colloid. The entire curing process is completed within 1-2 seconds.

[0134] The workflow of the dynamic curation module is as follows:

[0135] First, a miniature ceramic resistance heater provides non-contact radiant heating to the freshly applied liquid adhesive strip. The heater's power output is set to a fixed pulse mode, with energy just enough to raise the gel temperature above the activation temperature of the thermal initiator within a time interval of Δt_gel (e.g., 500 milliseconds), thus completing the initial gelation.

[0136] Next, a UV-LED array located after the heater irradiates the gelled adhesive strip with high-intensity ultraviolet light. The light spot of the UV-LED array is focused and homogenized by a special optical lens group, forming a rectangular light spot that matches the width of the adhesive strip. Its light power density is determined by the curing energy parameter E calculated in step S313. c (t) Real-time control. The control system will control E c (t)(Unit: W / cm²) 2 This is converted into the duty cycle of the pulse width modulation (PWM) signal for the UV-LED driver power supply, thereby linearly adjusting its actual output optical power.

[0137] Continue using the calculation example from step S313:

[0138] In scenario one (sparse region with straight lines), E c (t)=8.67W / cm 2 The controller outputs the corresponding PWM duty cycle to drive the UV-LED array to provide illumination at that power density.

[0139] In Scenario 2 (sharp turn in dense area), E c (t)=189.75W / cm 2 The controller adjusts the PWM duty cycle to an extremely high level, enabling the UV-LED array to output instantaneous peak power to meet the energy demands generated by the surge in adhesive volume and high stress requirements.

[0140] After this step, the protective adhesive transforms into a hard, solid protective layer. Once fully cured, this layer has a Shore D hardness of at least 80, providing sufficient rigidity to resist the lateral cutting forces of the milling cutter. Simultaneously, its peel strength to the gold surface, optimized through formulation, is controlled to below 5 Newtons per centimeter, ensuring ease of operation and non-destructive nature in subsequent peeling steps.

[0141] As a further embodiment of the present invention, after step S5 and before step S6, the system further performs a closed-loop feedback control step. (See also...) Figure 5 This step aims to further improve the consistency and accuracy of the protective layer morphology.

[0142] In step S5_feedback, after the dynamic curing module, a laser displacement sensor is set up. This sensor is used to measure the surface contour height h of the cured protective layer in real time. measured (t); The system will measure the height h. measured (t) and the target height h calculated based on dynamic protection parameters target By comparing (t), the height error e(t) is obtained; the system adopts a proportional-integral-derivative (PID) control algorithm to generate a flow regulation amount ΔQ based on the height error e(t). feedback And the subsequent colloidal flow rate parameter Q(t+Δt) feedback Compensation and correction will be made.

[0143] Specifically, the laser displacement sensor can be a high-frequency laser triangulation sensor or a confocal dispersion sensor, with a measurement frequency of up to 10 kHz and a height resolution better than 1 micrometer. This sensor physically follows the dynamic curing module, maintaining a fixed geometric relationship with the application nozzle and the curing module.

[0144] Target height h target (t) is not a fixed value, but is dynamically calculated by the system based on the current colloid flow rate parameter Q(t) and the application speed v(t). Its physical meaning is the cross-sectional height of the ideal cured adhesive strip formed under the current Q(t) and v(t). This relationship h targetThe function f(Q(t), v(t)) is obtained by fitting pre-existing experimental data, where the function f describes the relationship between the cross-sectional geometry of the adhesive strip and the flow rate and velocity. In a simplified model, if the cross-section of the adhesive strip is assumed to be semi-circular, then the cross-sectional area A = Q / v, and the height h = sqrt(2A / π). Therefore...

[0145]

[0146] The system calculates the height error e(t) = h in real time. target (t)-h measured (t). This error signal is input to a digital PID controller.

[0147] PID controller calculates flow regulation amount

[0148]

[0149] Among them, K p K i K d These are the proportional, integral, and derivative coefficients, which are tuned online using the Ziegler-Nichols method or relay feedback method to obtain a stable and fast response.

[0150] K p The term responds to the current error, if h measured If the flow rate is too low, increase it immediately.

[0151] K i The integral term is used to eliminate steady-state error. If the height remains low, the integral term will accumulate, causing the flow rate to increase continuously until the error is eliminated.

[0152] K d The term is used to predict the trend of error changes and suppress overshoot. If the altitude is rapidly approaching the target value, the derivative term will have a counteracting effect, preventing the flow from increasing too quickly and causing overshoot.

[0153] Finally, the corrected flow parameter is Q. corrected (t future )=Q originak (t future )+ΔQ feedback .

[0154] It should be noted that the correction here applies to a future time point t. future . t future The delay Δt relative to the current measurement time t feedback This is equal to the physical distance from the measurement point to the laying nozzle divided by the current laying speed. This ensures that feedback corrections are applied to the corresponding path segments that caused the error.

[0155] Calculation example:

[0156] Assume that at time t, Q(t) = 10 μL / s and v(t) = 20 mm / s.

[0157] Target height h target (t)=sqrt(2*10 / (π*20))≈0.564mm.

[0158] The laser sensor measures h measured (t) = 0.550 mm.

[0159] Error e(t) = 0.564 - 0.550 = 0.014 mm.

[0160] Assume the PID parameter is K p =50,K i =10,K d =5, and the current error changes steadily (de / dt≈0), with the cumulative value of the integral term being 0.02.

[0161] ΔQ feedback =50*0.014+10*0.02+5*0=0.7+0.2=0.9μL / s.

[0162] The system will compensate for future flow parameters, increasing the planned flow value by 0.9 μL / s, so that the height of the subsequently laid rubber strips will be closer to the target value.

[0163] Step S6: Drive the electric milling execution module so that its tool strictly follows the electric milling tool path function Pm(t) to mill the gold finger area covered and supported by the solid protective layer; after the machining is completed, the solidified protective layer is removed from the circuit board surface as a whole by a special peeling mechanism.

[0164] Specifically, the electric milling execution module includes a high-speed electric milling spindle with an adjustable speed between 30,000 and 60,000 RPM. The spindle is equipped with micro-milling cutters specifically designed for milling PCB gold fingers, such as 0.8 mm diameter fishtail or end mills. The motion control system of this module is strictly synchronized with the motion control system of the protective adhesive application module, both sharing the same master clock signal to ensure that the position of the electric milling cutter at any given time t precisely corresponds to its path function P_m(t). (Refer to...) Figure 4Because the application and curing of the protective adhesive precede the milling process, by the time the milling cutter reaches any position, the edge of the gold finger at that location is already covered and supported by a hardened protective layer. During the cutting process, the cutting edge of the cutter simultaneously cuts the copper / nickel / gold plating of the gold finger and the adjacent hardened protective layer. The hard protective layer acts as a "sacrificial support," providing strong lateral rigidity support for the soft edge of the gold finger, effectively inhibiting plastic flow, curling, or tearing of the metal material under cutting forces, thus fundamentally eliminating the generation of burrs and metal wires. The milled edge contour is clear, steep, and highly perpendicular.

[0165] After all milling operations are completed, a dedicated peeling mechanism is activated. In one specific embodiment, this peeling mechanism employs cryogenic assisted peeling technology. The mechanism includes a precisely positioned liquid nitrogen nozzle. This nozzle moves rapidly along the path of the cured protective layer, spraying a small amount of liquid nitrogen (-196°C). The extremely low temperature causes the cured epoxy protective layer to undergo rapid cold shrinkage and glass transition, making it extremely brittle. Because the linear thermal expansion coefficient of epoxy resin (approximately 55-65 ppm / °C) is much greater than that of the circuit board substrate FR-4 (approximately 14 ppm / °C) and copper (approximately 17 ppm / °C), this drastic temperature difference generates enormous shear stress at the interface between the protective layer and the substrate, exceeding the relatively low peel strength of the protective adhesive itself (<5 N / cm). This leads to interfacial debonding between the protective layer and the gold finger surface, and may cause microcracks.

[0166] Subsequently, a robotic arm equipped with a low-hardness polymer scraper or a row of synchronized high-pressure air nozzles moves along the path, gently removing the loosened and brittle protective layer from the circuit board surface in whole or in sections. The removed gold finger surface is clean, without any adhesive residue, requiring no additional cleaning process.

[0167] As another further embodiment of the present invention, the execution of step S6 also includes an acoustic emission signal monitoring sub-step, which is used to realize process monitoring and offline self-optimization of the model.

[0168] Step S6_monitor: A high-frequency acoustic emission sensor is installed on the spindle of the electric milling execution module. This sensor is used to collect acoustic emission signals during the cutting process of the tool cutting the gold fingers and curing the protective layer. The system analyzes the collected acoustic emission signals. When a characteristic signal related to burr generation is detected, the relevant parameters of the event are recorded in a self-optimizing database.

[0169] Specifically, the acoustic emission sensor is a broadband piezoelectric ceramic sensor with a response frequency range covering 100 kHz to 1 MHz. The sensor is secured to a housing near the spindle bearing by a magnetic base or screws to maximize signal coupling efficiency. The weak electrical signal output by the sensor is amplified by a low-noise preamplifier (40 dB gain) and then digitized by a high-speed data acquisition card at a sampling rate of at least 2 MHz.

[0170] The system performs real-time time-frequency analysis on the acquired digital signal stream using Short-Time Fourier Transform (STFT) or Wavelet Packet Transform (WPT). Research shows that microscopic events such as fracture and burr formation in metal materials during cutting release high-frequency stress waves, with their energy concentrated primarily in a specific frequency band, such as 300 kHz to 500 kHz. The system continuously calculates the root mean square (RMS) value or amplitude of the signal energy within this characteristic frequency band.

[0171] The system presets an energy threshold. AE This threshold is determined by acquiring a large number of baseline signals under normal, defect-free cutting conditions and statistically analyzing their energy distribution (e.g., taking the mean plus six standard deviations). During the machining process, if the characteristic frequency band energy E is calculated in real time... AE (t) shows a mutation spike exceeding the threshold, i.e., E AE (t)>Threshold AE The system determines this as a cutting instability event, a strong indication that burrs are about to form or have already formed.

[0172] Once such an event is detected, the system immediately logs the complete context information of the event to a dedicated self-optimizing database (e.g., an SQL or NoSQL database). The logged data entries include:

[0173] Event timestamp.

[0174] The precise three-dimensional coordinates (x, y, z) of the milling cutter at the time the event occurred.

[0175] The snapshot of the control parameters that caused the event is the value of the gold finger density index FDI(t) and geometric stress factor GSF(t) that the system was using at the time the event occurred.

[0176] The corresponding colloidal flow rate Q(t) and curing energy E c The set value of (t).

[0177] Characteristic parameters of acoustic emission signals, such as peak energy, duration, and ring count.

[0178] When the system is idle or at predetermined intervals, an offline optimization program analyzes all records in the database. This program employs machine learning algorithms, such as Gradient Boosting Decision Trees or Support Vector Regression, to build a predictive model from input parameters (FDI, GSF) to output (probability of acoustic emission event occurrence or peak energy). Then, by performing sensitivity analysis on this model or using optimization algorithms (such as genetic algorithms or Bayesian optimization), the control model coefficients that minimize the predicted instability probability (i.e., α, β, γ, δ in steps S312 and S313) are solved in reverse. The optimized coefficients are then updated in the online control system, enabling the system to automatically adopt better and more robust protection parameters when encountering similar layout features and path geometry in subsequent processing, achieving self-learning and iterative performance improvement.

[0179] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for protecting peelable adhesive gold fingers during electric milling operations, characterized in that, Includes the following steps: S1: Obtain the layout design data and reference point coordinates of the printed circuit board to be processed, and use the machine vision system to position and correct the posture of the printed circuit board to generate a gold finger position mapping map in the global coordinate system. S2: Parse the preset electric milling tool path planning file, convert it into an electric milling tool motion trajectory function, and combine it with the protection lead time to calculate and generate a protective adhesive application path function that is synchronized with the electric milling tool path but ahead in time. S3: Based on the gold finger position mapping map and the electric milling cutter motion trajectory function, a prospective analysis window is established to analyze the geometric features and path curvature of the gold finger within the window, generate the gold finger density index and geometric stress factor, and calculate a set of dynamic protection parameters accordingly. S4: Based on the protective adhesive application path function and the dynamic protection parameters, drive the protective adhesive application module and the dynamic curing module to work together to accurately apply and cure the protective adhesive in the milling area of ​​the gold finger to form a solid protective layer. The specific steps for obtaining the protective adhesive application path function are as follows: S211: Parse the G-code instruction sequence in the electric milling tool path planning file and convert it into the electric milling tool motion trajectory function that describes the position of the tool center point in the global coordinate system of the equipment, with time as the parameter; S212: Based on the system response delay, colloid stabilization time and initial gelation time of the protective adhesive application module, and setting the protective advance amount, the electric milling cutter motion trajectory function is positively translated on the time axis to generate the protective adhesive application path function. The specific steps for obtaining the dynamic protection parameters are as follows: S311: Set a forward-looking analysis window along the motion trajectory function of the electric milling cutter, calculate the ratio of the total area of ​​the gold fingers falling into the window to the area of ​​the window to obtain the gold finger density index, and obtain the path curvature by performing differential operation on the path function to obtain the geometric stress factor. S312: Integrate the gold finger density index and the geometric stress factor, substitute them into the preset control model, and calculate and generate the time-related dynamic protection parameters that include the colloidal flow rate parameters and curing energy parameters; The specific steps for forming the solid protective layer are as follows: S411: Drive the controllable phase change protective adhesive application module so that its nozzles strictly follow the protective adhesive application path function to precisely apply liquid protective adhesive to the path edge of the gold finger to be milled area. S412: Based on the colloid flow rate parameter in the dynamic protection parameters, adjust the application amount of the protective adhesive application module in real time to match the protection requirements of different areas; S413: Drives the following dynamic curing module to provide instantaneous and localized energy input to the applied liquid protective adhesive according to the curing energy parameter in the dynamic protection parameters, so that it can quickly undergo phase change curing to form the solid protective layer.

2. The method for protecting peelable adhesive gold fingers during electric milling operations according to claim 1, characterized in that, The gold finger position mapping map contains the geometric topology information of the gold fingers after coordinate transformation. The protective adhesive application path function characterizes the spatiotemporal motion trajectory of the protective adhesive nozzle in the equipment coordinate system. The gold finger density index and geometric stress factor respectively characterize the local gold finger density and the stress concentration risk of the milling path. The dynamic protection parameters include the colloid flow rate parameter and the curing energy parameter.

3. The method for protecting peelable adhesive gold fingers during electric milling operations according to claim 1, characterized in that, The specific steps for obtaining the gold finger position mapping map are as follows: S111: Obtain layout design data containing geometric information of the gold finger region, and capture actual images of at least two preset reference points on the printed circuit board through a machine vision system; S112: Based on the actual image of the reference point and the preset design coordinates, calculate and generate an affine transformation matrix for mapping the layout design coordinate system to the device global coordinate system; S113: Apply the affine transformation matrix to all the gold finger geometric information in the layout design data, transform its vertex coordinates to the device global coordinate system, and generate the gold finger position mapping map.

4. The method for protecting peelable adhesive gold fingers during electric milling operations according to claim 1, characterized in that, The method also includes step S5: S5: After dynamic curing and before electric milling, the surface contour height of the cured protective layer is measured by a laser displacement sensor, and the height error is obtained by comparing it with the target height. Based on the height error, the subsequent colloid flow parameters are compensated and corrected to generate a protective layer morphology correction command. The protective layer morphology correction command includes real-time measured height, target contour height, height error value, and PID flow adjustment amount.

5. The method for protecting peelable adhesive gold fingers during electric milling operations according to claim 4, characterized in that, The specific steps for obtaining the protective layer topography correction command are as follows: S511: Deploy a laser displacement sensor after the dynamic curing module to scan the surface of the cured protective layer in real time and obtain a real-time measurement sequence of the surface contour height. S512: Compare the real-time measurement value sequence with the target height sequence calculated based on dynamic protection parameters point by point to generate a height error sequence that changes over time; S513: Using a proportional-integral-derivative PID control algorithm, a flow rate adjustment is generated based on the height error sequence, and subsequent colloidal flow rate parameters are superimposed and corrected to obtain the protective layer morphology correction command.

6. The method for protecting peelable adhesive gold fingers during electric milling operations according to claim 1, characterized in that, The method also includes step S6: S6: During the electric milling process, acoustic emission signals during the cutting process are collected by an acoustic emission sensor installed on the electric milling spindle. The signals are analyzed in the frequency domain to identify signs of cutting instability or impending burr formation. The coordinates, gold finger density index and geometric stress factor corresponding to the signs are recorded in the self-optimization database.

7. A peelable adhesive gold finger protection system for electric milling operations, characterized in that, The system is used to implement the method for protecting peelable gold fingers during electric milling operations as described in any one of claims 1-6, the system comprising: The substrate positioning and mapping module is used to acquire the layout design data of the printed circuit board, perform precise positioning and posture correction through the machine vision system, and generate a gold finger position mapping map in the global coordinate system. The path and parameter generation module is used to parse the electric milling tool path to generate a time-advanced protective adhesive application path function, and to establish a forward-looking analysis window to calculate and generate a dynamic protective parameter sequence based on the gold finger density and path curvature. The protective layer application and curing module is used to follow the application path function of the protective adhesive and, based on the dynamic protective parameter sequence, precisely control the application amount and curing energy of the protective adhesive to form a solid protective layer in the gold finger area. The topography closed-loop feedback module is used to measure the surface contour of the cured protective layer through a laser displacement sensor, calculate its height error with the target contour, and generate flow compensation correction instructions through a PID controller. The cutting condition monitoring module is used to collect stress wave signals during the electric milling process through acoustic emission sensors, analyze signal characteristics to identify cutting abnormalities, and record event data to a database for offline optimization of the control model.

Images