A High-Precision Placement Control Method for LED Display Circuit Boards Based on Placement Machine Positioning
By combining and analyzing driving current, air pressure, and texture feature points, a correction matrix is generated, which solves the problem of distinguishing between the heat dissipation pads and signal pin pressure requirements of LED display circuit boards. This enables high-precision placement and defect pre-control, improving placement quality and efficiency.
Patent Information
- Application Number
- CN202611574475.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-12
- Publication Date
- 2026-11-17
AI Technical Summary
Existing surface mount technology (SMT) methods cannot accurately distinguish the different pressure requirements of heat dissipation pads and signal pins on LED display circuit boards, resulting in mounting quality relying on AOI inspection after high-temperature solder curing, making rework impossible.
By acquiring the driving current timing and Z-axis position value when the placement head picks up components, the pre-compensation positioning coordinates are generated; the timing of the descent air pressure fluctuation of the vacuum nozzle is acquired to generate a dynamic approach displacement curve; the texture image around the solder pad is acquired, the surface texture feature matrix is extracted, and the solder pad coordinate system is analyzed; the impact deviation and adsorption deviation are separated, a correction matrix is generated, and the servo drive parameters and solder paste thixotropic state are adjusted.
It enables real-time adjustment of pressing depth and speed during the placement process to meet the different needs of heat dissipation pads and signal pins, improves placement quality, and allows for immediate detection and pre-control of defects after placement, avoiding reliance on inspection after high-temperature solder curing.
Smart Images

Figure CN123124329A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of SMT (Surface Mount Technology) technology, specifically a high-precision placement control method for LED display circuit boards based on placement machine positioning. Background Technology
[0002] LED display circuit board surface mount technology (SMT) is a crucial step in mounting LED driver chips (QFN packages) onto the circuit board pads, and it's part of the "solder paste printing → surface mount → high-temperature soldering → LED chip mounting → inspection" process chain. Unlike ordinary circuit boards, LED display circuit boards use a black solder mask layer, resulting in a grayscale contrast between the pads and the solder mask layer that is only 40% to 60% of that of ordinary circuit boards. Traditional visual positioning algorithms struggle to accurately identify the pad boundaries. The bottom of the QFN package chip includes a central heat dissipation pad and surrounding signal pins, with the heat dissipation pad recessed 0.02 to 0.05 mm relative to the pins. The heat dissipation pad requires a greater depth of contact to ensure sufficient solder paste spread for heat conduction, while the signal pins only need slight contact to conduct the circuit; therefore, the required depth of contact differs between the two.
[0003] Existing surface mount technology (SMT) methods have the following shortcomings: the servo current signal contains both the frictional resistance of the drive train and the contact force between the chip and the solder paste, which traditional methods cannot distinguish, leading to inaccuracies in contact position determination. Using the same depth for all pads cannot simultaneously meet the different depth requirements of heat dissipation pads and signal pins. Mounting quality relies on AOI inspection after high-temperature solder curing, at which point the solder paste has solidified and the chip is fixed, making rework impossible. Therefore, this invention proposes a high-precision SMT control method for LED display circuit boards based on SMT machine positioning. Summary of the Invention
[0004] This invention aims to at least solve one of the technical problems existing in the prior art. To this end, this invention proposes a high-precision chip placement control method for LED display circuit boards based on chip mounter positioning.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: A high-precision placement control method for LED display circuit boards based on placement machine positioning includes: S1. Obtain the driving current timing and Z-axis position value when the placement head picks up the component, and generate the positioning coordinates before compensation by time domain decomposition. S2. Obtain the timing sequence of the descending air pressure fluctuation of the vacuum nozzle, and generate a dynamic approach displacement curve through air film damping inversion; S3. Obtain the texture image around the pad, extract the surface texture feature point matrix through multi-scale feature point extraction, and analyze the surface texture feature point matrix in combination with the dynamic approach displacement curve to obtain the instantaneous pad coordinate system. S4. Perform dynamic deviation separation between the pre-compensation positioning coordinates and the final convergence value of the dynamic approach displacement curve to generate impact deviation component and adsorption deviation component. S5. Map the impact deviation component to the servo drive parameter adjustment amount, and convert the adsorption deviation component into the solder paste thixotropic state evaluation index through solder paste thixotropic evaluation. Then, weight and fuse the servo drive parameter adjustment amount and the solder paste thixotropic state evaluation index to generate a correction matrix. S6. Obtain the rising air pressure fluctuation time sequence of the vacuum nozzle, extract the equivalent air tightness feature value after mounting through air tightness feature extraction, compare the deviation with the expected feature range derived from the correction matrix, mark the deviation as a suspected unstable solder joint, and perform correlation backtracking adjustment on the correction matrix parameters of adjacent subsequent mounting points.
[0006] Furthermore, the pick-and-place machine includes a three-dimensionally movable placement head, which is used to pick up components and move them above the target pads on the circuit board for placement. The placement head is driven by a servo motor equipped with a servo driver, which integrates a current feedback unit to output the real-time drive current value of the servo motor at a fixed acquisition cycle. A grating ruler is installed on the Z-axis of the placement head to measure the Z-axis position of the placement head. A vacuum nozzle is installed at the end of the placement head to pick up and release components. The vacuum nozzle is connected to a vacuum generator through an air path, in which a pressure sensor is installed to detect the internal air pressure of the vacuum nozzle. The pick-and-place machine itself is equipped with a vision imaging module. The vision imaging module is a standard downward imaging system for the pick-and-place machine, which includes an industrial camera and a coaxial light source.
[0007] Further, S1 includes: When the placement head picks up a component, obtain the timing of the drive current of the placement head; The driving current timing sequence is decomposed in the time domain to obtain the timing sequence of the fluctuation component corresponding to the driving current timing sequence. The Z-axis position value of the mounting head is acquired synchronously, and the position fluctuation mapping sequence is obtained based on the fluctuation component timing and the Z-axis position value of the mounting head. Each fluctuation component contained in the position fluctuation mapping sequence is converted into a spatial displacement deviation value corresponding to the Z-axis position value; The average bottom surface deviation value is obtained based on the spatial displacement deviation value corresponding to all Z-axis position values. The negative value of the average bottom surface deviation value is used as the Z-axis correction amount. The Z-axis correction amount is superimposed on the Z component of the theoretical positioning coordinates to generate the positioning coordinates before compensation.
[0008] Further, S2 includes: During the descent phase of the placement head, the Z-axis position of the placement head is monitored in real time. When the placement head enters the effective descent range, the air pressure sensor is activated to collect air pressure values, forming a descent air pressure fluctuation sequence. The time series of descending air pressure fluctuations is subjected to sliding window averaging filtering to remove high-frequency noise components, resulting in a filtered descending air pressure sequence. Perform air film damping inversion processing on the filtered descending air pressure sequence to obtain the air film thickness value corresponding to each air pressure value; Connect all the air film thickness values collected at all times in the order of the mounting head's descent stroke to generate a dynamic approach displacement curve, and record the convergence value at the end of the curve.
[0009] Further, the process of performing air film damping inversion on the filtered descending air pressure sequence to obtain the air film thickness value corresponding to each air pressure value includes: The first pressure value in the filtered descending pressure sequence is taken as the reference pressure value. The initial air film thickness corresponding to this reference pressure value is equal to the initial gap between the bottom surface of the component and the surface of the pad at the critical mounting height. For each pressure value in the filtered descending pressure sequence, the absolute value of the pressure value is used as the denominator, and the product of the reference pressure value and the initial air film thickness is used as the numerator. The two are divided to obtain the air film thickness value at the acquisition time. When the pressure value in the filtered descending pressure sequence approaches zero, the air film thickness value corresponding to the pressure value approaches the initial air film thickness, and the air film thickness value is set as the saturation value.
[0010] Further, S3 includes: When the placement head descends to the shooting height, the industrial camera is triggered to capture a grayscale image as the texture image around the pads; Perform contrast stretching preprocessing on the grayscale image to enhance the contrast between the pad edges and the solder mask layer; Feature point detection is performed on the preprocessed image at three neighborhood radius scales. The local response value of each pixel is calculated at each scale and non-maximum suppression is performed to screen candidate feature points. The candidate feature points selected from the three scales are merged and deduplicated to generate a surface texture feature point array. Multiply the image coordinates of each surface texture feature point in the surface texture feature point array by the pixel equivalent coefficient to convert them into physical coordinates of the pad. Obtain the terminal convergence value of the dynamic approach displacement curve, calculate the depth weighting factor, and weight the physical coordinates of the pads of the surface texture feature matrix to obtain the weighted origin coordinates. The instantaneous pad coordinate system is obtained based on the physical coordinates of the pads and the weighted origin of the surface texture feature matrix.
[0011] Further, obtaining the instantaneous pad coordinate system based on the physical coordinates of the pads using the surface texture feature point matrix and the weighted origin includes: Calculate the X-direction and Y-direction offsets of each surface texture feature point relative to the weighted origin; construct a 2×2 covariance matrix, the four elements of which are the sum of the squares of the X-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, and the sum of the squares of the Y-direction offsets of all surface texture feature points; calculate the two eigenvectors of the covariance matrix, the first eigenvector corresponding to the larger eigenvalue is used as the X-axis direction of the instantaneous pad coordinate system; the second eigenvector corresponding to the smaller eigenvalue is used as the Y-axis direction of the instantaneous pad coordinate system; use the weighted origin as the origin of the coordinate system, and normalize the two eigenvector directions to use them as the unit vectors of the coordinate axes.
[0012] Further, S4 includes: When the placement head reaches the critical placement height, obtain the Z component of the positioning coordinates before compensation and the convergence value at the end of the dynamic approach displacement curve; The actual height of the pad surface is obtained by subtracting the convergence value at the end of the dynamic approach displacement curve from the Z component of the positioning coordinates before compensation, and then subtracting the theoretical mounting height to obtain the comprehensive deviation value. In the first acquisition cycle after the placement head reaches the critical placement height, the speed difference between the actual instantaneous speed of the placement head and the commanded speed is calculated, and the speed difference is multiplied by the speed loop response time constant to obtain the impact deviation component. The adsorption deviation component is obtained by subtracting the impact deviation component from the overall deviation value.
[0013] Further, S5 includes: When the placement head presses in the solder paste, the impact deviation component and the adsorption deviation component are obtained; Map the impact deviation component to the servo drive parameter adjustment amount; Thixotropic evaluation of solder paste was performed on the adsorption deviation component to obtain the thixotropic state evaluation index of solder paste. Calculate the correction amount for the end position of the downward pressure and the correction amount for the downward pressure speed curve, combine them into a correction matrix, and send it to the position loop and speed loop of the servo system.
[0014] Further, S6 includes: During the rising phase of the placement head, the air pressure value inside the vacuum nozzle is collected to form the rising air pressure fluctuation sequence. Extract the recovery slope value and steady-state offset value from the time series of rising air pressure fluctuations; The normalized recovery slope value is obtained by dividing the recovery slope value by the baseline recovery slope value, and the normalized steady-state offset value is obtained by dividing the steady-state offset value by the baseline steady-state offset value. Calculate the equivalent airtightness characteristic value after mounting; Based on the correction matrix, the correction amount of the downward pressure endpoint position and the correction amount of the downward pressure speed curve are extracted, and the lower limit and upper limit of the expected characteristic interval are derived. The equivalent airtightness characteristic value after mounting is compared with the expected characteristic range. When the equivalent airtightness characteristic value after mounting is lower than the lower limit of the expected characteristic range or higher than the upper limit of the expected characteristic range, the current mounting point is marked as a suspected unstable solder joint. When a suspected unstable solder joint is marked, the Euclidean distance between the suspected unstable solder joint and subsequent unattached joints is calculated. For subsequent joints whose Euclidean distance is less than the preset neighborhood radius, the backtracking adjustment range is determined by segmenting them according to the Euclidean distance. The correction matrix parameters of the current solder joint are multiplied by the backtracking adjustment range and then superimposed onto the original correction matrix of the subsequent placement points to generate the updated correction matrix of the subsequent placement points.
[0015] Compared with the prior art, the beneficial effects of the present invention are: 1) This invention obtains the driving current timing and Z-axis position value when the mounting head picks up the component and performs time-domain decomposition to extract the fluctuation component caused by the change in the bottom surface shape of the component in the driving current value; obtains the real-time distance between the component and the pad by obtaining the timing of the descending air pressure fluctuation and generating a dynamic approach displacement curve through air film damping inversion; and dynamically separates the impact deviation component from the dynamic inertia of the machine and the adsorption deviation component from the material properties in the comprehensive deviation value by performing dynamic deviation separation between the pre-compensation positioning coordinates and the final convergence value of the dynamic approach displacement curve. This allows the two types of information mixed in the driving current signal to be gradually separated in different steps, avoiding the error in the correction direction caused by confusion of the causes of deviation. 2) This invention obtains the real-time distance between the component and the pad through air film damping inversion, providing a basis for determining when the bottom surface of the component contacts the solder paste; by separating the comprehensive deviation value into impact deviation component and adsorption deviation component, corresponding to machine dynamic characteristics and material characteristics respectively; by mapping the impact deviation component to servo drive parameter adjustment amount to control the buffer speed, and converting the adsorption deviation component into solder paste thixotropic state evaluation index to control the pressing depth, the two are weighted and fused to generate a correction matrix; this enables the placement head to adjust the buffer speed and pressing depth according to the machine dynamic state and solder paste rheological state during the contact and pressing process of solder paste, thereby simultaneously meeting the different pressing requirements of heat dissipation pads and signal pins; 3) This invention uses the generated adsorption deviation component to evaluate the thixotropic state of solder paste; it provides execution parameters for the mounting action through the generated correction matrix; by acquiring the rising air pressure fluctuation time sequence, the equivalent air tightness feature value after mounting is obtained through air tightness feature extraction, and the deviation is compared with the expected feature range derived from the correction matrix. When there is a deviation, it is marked as a suspected unstable solder joint, and the correction matrix parameters of adjacent subsequent mounting points are adjusted in a correlated retrospective manner; so that the mounting quality can be known in a short time after the completion of the mounting, and the adjacent subsequent mounting points can be proactively controlled before high-temperature soldering, and defects can be detected and controlled without waiting for the automatic optical inspection after high-temperature soldering curing. Attached Figure Description
[0016] Figure 1 This is a schematic diagram of the process for high-precision placement control of LED display circuit boards based on placement machine positioning according to the present invention. Detailed Implementation
[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.
[0018] like Figure 1 As shown, a high-precision placement control method for LED display circuit boards based on pick-and-place machine positioning includes: S1. Obtain the driving current timing and Z-axis position value when the placement head picks up the component, and generate the positioning coordinates before compensation by time domain decomposition. S2. Obtain the timing sequence of the descending air pressure fluctuation of the vacuum nozzle, and generate a dynamic approach displacement curve through air film damping inversion; S3. Obtain the texture image around the pad, extract the surface texture feature point matrix through multi-scale feature point extraction, and analyze the surface texture feature point matrix in combination with the dynamic approach displacement curve to obtain the instantaneous pad coordinate system. S4. Perform dynamic deviation separation between the pre-compensation positioning coordinates and the final convergence value of the dynamic approach displacement curve to generate impact deviation component and adsorption deviation component. S5. Map the impact deviation component to the servo drive parameter adjustment amount, and convert the adsorption deviation component into the solder paste thixotropic state evaluation index through solder paste thixotropic evaluation. Then, weight and fuse the servo drive parameter adjustment amount and the solder paste thixotropic state evaluation index to generate a correction matrix. S6. Obtain the rising air pressure fluctuation time sequence of the vacuum nozzle, extract the equivalent air tightness feature value after mounting through air tightness feature extraction, compare the deviation with the expected feature range derived from the correction matrix, mark the deviation as a suspected unstable solder joint, and perform correlation backtracking adjustment on the correction matrix parameters of adjacent subsequent mounting points.
[0019] It should be noted that, in the embodiments of the present invention, the pick-and-place machine includes a three-dimensionally movable placement head, which is used to pick up components and move them above the target pads on the circuit board for placement; the placement head is driven by a servo motor, which is equipped with a servo driver, and the servo driver integrates a current feedback unit for outputting the real-time drive current value of the servo motor at a fixed acquisition cycle; a grating ruler is mounted on the Z-axis of the placement head, which includes a reading head fixed on the motion guide rail of the placement head and a scale grating mounted on the body of the placement head. The reading head reads the scale lines on the scale grating by photoelectric scanning and converts them into a digital position signal output for measuring the Z-axis position of the placement head; a vacuum nozzle is mounted at the end of the placement head, which is used to adsorb and release components. The vacuum nozzle is connected to a vacuum generator through an air path, and the vacuum generator is used to generate negative pressure in the air path. A pressure sensor is installed in the air path for detecting the internal air pressure of the vacuum nozzle; In embodiments of the present invention, the driving current timing and Z-axis position value when the placement head picks up the component are obtained, and the positioning coordinates before compensation are generated through time-domain decomposition; specifically including: S11. When the placement head picks up a component, obtain the timing of the drive current of the placement head; The timing of the drive current of the placement head is obtained as follows: the drive current value is obtained through the current feedback unit of the servo driver. Specifically, the acquisition start condition is set to the placement head Z-axis position reaching the pick-up start height, which is, for example, 2 mm above the surface of the feeder component; the acquisition duration covers the entire pick-up action, for example, starting 10 milliseconds before the vacuum nozzle contacts the component and ending 10 milliseconds after the adsorption is completed, for a total of 50 milliseconds; the acquisition period is consistent with the acquisition period of the current feedback unit of the servo driver, for example, 0.1 milliseconds; the acquisition period is the time interval between two adjacent acquisitions. The collection range is limited to all drive current values from the moment the vacuum nozzle contacts the component until the component is completely separated from the feeder film. After extracting all drive current values, the drive current values during the no-load movement of the placement head are removed to generate the drive current timing sequence. S12. Perform time-domain decomposition on the drive current timing sequence to obtain the timing sequence of the fluctuation components corresponding to the drive current timing sequence; specifically including: The arithmetic mean of all drive current values included in the drive current timing sequence is calculated as the mean component; the mean component is subtracted from each drive current value to obtain the fluctuation component of the corresponding drive current value. If the fluctuation component of the corresponding drive current value is positive, it indicates that there is a bulge on the bottom surface of the component at the corresponding acquisition time; if the fluctuation component of the corresponding drive current value is negative, it indicates that there is a depression on the bottom surface of the component at the corresponding acquisition time. Understandably, if the fluctuation component of the corresponding drive current value is positive, it means that the corresponding drive current value is higher than the average of all drive current values. The larger the drive current value, the greater the load on the servo motor, which means the greater the force on the vacuum nozzle and the tighter the contact between the vacuum nozzle and the component. In this case, the bottom surface of the component contacted by the vacuum nozzle is thicker, meaning there is a protrusion in the bottom surface. If the fluctuation component of the corresponding drive current value is negative, it means that the corresponding drive current value is lower than the average of all drive current values. The smaller the drive current value, the smaller the load on the servo motor, which means the smaller the force on the vacuum nozzle and the looser the contact between the vacuum nozzle and the component. In this case, the bottom surface of the component contacted by the vacuum nozzle is thinner, meaning there is a depression in the bottom surface. Furthermore, it can be understood that when the pick-and-place machine in this invention is working, the vacuum nozzle contacts the component from above. That is, the vacuum nozzle contacts the upper surface of the component, while the lower surface of the component faces the circuit board pads. In the embodiments of this invention, the lower surface of the component is referred to as the bottom surface of the component. There is a certain thickness between the upper surface and the bottom surface of the component. When the vacuum nozzle contacts the upper surface of the component from above, if there is a bulge in a certain area of the bottom surface of the component, that is, the area is thicker than the surrounding area, the upper surface of that area will also rise accordingly. When the vacuum nozzle at the end of the placement head descends to a certain fixed height, the upper surface of the component will contact the vacuum nozzle earlier and more tightly. At this time, the reaction force on the vacuum nozzle is greater, the load on the servo motor is greater, and the drive current value is greater. Conversely, the situation where there is a depression in a certain area of the bottom surface is not described in detail.
[0020] The fluctuation components corresponding to all drive current values in the drive current timing sequence are arranged according to the acquisition time sequence to obtain the fluctuation component timing sequence. S13. Synchronously acquire the Z-axis position value of the mounting head, and obtain the position fluctuation mapping sequence based on the fluctuation component timing and the Z-axis position value of the mounting head; including: The Z-axis position value of the mounting head is acquired by a grating ruler data acquisition device, with the acquisition period consistent with the acquisition period of the current feedback unit of the servo driver, for example, 0.1 milliseconds. The fluctuation component at each acquisition moment is bound to the corresponding Z-axis position value to form a data pair sequence. The Z-axis position values are reordered from high to low to obtain the force deviation distribution of the component bottom surface at different Z-axis position values. The average value is taken for cases where there are multiple fluctuation components at the same Z-axis position value to obtain the position fluctuation mapping sequence of the component bottom surface at different Z-axis position values. S14. Convert each fluctuation component contained in the position fluctuation mapping sequence into a spatial displacement deviation value corresponding to the Z-axis position value; including: The force deviation value is obtained by multiplying the fluctuation component value by the servo motor torque constant; the servo motor torque constant is provided by the motor manufacturer, for example, 0.1 N / Ampere; the spatial displacement deviation value is obtained by dividing the force deviation value by the stiffness coefficient of the mounting head drive chain; the drive chain stiffness coefficient is determined during the initial calibration of the equipment, for example, 0.2 N / mm, and is recalibrated when the drive chain is replaced or when the cumulative operation exceeds a set cycle (for example, 1000 hours); Since both the servo motor torque constant and the transmission chain stiffness coefficient are positive, and the sign of the spatial displacement deviation value is consistent with the sign of the fluctuation component, then: if the spatial displacement deviation value corresponding to the Z-axis position value is positive, it indicates that there is a bulge on the bottom surface of the component corresponding to the Z-axis position value; if the spatial displacement deviation value corresponding to the Z-axis position value is negative, it indicates that there is a depression on the bottom surface of the component corresponding to the Z-axis position value; the unit of the spatial displacement deviation value is millimeters, and the spatial displacement deviation value refers to the height deviation of the component bottom surface area corresponding to the Z-axis position value from the average plane of the component bottom surface, that is, how many millimeters the area is higher or lower than the average plane; where the average plane of the component bottom surface refers to a virtual plane formed by the average height of all positions on the component bottom surface; S15. Based on the spatial displacement deviation values corresponding to all Z-axis position values, obtain the average bottom surface deviation value, use the negative value of the average bottom surface deviation value as the Z-axis correction amount, and superimpose the Z-axis correction amount onto the Z component of the theoretical positioning coordinates to generate the positioning coordinates before compensation; including: The average bottom surface deviation value is obtained by summing the spatial displacement deviation values corresponding to all Z-axis position values and dividing by the total number of spatial displacement deviation values. The average bottom surface deviation value represents the difference between the average height of the component bottom surface area corresponding to all Z-axis position values and the average plane of the component bottom surface. The average plane of the component bottom surface refers to the virtual plane where the average height of the component bottom surface area corresponding to all Z-axis position values is located. This plane serves as a reference benchmark for judging the height of each area; that is, whether the overall component bottom surface is too thick or too thin, and by how many millimeters. If the average bottom surface deviation is positive, it indicates that the bottom surface of the component is thicker than the average plane. In this case, the Z-axis correction is taken as the negative value of the average bottom surface deviation, so that the placement head presses down a corresponding distance less. If the average bottom surface deviation is negative, it indicates that the bottom surface of the component is thinner than the average plane. In this case, the Z-axis correction is taken as the absolute value of the average bottom surface deviation, so that the placement head presses down a corresponding distance more. The Z component of the theoretical positioning coordinate preset by the placement program is added to the Z-axis correction to generate the Z component of the positioning coordinate before compensation. The X and Y components of the theoretical positioning coordinate remain unchanged. When the Z-axis correction exceeds the set tolerance, such as ±0.05 mm, an alarm is triggered to indicate that the component is abnormal.
[0021] In embodiments of the present invention, the timing sequence of the decreasing air pressure fluctuation of the vacuum nozzle is obtained, and a dynamic approach displacement curve is generated through air film damping inversion; specifically including: S21. During the descent phase of the placement head, monitor the Z-axis position of the placement head in real time. When the placement head enters the effective descent range, activate the air pressure sensor to collect air pressure values and form a descent air pressure fluctuation sequence. The placement head descent phase refers to the process of the placement head descending from the initial height to the theoretical placement height; the effective descent range refers to the range between the critical placement height and the theoretical placement height; the critical placement height is the theoretical placement height plus a set spacing value, such as 0.5 mm; position monitoring is achieved by reading the Z-axis position of the placement head through a grating ruler data acquisition device; the acquisition trigger condition is that the Z-axis position value equals the theoretical placement height plus the set spacing value, and the acquisition stop condition is that the Z-axis position value equals the theoretical placement height; the acquisition period is consistent with the acquisition period of the current feedback unit of the servo driver, such as 0.1 milliseconds; the acquired air pressure values are arranged in the order of acquisition time to form the descent air pressure fluctuation sequence; the average speed of the placement head within the effective descent range is determined by the servo system speed command, such as 25 mm / s; when the speed fluctuation exceeds the set range, such as ±5 mm / s, the corresponding air pressure value is marked as invalid and discarded; when the speed fluctuation is within the set range, the corresponding air pressure value is marked as valid and retained; S22. Perform sliding window averaging filtering on the descending pressure fluctuation time series to remove high-frequency noise components, obtaining the filtered descending pressure sequence; specifically including: The sliding window length is set to one-twentieth of the total number of data points in the descending pressure fluctuation time series; when the total number of data points is 200, the window length is 10 points; for each data point, the arithmetic mean of all data points within half the window length before and after that data point is taken as the filtered value of that data point; the length of the filtered descending pressure sequence is consistent with the original descending pressure fluctuation time series, and the boundary points at both ends are directly taken as the original values; when there is still obvious noise in the filtered pressure sequence, the window length is increased, for example, from one-twentieth to one-tenth; after each increase in the window length, the filtering process is repeated and the noise is re-evaluated until the noise meets the requirements or the window length reaches the upper limit, which is one-fifth of the total number of data points; S23. Perform air film damping inversion processing on the filtered descending air pressure sequence to obtain the air film thickness value corresponding to each air pressure value; The gas film damping inversion process is based on Boyle's law, which states that under isothermal conditions, the pressure of a gas is inversely proportional to its volume. In this step, the gas film pressure is represented by the gas pressure value, and the gas film volume is represented by the product of the gas film thickness and the effective gas film area. When the effective gas film area remains constant, the gas film thickness is inversely proportional to the gas pressure value. The specific inversion process is as follows: the first gas pressure value of the filtered decreasing gas pressure sequence is taken as the reference gas pressure value. The initial gas film thickness corresponding to this reference gas pressure value is equal to the initial gap between the bottom surface of the component and the surface of the pad at the critical mounting height, i.e. 0.5 mm; For each pressure value in the filtered descending pressure sequence, the absolute value of the pressure value is used as the denominator, and the product of the reference pressure value and the initial gas film thickness is used as the numerator. The two are divided to obtain the gas film thickness value at that acquisition time; the gas film thickness value represents the real-time distance between the bottom surface of the component and the surface of the pad at that acquisition time; when the pressure value in the filtered descending pressure sequence approaches zero, the gas film thickness value corresponding to the pressure value approaches the initial gas film thickness, i.e., 0.5 mm. At this time, the gas film thickness value is set to 0.5 mm as the saturation value; S24. Connect all the air film thickness values collected at all times in the order of the descent stroke of the mounting head to generate a dynamic approach displacement curve, and record the convergence value at the end of the curve. The dynamic approach displacement curve has the horizontal axis representing the placement head's descent position, from the critical placement height to the theoretical placement height, and the vertical axis representing the air film thickness at each acquisition moment. Connecting the air film thickness values at all acquisition moments in descending order of the placement head's descent position generates the dynamic approach displacement curve. The vertical axis value of the last data point in the dynamic approach displacement curve (i.e., the data point when the placement head reaches the theoretical placement height) is taken as the final convergence value. When the final convergence value is less than the pre-contact threshold, for example, 0.005 mm, it is determined that the component has prematurely contacted the solder paste, triggering a speed reduction command to decrease the descent speed to a slow descent speed value, for example, 10 mm / s. When the final convergence value is greater than the abnormal gap threshold, for example, 0.05 mm, it is determined that there is an abnormal gap between the component and the pad, triggering a position recalibration command. The pre-contact threshold of 0.005 mm corresponds to the minimum allowable gap between the bottom surface of the component and the surface of the pad. This value is determined by the maximum allowable indentation depth of the LED driver chip pin and the minimum solder paste spreading height in this invention. When the gas film thickness is less than this value, it indicates that the bottom surface of the chip has approached or contacted the solder paste surface. Continuing to descend at the original speed may cause excessive indentation or impact damage to the chip pin. The abnormal gap threshold of 0.05 mm corresponds to the maximum allowable gap between the bottom surface of the component and the surface of the pad. This value is determined by adding the normal upper limit of the solder paste printing thickness fluctuation (e.g., 0.03 mm) to the normal upper limit of the circuit board warpage (e.g., 0.02 mm). When the gas film thickness is greater than this value, it indicates that the circuit board is warped or the solder paste printing thickness exceeds the normal range. Continuing to mount may cause the chip pin to fail to effectively contact the solder paste. The slow descent speed of 10 mm / s is 40% of the original descent speed (e.g., 25 mm / s). This value ensures that the component descends at a low speed when it may contact the solder paste, avoiding impact damage. After the position recalibration command is triggered, the mounting head will return to a safe height and re-execute the alignment and descent process.
[0022] In embodiments of the present invention, a texture image of the area surrounding the solder pad is acquired, and a surface texture feature point matrix is obtained through multi-scale feature point extraction. This surface texture feature point matrix is then analyzed using a dynamic approach displacement curve to obtain an instantaneous solder pad coordinate system. Specifically, this includes: The pick-and-place machine is equipped with a vision imaging module; the vision imaging module is a standard downward imaging system for the pick-and-place machine, which includes an industrial camera and a coaxial light source; the industrial camera uses a complementary metal-oxide-semiconductor image sensor, and the coaxial light source is installed around the lens of the industrial camera. S31. When the placement head descends to the shooting height, the industrial camera is triggered to capture a grayscale image as the texture image around the pad. The image capture height is defined as a set value above the critical mounting height, for example, 1.0 mm, which is 1.5 mm above the theoretical mounting height; the trigger condition is that the Z-axis position value equals the theoretical mounting height plus 1.5 mm; the industrial camera exposure time is set to a fixed value, for example, 5 milliseconds; the light source brightness is automatically adjusted according to the ambient light to ensure that the average grayscale value of the image is within the set target range, for example, 120 to 160; the resolution of the captured image is a set value, for example, 2592×1944 pixels, covering a physical area of approximately 12 mm × 9 mm for the pad area; when the average grayscale value of three consecutive captured images is lower than the set lower limit, for example, 80, the light source brightness is automatically increased by a set step size, for example, 10%. S32. Perform contrast stretching preprocessing on the grayscale image to enhance the contrast between the pad edge and the solder mask layer. The algorithm iterates through all pixels to find the minimum and maximum grayscale values. For each pixel, it maps the original grayscale value to the stretched grayscale value using a linear mapping relationship. The mapping relationship is that the stretched grayscale value equals 255 multiplied by the difference between the original grayscale value and the minimum value, and then divided by the difference between the maximum value and the minimum value. After preprocessing, the grayscale difference between the pad edge and the solder mask layer in the image is amplified, and the overall grayscale value distribution is expanded from the original range to the full range of 0 to 255. When the difference between the maximum and minimum values is less than a set threshold, such as 50, the image contrast is determined to be too low, triggering re-acquisition and increasing the exposure time, for example, to 10 milliseconds. S33. Perform feature point detection on the preprocessed image at three neighborhood radius scales respectively, calculate the local response value of each pixel at each scale and perform non-maximum suppression to screen candidate feature points; Three neighborhood radius scales are set to 1 pixel, 2 pixels, and 3 pixels. For each pixel, the sum of the absolute values of its grayscale differences with all pixels in its neighborhood is calculated as the response value of that pixel at that scale. The larger the response value, the more drastic the grayscale changes around the pixel. Non-maximum suppression is performed on the response value matrix at each scale. The non-maximum suppression process is as follows: for each pixel in the response value matrix, the response value of the pixel is compared with the response values of all pixels in its eight neighborhoods, where the eight neighborhoods refer to the eight adjacent pixels surrounding the target pixel: top, bottom, left, right, upper left, upper right, lower left, and lower right. If the response value of the pixel is less than the response value of any pixel in its eight neighborhoods, the pixel is discarded. If the response value of the pixel is greater than or equal to the response values of all pixels in its eight neighborhoods, the pixel is retained as a candidate feature point. Non-maximum suppression is performed independently at each of the three scales, generating three sets of candidate feature points. S34. Merge and remove duplicate candidate feature points selected from the three scales to generate a surface texture feature point array. Take the union of three sets of candidate feature points; for any two candidate feature points in the union, calculate the difference between their X and Y coordinates in the image coordinate system. If the absolute value of the difference between their X and Y coordinates is less than a set value, such as 1 pixel, then the two candidate feature points are determined to be the same feature point and merged into one point. The coordinates of the merged point are the average of the coordinates of the two candidate feature points. After performing the above merging and deduplication process on all candidate feature points, a surface texture feature point array is obtained. Each surface texture feature point contains its X and Y coordinates in the image coordinate system. When the number of surface texture feature points in the surface texture feature point array is lower than a set lower limit, such as 50, it is determined that the pad area texture is insufficient, triggering visual supplementary lighting adjustment, increasing the light source brightness setting step size, such as 20%, and re-acquiring. S35. Multiply the image coordinates of each surface texture feature point in the surface texture feature point matrix by the pixel equivalent coefficient to convert them into pad physical coordinates. The pixel equivalent coefficient is obtained through camera calibration: a calibration board of known size, such as a grid calibration board with a spacing of 1 mm, is photographed at the camera's working distance; the pixel distance between adjacent grid points is measured, and the pixel equivalent coefficient is obtained by dividing the physical distance of 1 mm by the pixel distance; the pixel equivalent coefficient is determined when the device is first installed, and recalibrated when the camera position or working distance changes; after conversion, the physical coordinate distribution of the pads of the surface texture feature point matrix is obtained, and each surface texture feature point corresponds to a set of pad physical coordinates, including X and Y components; S36. Obtain the terminal convergence value of the dynamic approach displacement curve, calculate the depth weighting factor, and weight the physical coordinates of the pads of the surface texture feature matrix to obtain the weighted origin coordinates. The depth weighting factor is equal to 1 minus the ratio of the terminal convergence value to the total length of the effective descent interval. The total length of the effective descent interval is the critical mounting height minus the theoretical mounting height, for example, 0.5 mm. The physical coordinates X and Y of each surface texture feature point in the surface texture feature matrix are multiplied by the depth weighting factor, summed, and then divided by the total number of surface texture feature points in the surface texture feature matrix to obtain the weighted origin coordinates. The weighted origin coordinates contain two components: weighted origin X and weighted origin Y, representing the estimated position of the pad center in the machine coordinate system. When the depth weighting factor is less than the set lower limit, the spatiotemporal fusion process is terminated and the process returns to the waiting state. The instructions state that the depth weight factor is set to a lower limit of 0.5, corresponding to a convergence value of half the total length of the effective descent interval (i.e., 0.25 mm). At this point, the component is too far from the pad, and the correspondence between the surface texture image and the actual position of the component is unreliable. When the depth weight factor is greater than or equal to 0.5, the spatiotemporal fusion processing continues. The upper limit of the convergence value is determined by the abnormal gap threshold of 0.05 mm. When the convergence value exceeds 0.05 mm, the position recalibration instruction has already been triggered in step S24, and the processing will not be repeated here. If the convergence value is normal (≤0.05 mm), the depth weight factor ranges from 0.9 to 1.0, and the fusion processing is executed normally. S37. Obtain the instantaneous pad coordinate system based on the physical coordinates and weighted origin of the pads using surface texture feature points; including: Calculate the X-direction offset and Y-direction offset of each surface texture feature point relative to the weighted origin. The X-direction offset is the physical coordinate X of the pad of the surface texture feature point minus the X coordinate of the weighted origin, and the Y-direction offset is the physical coordinate Y of the pad of the surface texture feature point minus the Y coordinate of the weighted origin. Construct a 2×2 covariance matrix, the four elements of which are: the sum of squares of the X-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, and the sum of squares of the Y-direction offsets of all surface texture feature points. Calculate the two eigenvectors of the covariance matrix. The first eigenvector corresponds to the larger eigenvalue and points in the direction where the surface texture feature point distribution is most dispersed, i.e., the direction where the pad area extends the longest, and serves as the X-axis direction of the instantaneous pad coordinate system. The second eigenvector corresponds to the smaller eigenvalue and points in the direction perpendicular to the first eigenvector, i.e., the direction where the pad area extends the shortest, and serves as the Y-axis direction of the instantaneous pad coordinate system. The weighted origin is used as the origin of the coordinate system, and the two eigenvector directions are normalized and used as the unit vectors of the coordinate axes; the structure of the instantaneous pad coordinate system includes the origin, the X-axis direction vector, and the Y-axis direction vector.
[0023] In embodiments of the present invention, dynamic deviation separation is performed between the pre-compensation positioning coordinates and the terminal convergence value of the dynamic approach displacement curve to generate impact deviation components and adsorption deviation components; specifically including: S41. When the mounting head reaches the critical mounting height, obtain the Z component of the positioning coordinates before compensation and the convergence value at the end of the dynamic approach displacement curve. When the Z-axis position value of the mounting head is equal to the theoretical mounting height plus 0.5 mm, the Z component of the pre-compensation positioning coordinates and the convergence value at the end of the dynamic approach displacement curve are obtained within the same acquisition cycle. S42. Subtract the convergence value at the end of the dynamic approach displacement curve from the Z component of the positioning coordinates before compensation to obtain the actual height of the pad surface, and then subtract the theoretical mounting height to obtain the comprehensive deviation value. The actual height of the pad surface is equal to the Z component of the positioning coordinates before compensation minus the convergence value at the end of the dynamic approach displacement curve; the comprehensive deviation value is equal to the actual height of the pad surface minus the theoretical mounting height; the theoretical mounting height refers to the target height position that the bottom surface of the component should ultimately reach, preset by the mounting program, represented by the Z-axis coordinate value based on the machine zero point, derived from the circuit board design file, and set by the pick-and-place machine programmer when compiling the mounting program, and is directly read from the mounting program in this invention; a positive comprehensive deviation value indicates that the actual height of the pad surface is higher than the theoretical mounting height, corresponding to a pad that is too high or solder paste that is too thick; a negative comprehensive deviation value indicates that the actual height of the pad surface is lower than the theoretical mounting height, corresponding to a pad that is too low or solder paste that is too thin; when the comprehensive deviation value exceeds the preset tolerance, for example, ±0.1 mm, an alarm is triggered, indicating abnormal circuit board incoming materials or abnormal solder paste printing; S43. In the first acquisition cycle after the placement head reaches the critical placement height, calculate the speed difference between the actual instantaneous speed of the placement head and the commanded speed, and multiply the speed difference by the speed loop response time constant to obtain the impact deviation component. The command speed refers to the target speed value preset by the servo system during the descent phase of the placement head, measured in millimeters per second. Defined in the placement program or servo drive parameters, it is the speed reference value that the servo drive expects to maintain during the placement head's descent. The actual instantaneous speed is calculated as: the Z-axis position value of the placement head at the current acquisition moment minus the Z-axis position value of the placement head at the previous acquisition moment, divided by the acquisition period. The speed difference is the actual instantaneous speed minus the command speed. A positive speed difference indicates overshoot, meaning the actual instantaneous speed of the placement head is faster than the command speed; a negative speed difference indicates lag, meaning the actual instantaneous speed of the placement head is slower than the command speed. The speed loop response time constant is an inherent parameter of the servo drive, for example, 0.025 seconds, measured during the initial debugging of the servo system and remeasured after servo drive replacement or parameter adjustment. When the speed difference exceeds the speed difference setting threshold (e.g., ±10 mm / s) three times consecutively, the servo system self-test process is triggered. The speed difference setting threshold is determined based on the maximum allowable speed deviation of the placement head's mechanical structure. S44. Subtract the impact deviation component from the comprehensive deviation value to obtain the adsorption deviation component; The adsorption deviation component represents the remaining part of the overall deviation value excluding the impact deviation component. It originates from material-related factors such as pad height deviation, solder paste thickness deviation, and surface tension difference between the component bottom surface and the solder paste. When the adsorption deviation component and the impact deviation component have opposite signs, it is determined that the machine dynamic deviation and the material deviation have a mutual cancellation relationship and are handled independently in subsequent corrections.
[0024] In embodiments of the present invention, the impact deviation component is mapped to a servo drive parameter adjustment amount, and the adsorption deviation component is converted into a solder paste thixotropic state evaluation index through solder paste thixotropic evaluation. Then, the servo drive parameter adjustment amount and the solder paste thixotropic state evaluation index are weighted and fused to generate a correction matrix; specifically including: S51. When the placement head presses in solder paste, obtain the impact deviation component and the adsorption deviation component. S52. Map the impact deviation component to the servo drive parameter adjustment amount; The servo drive parameter adjustment includes speed loop gain adjustment and acceleration feedforward coefficient adjustment. Speed loop gain is the response strength parameter of the servo driver controlling the motor speed, and acceleration feedforward coefficient is the torque compensation parameter of the servo system during acceleration and deceleration. Both are inherent adjustable control parameters within the servo driver. Speed loop gain adjustment equals the negative impact deviation component multiplied by the first mapping coefficient, and acceleration feedforward coefficient adjustment equals the negative impact deviation component multiplied by the second mapping coefficient. Both adjustments are percentages relative to the current set value. The speed loop gain adjustment is expressed as a percentage... The ratio is superimposed onto the current set value of the speed loop gain, and the acceleration feedforward coefficient adjustment is superimposed onto the current set value of the acceleration feedforward coefficient in percentage form. The superimposed values are used as the new speed loop gain parameter and the new acceleration feedforward coefficient parameter for the placement head descent speed control, respectively, and are used by the servo driver to control the movement speed of the placement head during the contact and pressing process of the solder paste. Among them, the first mapping coefficient is determined by the ratio of the back electromotive force constant to the torque constant of the servo motor, for example, 0.5 / mm; the second mapping coefficient is determined by multiplying the ratio of the armature inductance to the resistance of the servo motor by the mechanical time constant, for example, 0.3 / mm. A positive impact deviation component indicates that the actual instantaneous speed of the placement head is faster than the commanded speed. In this case, both the speed loop gain adjustment and the acceleration feedforward coefficient adjustment are negative, meaning that reducing the speed loop gain and acceleration feedforward coefficient reduces the servo system response speed, slows down the placement head descent speed, and avoids impact damage to components. A negative impact deviation component indicates that the actual instantaneous speed of the placement head is slower than the commanded speed. In this case, both the speed loop gain adjustment and the acceleration feedforward coefficient adjustment are positive, meaning that increasing the speed loop gain and acceleration feedforward coefficient improves the servo system response speed, speeds up the placement head descent speed, and avoids incomplete placement. The mapping coefficient is measured during the initial debugging of the servo system and recalibrated after the servo motor or servo driver is replaced. S53. Perform thixotropic evaluation on the adsorption deviation component of the solder paste to obtain the thixotropic state evaluation index of the solder paste. Among them, the thixotropic evaluation of solder paste is achieved through a preset adsorption viscosity mapping relationship. The adsorption viscosity mapping relationship refers to the correspondence between the adsorption deviation component and the apparent viscosity of the solder paste. This correspondence is based on the following physical principle: the apparent viscosity of solder paste at a constant shear rate is directly related to its internal structural state. The higher the apparent viscosity of the solder paste, the greater the resistance encountered when the component is pressed in. This resistance is reflected in the adsorption deviation component as an increase in the positive value of the adsorption deviation component. The specific form of the adsorption viscosity mapping relationship is as follows: the solder paste thixotropic state evaluation index is equal to 0.5 plus the adsorption deviation component multiplied by the third conversion coefficient; the third conversion coefficient is determined by dividing the slope of the viscosity-shear rate curve provided by the solder paste manufacturer by the ratio of the inner diameter of the vacuum nozzle to the bottom surface area of the component, for example, 3.5 / mm; the saturation range of the solder paste thixotropic state evaluation index value is limited to 0.1 to 0.9. When the calculated value of the solder paste thixotropic state evaluation index is less than 0.1, it is taken as 0.1, and when the calculated value of the solder paste thixotropic state evaluation index is greater than 0.9, it is taken as 0.9; a solder paste thixotropic state evaluation index value of 0.5 corresponds to standard viscosity. A solder paste thixotropic state evaluation index value greater than 0.5 indicates that the solder paste is too thick, and a solder paste thixotropic state evaluation index value less than 0.5 indicates that the solder paste is too thin; S54. Calculate the correction amount for the end position of the downward pressure and the correction amount for the downward pressure speed curve, combine them into a correction matrix, and send it to the position loop and speed loop of the servo system; specifically including: Subtract 0.5 from the solder paste thixotropic state evaluation index to obtain the difference. Multiply this difference by the base pressure depth and then by 2 to obtain the pressure endpoint position correction amount. The base pressure depth is given by the component manufacturer in the component datasheet, for example, 0.06 mm. The positive and negative meanings of the pressure endpoint position correction amount are as follows: a positive value indicates that the placement head needs to descend a certain distance further than the theoretical pressure endpoint position, and a negative value indicates that the placement head needs to descend a certain distance less than the theoretical pressure endpoint position. Multiply the absolute value of the velocity loop gain adjustment by 0.5, and add the absolute value of the acceleration feedforward coefficient adjustment multiplied by 0.5 to obtain the downward velocity curve correction amount; the downward velocity curve correction amount is a dimensionless percentage, representing the reduction in the slope of the velocity curve, for example, 0.03 represents a reduction of 3%; The correction matrix is a 2-row, 1-column matrix. The first row of the correction matrix is the correction amount for the downward pressure endpoint position, in millimeters. The second row of the correction matrix is the correction amount for the downward pressure speed curve, which is a dimensionless percentage. The first row of the correction matrix is sent to the position loop of the servo system as the new downward pressure endpoint position offset, and the second row of the correction matrix is sent to the speed loop of the servo system as the speed curve slope adjustment ratio.
[0025] In embodiments of the present invention, the rising air pressure fluctuation time sequence of the vacuum nozzle is obtained, and the equivalent air tightness feature value after mounting is obtained through air tightness feature extraction. This value is then compared with the expected feature range derived from the correction matrix. Deviations are marked as suspected unstable solder joints, and the correction matrix parameters of adjacent subsequent mounting points are adjusted retrospectively. Specifically, this includes: S61. During the rising phase of the placement head, the internal air pressure value of the vacuum nozzle is collected to form the rising air pressure fluctuation sequence; specifically: The placement head rising phase refers to the process where the placement head begins to rise and detach from the pad after reaching the corrected downward pressure endpoint and maintaining solder paste wetting time. After the placement head reaches the corrected downward pressure endpoint, it maintains this position for a preset solder paste wetting time, such as 50 milliseconds. The solder paste wetting time is determined by the rheological properties parameters provided by the solder paste manufacturer. When the ambient temperature deviates from the standard value, such as 25°C, the solder paste wetting time is adjusted according to the temperature compensation coefficient, such as ±10 milliseconds for every ±5°C. After the solder paste wetting time is completed, the placement head is driven to rise at a preset speed, such as 50 mm / s. At the same time as the rise starts, the air pressure sensor is activated to collect air pressure values, with a collection period set to 0.1 milliseconds. The collection is sustained when the air pressure value returns to the ambient atmospheric pressure, i.e., gauge pressure 0 kPa, and the fluctuation amplitude of 10 consecutive collection points is less than the set tolerance, such as ±0.5 kPa. All collected air pressure values are arranged in chronological order to form the rising air pressure fluctuation sequence. S62. Extract the recovery slope value and steady-state offset value from the time series of rising air pressure fluctuations; The method for extracting the recovery slope value is as follows: starting from the lowest value of the air pressure in the rising air pressure fluctuation sequence, all sampling points are traversed, and the slope between adjacent sampling points is calculated; the slope is the air pressure value of the next sampling point minus the air pressure value of the previous sampling point, and then divided by the sampling interval of 0.1 milliseconds; the maximum value among all slope values is taken as the recovery slope value; the recovery slope value represents the degree of detachment between the vacuum nozzle and the component; The method for extracting the steady-state offset value is as follows: take the pressure values of 10 consecutive collection points at the end of the rising air pressure fluctuation time series and calculate the arithmetic mean; subtract the ambient atmospheric pressure, i.e., 0 kPa gauge pressure, from the arithmetic mean and take the absolute value of the difference as the steady-state offset value; the steady-state offset value represents whether there is residual negative pressure inside the vacuum nozzle, corresponding to partial blockage of the vacuum nozzle or incomplete detachment of components. S63. Divide the recovery slope value by the reference recovery slope value to obtain the normalized recovery slope value, and divide the steady-state offset value by the reference steady-state offset value to obtain the normalized steady-state offset value. The baseline recovery slope value is determined during equipment calibration: a standard recovery test is performed with the vacuum nozzle fully exposed to the atmosphere and the air path clean, recovering from -85 kPa to 0 kPa, and the maximum slope value is recorded; the baseline recovery slope value is set to 95% of the measured value, with a 5% tolerance; the baseline steady-state offset value is set to 5% of the absolute value of the lower limit of the barometric pressure sensor range; the barometric pressure sensor range is -101 kPa to 0 kPa, the absolute value of the lower limit is 101 kPa, and 5% of it is 5.05 kPa; the baseline steady-state offset value represents the critical value of steady-state offset when the vacuum nozzle is completely blocked; the baseline recovery slope value or the baseline steady-state offset value is re-measured when it changes due to air path maintenance or component replacement; S64. Calculate the equivalent airtightness characteristic value after mounting; The equivalent airtightness characteristic value after mounting is equal to the normalized recovery slope value multiplied by the first weighting coefficient 0.6, plus the normalized steady-state offset value multiplied by the second weighting coefficient 0.4. The sum of the first and second weighting coefficients is 1. The first weighting coefficient is greater than the second weighting coefficient, reflecting that the recovery slope value contributes more to the judgment of mounting quality. The equivalent airtightness characteristic value after mounting is between 0 and 1. The higher the value of the equivalent airtightness characteristic value after mounting, the better the mounting quality. After the calculation is completed, the equivalent airtightness characteristic value after mounting is associated with the mounting point identifier. The first weighting coefficient of 0.6 and the second weighting coefficient of 0.4 are determined based on the physical characteristics of the vacuum nozzle air path system: the recovery slope value directly reflects the degree of smoothness of the vacuum nozzle's separation from the component. When the component is well mounted, there is no residual adsorption force between the vacuum nozzle and the component, the air pressure recovers quickly, and the recovery slope value is high; the steady-state offset value reflects whether there is residual negative pressure inside the vacuum nozzle. This value is greatly affected by the inherent leakage of the air path system and the zero drift of the sensor, and its signal-to-noise ratio is lower than that of the recovery slope value; therefore, the recovery slope value contributes more to the judgment of the mounting quality, so it is given a higher weight of 0.6, and the steady-state offset value is given a lower weight of 0.4; this weighting is determined based on the measured signal-to-noise ratio of the air path system during the initial calibration of the equipment, and is recalibrated after the air path system is replaced or repaired. S65. Based on the correction matrix, extract the correction amount of the downward pressure endpoint position and the correction amount of the downward pressure speed curve, and derive the lower limit and upper limit of the expected characteristic interval. The expected lower limit of the characteristic interval is equal to 0.5 minus the absolute value of the correction amount for the pressure endpoint position divided by the first reference coefficient, and then minus the correction amount for the pressure speed curve divided by the second reference coefficient; the first reference coefficient is, for example, 0.05 mm, and the second reference coefficient is, for example, 0.1. The expected upper limit of the characteristic interval is equal to 0.5 plus the absolute value of the correction amount for the pressure endpoint position divided by the first reference coefficient, plus the correction amount for the pressure speed curve divided by the second reference coefficient; the first and second reference coefficients are determined by the viscosity-temperature relationship curve of the solder paste material and the flow characteristics of the vacuum nozzle air path, and are measured during the initial calibration of the equipment; the expected lower limit of the characteristic interval is saturated and limited to 0.1, and when the expected lower limit of the characteristic interval is less than 0.1, it is taken as 0.1; the expected upper limit of the characteristic interval is saturated and limited to 0.9, and when the expected upper limit of the characteristic interval is greater than 0.9, it is taken as 0.9. S66. Compare the equivalent airtightness characteristic value after mounting with the expected characteristic range. When the equivalent airtightness characteristic value after mounting is lower than the lower limit of the expected characteristic range or higher than the upper limit of the expected characteristic range, mark the current mounting point as a suspected unstable solder joint. When the equivalent airtightness characteristic value after mounting is greater than or equal to the lower limit of the expected characteristic range and less than or equal to the upper limit of the expected characteristic range, the mounting quality is considered normal and no marking operation is performed; when the equivalent airtightness characteristic value after mounting is less than the lower limit of the expected characteristic range or greater than the upper limit of the expected characteristic range, the mounting quality is considered abnormal and marked as a suspected unstable solder joint; the marking information includes the mounting point identifier, the mounting point position coordinates (X, Y values), the equivalent airtightness characteristic value after mounting, the upper limit of the expected characteristic range, the lower limit of the expected characteristic range, and the current correction matrix parameters; S67. When a suspected unstable solder joint is marked, calculate the Euclidean distance between the suspected unstable solder joint and subsequent unmounted joints. For subsequent mounting joints whose Euclidean distance is less than the preset neighborhood radius, determine the retrospective adjustment range according to the Euclidean distance in segments. Extract the location coordinates of suspected unstable solder joints from the marking information; obtain the location coordinates of all unmounted points on the current circuit board; calculate the Euclidean distance between the suspected unstable solder joints and each unmounted point using Euclidean distance; the preset neighborhood radius is 1.5 times the average spacing of the mounting points; the average spacing of the mounting points is obtained by averaging the pairwise distances calculated from the location coordinates of all mounting points on the current circuit board; for subsequent mounting points whose Euclidean distance value is less than the neighborhood radius: the backtracking adjustment is 100% when the Euclidean distance value is less than one-third of the neighborhood radius, 60% when the Euclidean distance value is between one-third and two-thirds of the neighborhood radius, and 30% when the Euclidean distance value is between two-thirds and one times the neighborhood radius; S68. Multiply the correction matrix parameters of the current solder joint by the backtracking adjustment range and then add them to the original correction matrix of the subsequent placement points to generate the updated correction matrix of the subsequent placement points. The correction matrix parameter of the current solder joint is multiplied by the retrospective adjustment range of the subsequent placement point, and then added item by item to the original correction matrix parameter of the subsequent placement point to generate the updated correction matrix of the subsequent placement point. The updated correction matrix is stored in the machine's memory address, overwriting the original value, and is read by the subsequent placement point when it is placed. When the same subsequent placement point is affected by multiple suspected unstable solder joints, the retrospective adjustment amount of the suspected unstable solder joint with the closest Euclidean distance is taken, and multiple retrospective adjustment amounts are not superimposed. After the correlational retrospective adjustment is completed, the marking information of the current suspected unstable solder joint is written to the equipment log database for offline quality traceability.
[0026] In the embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the method in this embodiment according to actual needs.
[0027] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A high-precision patching management method for LED display screen circuit boards based on patch machine positioning, characterized in that, include: S1. Obtain the driving current timing and Z-axis position value when the placement head picks up the component, and generate the positioning coordinates before compensation by time domain decomposition. S2. Obtain the timing sequence of the descending air pressure fluctuation of the vacuum nozzle, and generate a dynamic approach displacement curve through air film damping inversion; S3. Obtain the texture image around the pad, extract the surface texture feature point matrix through multi-scale feature point extraction, and analyze the surface texture feature point matrix in combination with the dynamic approach displacement curve to obtain the instantaneous pad coordinate system. S4. Perform dynamic deviation separation between the pre-compensation positioning coordinates and the final convergence value of the dynamic approach displacement curve to generate impact deviation component and adsorption deviation component. S5. Map the impact deviation component to the servo drive parameter adjustment amount, and convert the adsorption deviation component into the solder paste thixotropic state evaluation index through solder paste thixotropic evaluation. Then, weight and fuse the servo drive parameter adjustment amount and the solder paste thixotropic state evaluation index to generate a correction matrix. S6. Obtain the rising air pressure fluctuation time sequence of the vacuum nozzle, extract the equivalent air tightness feature value after mounting through air tightness feature extraction, compare the deviation with the expected feature range derived from the correction matrix, mark the deviation as a suspected unstable solder joint, and perform correlation backtracking adjustment on the correction matrix parameters of adjacent subsequent mounting points.
2. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 1, characterized in that, The pick-and-place machine includes a three-dimensionally movable placement head, which picks up components and moves them above target pads on a circuit board for placement. The placement head is driven by a servo motor equipped with a servo driver. The servo driver integrates a current feedback unit to output the real-time drive current value of the servo motor at a fixed acquisition cycle. A linear encoder is mounted on the Z-axis of the placement head to measure its Z-axis position. A vacuum nozzle is mounted at the end of the placement head to pick up and release components. The vacuum nozzle is connected to a vacuum generator via an air path, in which a pressure sensor is installed to detect the internal air pressure. The pick-and-place machine is equipped with a vision imaging module, a standard downward imaging system that includes an industrial camera and a coaxial light source.
3. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S1 includes: When the placement head picks up a component, obtain the timing of the drive current of the placement head; The driving current timing sequence is decomposed in the time domain to obtain the timing sequence of the fluctuation component corresponding to the driving current timing sequence. The Z-axis position value of the mounting head is acquired synchronously, and the position fluctuation mapping sequence is obtained based on the fluctuation component timing and the Z-axis position value of the mounting head. Each fluctuation component contained in the position fluctuation mapping sequence is converted into a spatial displacement deviation value corresponding to the Z-axis position value; The average bottom surface deviation value is obtained based on the spatial displacement deviation value corresponding to all Z-axis position values. The negative value of the average bottom surface deviation value is used as the Z-axis correction amount. The Z-axis correction amount is superimposed on the Z component of the theoretical positioning coordinates to generate the positioning coordinates before compensation.
4. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S2 includes: During the descent phase of the placement head, the Z-axis position of the placement head is monitored in real time. When the placement head enters the effective descent range, the air pressure sensor is activated to collect air pressure values, forming a descent air pressure fluctuation sequence. The time series of descending air pressure fluctuations is subjected to sliding window averaging filtering to remove high-frequency noise components, resulting in a filtered descending air pressure sequence. Perform air film damping inversion processing on the filtered descending air pressure sequence to obtain the air film thickness value corresponding to each air pressure value; Connect all the air film thickness values collected at all times in the order of the mounting head's descent stroke to generate a dynamic approach displacement curve, and record the convergence value at the end of the curve.
5. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 4, characterized in that, The process of performing air film damping inversion on the filtered descending air pressure sequence to obtain the air film thickness value corresponding to each air pressure value includes: The first pressure value in the filtered descending pressure sequence is taken as the reference pressure value. The initial air film thickness corresponding to this reference pressure value is equal to the initial gap between the bottom surface of the component and the surface of the pad at the critical mounting height. For each pressure value in the filtered descending pressure sequence, the absolute value of the pressure value is used as the denominator, and the product of the reference pressure value and the initial air film thickness is used as the numerator. The two are divided to obtain the air film thickness value at the acquisition time. When the pressure value in the filtered descending pressure sequence approaches zero, the air film thickness value corresponding to the pressure value approaches the initial air film thickness, and the air film thickness value is set as the saturation value.
6. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S3 includes: When the placement head descends to the shooting height, the industrial camera is triggered to capture a grayscale image as the texture image around the pads; Perform contrast stretching preprocessing on the grayscale image to enhance the contrast between the pad edges and the solder mask layer; Feature point detection is performed on the preprocessed image at three neighborhood radius scales. The local response value of each pixel is calculated at each scale and non-maximum suppression is performed to screen candidate feature points. The candidate feature points selected from the three scales are merged and deduplicated to generate a surface texture feature point array. Multiply the image coordinates of each surface texture feature point in the surface texture feature point array by the pixel equivalent coefficient to convert them into physical coordinates of the pad. Obtain the terminal convergence value of the dynamic approach displacement curve, calculate the depth weighting factor, and weight the physical coordinates of the pads of the surface texture feature matrix to obtain the weighted origin coordinates. The instantaneous pad coordinate system is obtained based on the physical coordinates of the pads and the weighted origin of the surface texture feature matrix.
7. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 6, characterized in that, The process of obtaining the instantaneous pad coordinate system based on the physical coordinates of the pads using the surface texture feature point matrix and the weighted origin includes: Calculate the X-direction and Y-direction offsets of each surface texture feature point relative to the weighted origin; construct a 2×2 covariance matrix, the four elements of which are the sum of the squares of the X-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, the sum of the products of the X-direction and Y-direction offsets of all surface texture feature points, and the sum of the squares of the Y-direction offsets of all surface texture feature points; calculate the two eigenvectors of the covariance matrix, the first eigenvector corresponding to the larger eigenvalue is used as the X-axis direction of the instantaneous pad coordinate system; the second eigenvector corresponding to the smaller eigenvalue is used as the Y-axis direction of the instantaneous pad coordinate system; use the weighted origin as the origin of the coordinate system, and normalize the two eigenvector directions to use them as the unit vectors of the coordinate axes.
8. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S4 includes: When the placement head reaches the critical placement height, obtain the Z component of the positioning coordinates before compensation and the convergence value at the end of the dynamic approach displacement curve; The actual height of the pad surface is obtained by subtracting the convergence value at the end of the dynamic approach displacement curve from the Z component of the positioning coordinates before compensation, and then subtracting the theoretical mounting height to obtain the comprehensive deviation value. In the first acquisition cycle after the placement head reaches the critical placement height, the speed difference between the actual instantaneous speed of the placement head and the commanded speed is calculated, and the speed difference is multiplied by the speed loop response time constant to obtain the impact deviation component. The adsorption deviation component is obtained by subtracting the impact deviation component from the overall deviation value.
9. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S5 includes: When the placement head presses in the solder paste, the impact deviation component and the adsorption deviation component are obtained; Map the impact deviation component to the servo drive parameter adjustment amount; Thixotropic evaluation of solder paste was performed on the adsorption deviation component to obtain the thixotropic state evaluation index of solder paste. Calculate the correction amount for the end position of the downward pressure and the correction amount for the downward pressure speed curve, combine them into a correction matrix, and send it to the position loop and speed loop of the servo system.
10. The high-precision patching management method for the LED display screen circuit board based on the patch machine positioning according to claim 2, characterized in that, S6 includes: During the rising phase of the placement head, the air pressure value inside the vacuum nozzle is collected to form the rising air pressure fluctuation sequence. Extract the recovery slope value and steady-state offset value from the time series of rising air pressure fluctuations; The normalized recovery slope value is obtained by dividing the recovery slope value by the baseline recovery slope value, and the normalized steady-state offset value is obtained by dividing the steady-state offset value by the baseline steady-state offset value. Calculate the equivalent airtightness characteristic value after mounting; Based on the correction matrix, the correction amount of the downward pressure endpoint position and the correction amount of the downward pressure speed curve are extracted, and the lower limit and upper limit of the expected characteristic interval are derived. The equivalent airtightness characteristic value after mounting is compared with the expected characteristic range. When the equivalent airtightness characteristic value after mounting is lower than the lower limit of the expected characteristic range or higher than the upper limit of the expected characteristic range, the current mounting point is marked as a suspected unstable solder joint. When a suspected unstable solder joint is marked, the Euclidean distance between the suspected unstable solder joint and subsequent unattached joints is calculated. For subsequent joints whose Euclidean distance is less than the preset neighborhood radius, the backtracking adjustment range is determined by segmenting them according to the Euclidean distance. The correction matrix parameters of the current solder joint are multiplied by the backtracking adjustment range and then superimposed onto the original correction matrix of the subsequent placement points to generate the updated correction matrix of the subsequent placement points.
