Circuit board detection system based on traceability code and control method

Abstract

The present application relates to a circuit board detection system and control method based on traceability code, the system comprises a business center, a code printing unit, a segmented conveying unit, a multi-station visual acquisition unit, a process parameter acquisition unit, a real-time error prevention and reporting unit, a manual review unit, a defective product partition processing unit, a function test unit and a cloud traceability repository, the segmented conveying unit is connected by multiple conveyors, and a visual acquisition unit is installed above each section, the visual acquisition unit includes an upper fixed camera and a lower movable guide rail camera and synchronously acquires double-sided images of the circuit board, the process parameter acquisition unit binds equipment parameters and traceability codes, the real-time error prevention and reporting unit has a multi-level weight determination model, the manual review unit completes false detection review, the defective product partition processing unit sets up an independent shunt channel, the function test unit acquires data, and the cloud traceability repository archives data according to traceability codes. The present application realizes digital management and control of the whole life cycle of the circuit board, and improves defect detection rate and traceability efficiency.

Description

Technical Field

[0001] This invention relates to the field of circuit board manufacturing technology, and in particular to a circuit board inspection system and control method based on traceability codes. Background Technology

[0002] With the intensification of market competition and the acceleration of the informatization process in the manufacturing industry, the market has placed higher demands on the control of product quality and production. Discrete manufacturing products have complex structures and many uncertainties in the production process, making it difficult to collect production data and diagnose quality problems. As an important electronic connector, the quality of printed circuit boards (PCBs) is related to the quality of the entire electronic product. In order to improve the level of product quality control, it is necessary to record production information to achieve full traceability and quality control. In the current PCB production process, some companies only print traceability codes on the surface of the PCB and enter basic information such as the production date. Key data such as production parameters, inspection images and equipment operating status of each process are not systematically collected and linked. This makes it impossible to accurately locate the specific process and technology link where the defect occurred when product quality problems occur. The traceability efficiency is low and it is difficult to meet the needs of full life cycle quality control. Summary of the Invention

[0003] The purpose of this invention is to provide a circuit board inspection system and control method based on traceability codes, to solve the technical problems of incomplete production data acquisition, difficulty in defect location, and low traceability efficiency in the prior art. The above-mentioned technical objective of this invention is achieved through the following technical solution: The circuit board inspection system based on traceability codes includes a business platform, a coding unit, a segmented conveyor unit, a multi-station visual acquisition unit, a process parameter acquisition unit, a real-time error prevention and reporting unit, a manual review unit, a defective product zoning processing unit, a functional testing unit, and a cloud-based traceability storage repository. The segmented conveyor unit consists of multiple independently controllable belt conveyors or roller conveyors connected in series. Single conveyor units are arranged at the junctions of adjacent processes. Each conveyor unit is independently equipped with the visual acquisition unit. The business platform is connected to the coding unit, segmented conveyor unit, visual acquisition unit, and process parameter acquisition unit. The circuit board includes a data acquisition unit, a real-time error prevention and reporting unit, a manual review unit, a functional testing unit, and a cloud-based traceability storage unit connected by electrical communication. The coding unit receives production work orders from the business platform, generates unique traceability codes in batches, and prints them onto the circuit board surface. The traceability codes are barcodes or QR codes. The vision acquisition unit includes an upper industrial camera, a lower mobile industrial camera, a guide rail adjustment module, and a station supplementary light. The upper industrial camera is fixed to capture images of the top surface of the circuit board, while the lower industrial camera, mounted on a mobile guide rail, achieves adaptive adjustment of its horizontal and vertical positions and, in conjunction with the supplementary light, simultaneously captures images of the bottom surface of the circuit board. The process parameter acquisition unit is independently configured for each workstation and is used to collect the operating parameters of the equipment and processing parameters of each process, and is bound to the corresponding circuit board traceability code. The real-time error prevention and reporting unit has a built-in image weight judgment algorithm and compares the circuit board image uploaded by the visual acquisition unit with the standard sample in real time. If the judgment threshold is exceeded, an error signal is output. The manual review unit receives the abnormal work orders pushed by the real-time error prevention and reporting unit, completes the manual review of visual mis-inspection, and synchronously sends the review results back to the business platform. The defective product zoning processing unit is set according to appearance defects, misplacement, welding defects, and electrical function defects. The system features an independent distribution channel. The functional testing unit collects electrical data on circuit board voltage, current, and conductivity performance, and uploads it to the cloud-based traceability repository by binding it with the traceability code. The cloud-based traceability repository establishes an independent data archive for each circuit board traceability code and archives images of the entire process, equipment parameters, test records, review records, and functional test data. It also supports full-process reverse traceability queries. The system also includes a circuit board repair workbench with a multi-directional elastic clamping mechanism. The elastic clamping mechanism is adapted to clamp and fix multi-layer circuit boards and circuit boards with irregular edges. The clamping mechanism includes anti-slip silicone pads and an angle adjustment shaft.

[0004] In a further embodiment, the coding unit includes a coding workbench for irregularly shaped parts, with flexible rubber positioning blocks pasted on the surface of the workbench. A dust removal cover is provided on the left side of the workbench, which is connected to a negative pressure dust extraction fan and used to adsorb coding dust. A horizontal adjustment screw is provided at the bottom of the workbench to adapt to the positioning of irregularly shaped circuit boards of different sizes.

[0005] In a further embodiment, the system also includes a circuit board welding fixture. The fixture is equipped with a liftable data interface pressing component. The pressing component includes two sets of symmetrical pressing blocks. The pressing blocks press down to fix the circuit board data interface through a spring buffer structure and prevent the interface from deforming and warping under the high temperature of welding. The side of the fixture integrates a data transmission interface and collects welding temperature data in real time and synchronizes it to the business middle platform.

[0006] In a further embodiment, the system also includes a circuit board welding blower module. The blower module is equipped with a magnetic windproof cover. The module housing has a built-in permanent magnet. The windproof cover can be quickly installed and removed by magnetic attraction. The windproof cover is equipped with a flow guide grid inside and constrains the hot air to act in the direction of the solder joint area to avoid the surrounding components from being overheated and having poor soldering.

[0007] In a further embodiment, the system also includes a circuit board positioning detection device. A transparent pressing plate is provided on the top of the detection device. The pressing plate controls the total pressing stroke by adjusting bolts. A liftable black positioning plate is provided at the bottom of the device. Positioning pins are distributed at the lower end of the positioning plate and are adapted to support the positioning of circuit boards of different thicknesses.

[0008] In a further embodiment, the system is also equipped with a circuit board inspection device with display function. The inspection device integrates an automated lifting module, which is driven by a servo motor. The front of the device is equipped with a touch screen display, which counts the number of products inspected on the production line in real time and visually displays the voltage and current detection values ​​of each circuit board. Data exceeding the standard is highlighted in a pop-up window.

[0009] In a further embodiment, the cloud-based traceability repository stores data in layers and is divided into a work order basic information layer, a process image layer, a process parameter layer, a defective product handling layer, and a finished product testing layer. The business middle platform is configured with a search engine and supports multi-dimensional retrieval of full-process data by traceability code, work order number, defect type, and production time period.

[0010] In a further embodiment, the real-time error prevention and reporting unit has a built-in multi-level weight judgment model. The first level is the standard image pixel matching weight, the second level is the component position offset weight, and the third level is the solder joint grayscale feature weight. If any of the three weights exceeds a preset threshold, an error signal is triggered.

[0011] This invention also provides a detection method for a circuit board inspection system based on traceability codes, comprising the following steps: S1, batch generation of work order traceability codes and circuit board coding: the business platform reads standardized production work orders and generates an independent traceability code for each circuit board based on the unique work order number, and sends it to the coding unit. The coding machine prints the traceability code on the blank area of ​​the circuit board. After coding is completed, the segmented conveying unit sends the circuit board into the first production process; S2, synchronous acquisition of images and process parameters throughout the process: as the circuit board flows through each segmented conveying unit, the vision acquisition unit synchronously acquires the top and bottom surfaces of the circuit board. The system provides high-definition images and process parameter acquisition. The system collects the current workstation's equipment speed, temperature, pressure, and welding time parameters, and all data is bound to the circuit board traceability code and uploaded to the business platform. S3: Real-time visual weighting and error prevention warning. The real-time error prevention and reporting unit retrieves the circuit board image bound to the traceability code and compares it with a multi-level weighting judgment model and standard samples. If placement misalignment, missing components, or solder joint defects are detected, an audible and visual error is immediately output, and the conveyor unit suspends transport. S4: Manual review and classification of abnormal work orders. The manual review unit receives error work orders, and operators retrieve images of the entire process. Process parameters are manually checked and distinguished between genuine defects and visual misdetections. Misdetected products are marked and returned to the production line, while genuine defects are pushed to the defective product zoning processing unit; S5, defective products are sorted and processed in separate channels. Four types of defective products—appearance defects, component misalignment, welding defects, and electrical faults—enter independent sorting channels. Process engineers and field application engineers conduct defect analysis and record the defect causes, binding traceability codes and storing them in the cloud; S6, semi-finished product electrical function testing: circuit boards that pass the appearance re-inspection flow into the functional testing unit and are tested for various electrical indicators, including voltage, etc. Current and continuity data are linked to traceability codes. Defective products are returned to the repair station and re-tested after repair. S7: Final inspection and traceability archives are filed. Circuit boards that pass the final inspection are packaged with anti-static material and sampled for shipment. All process images, equipment parameters, test records, and defect handling records are archived in the cloud traceability repository, and each circuit board forms a complete traceability archive. S8: Reverse traceability query. When a finished product has an after-sales quality problem, the circuit board traceability code is entered to retrieve the entire process of production, testing, and defect handling data and locate the defect-causing process, equipment, and operator.

[0012] In a further embodiment, in step S2, the lower movable industrial camera automatically adjusts the guide rail position according to the length and width of the circuit board, the fill light adjusts its brightness synchronously with the camera, the timestamps of the bottom image and the top image are completely synchronized, and the upper and lower images are bound to the same traceability code in a one-to-one correspondence. In step S3, the multi-level weight judgment model presets adjustable thresholds and sets three sets of differentiated weight threshold schemes for precision circuit boards, ordinary consumer electronics circuit boards and industrial control circuit boards. The business platform automatically switches the judgment standard according to the product type of the work order. In step S5, after the defective product is disposed of, the system automatically generates a rework work order. The rework work order carries the traceability code of the original circuit board, and the rework process data is superimposed on the original traceability file to form a secondary complete traceability record for defect repair. In step S8, during reverse traceability, the system automatically generates a visualized full-process timeline and displays each step's image, equipment operating parameters and detection judgment results in the order of the process, and supports image magnification, export of process parameter curves and one-click printing of defect reports.

[0013] In summary, the present invention has the following beneficial effects: 1. Through a segmented conveying unit combined with a multi-station visual acquisition unit and a process parameter acquisition unit, the entire process of circuit board image and process parameter archiving can be achieved. The structure of simultaneously acquiring double-sided images with a fixed upper industrial camera and a mobile lower industrial camera, combined with a multi-level weighted judgment model, can eliminate missed defects at bottom solder joints and improve the overall defect detection rate. The structure of setting a tolerance threshold to filter invalid interference through a real-time error prevention and reporting unit, combined with a standardized review process through a manual review unit, can reduce the workload of manual review and form a closed loop of dual interception between machine initial judgment and manual review. The structure of setting four independent diversion channels through a defective product zoning processing unit, along with engineer analysis and rework order traceability, can achieve closed-loop management of defective products throughout the entire process. The structure of hierarchical archiving in a cloud-based traceability repository, combined with a multi-dimensional search engine, allows for one-click retrieval of the entire process data by inputting the traceability code. Attached Figure Description

[0014] Figure 1 This is a flowchart of the circuit board testing system based on traceability codes according to the present invention.

[0015] Figure 2 This is a flowchart of the control method for the circuit board testing system based on traceability codes according to the present invention. Detailed Implementation

[0016] The present invention will be further described in detail below with reference to the accompanying drawings.

[0017] It should be noted that in the description of this invention, any descriptions of orientation, such as up, down, front, back, left, right, etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings. They are only for the purpose of facilitating the description of this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed or operated in a specific orientation, and should not be construed as a limitation of this invention.

[0018] Example 1: like Figures 1 to 2 As shown, a traceability code-based circuit board inspection system and control method are presented. A complete and practical traceability code-based circuit board inspection system is adapted to the entire process management of printed circuit boards from warehousing, placement, soldering, inspection to finished product delivery. The overall system architecture consists of a business platform, a coding unit, a segmented conveying unit, a multi-station visual acquisition unit, a process parameter acquisition unit, a real-time error prevention and reporting unit, a manual review unit, a defective product zoning processing unit, a functional testing unit, a cloud-based traceability storage unit, and seven types of supporting dedicated tooling equipment. All hardware units establish bidirectional communication with the business platform via industrial Ethernet, and the data transmission latency can be controlled within 200 milliseconds. The business platform utilizes an industrial-grade control host equipped with a Core i7 processor, 32GB of RAM, and a 4TB high-speed solid-state drive, along with a 23-inch industrial touchscreen display. It comes pre-installed with customized printed circuit board (PCB) traceability management software, which includes modules for work order management, traceability code generation, image acquisition management, process parameter monitoring, defective product traceability analysis, and cloud data synchronization. The work order management module supports importing bills of materials, engineering change orders, production sample drawings, and process standard documents. Process engineers integrate these documents into standardized production process documents and distribute them to all workstations on the production line. The traceability code generation module incorporates an encryption algorithm and assigns a unique 24-digit traceability code to each PCB under each production work order. This code includes five segments: work order number, production date, production line number, PCB serial number, and check digit, and can be converted to either a QR code or barcode format. The image acquisition management module receives high-definition images uploaded in real-time from the visual acquisition units at all workstations. The system features images with a resolution of 4,000 x 3,000 pixels, automatically bound to corresponding traceability codes and collection timestamps. Images are cached locally for seven days before being automatically synchronized to the cloud traceability repository and releasing local storage space. The process parameter monitoring module reads real-time operating data from each workstation and supports real-time curve display of over twenty process parameters, including temperature, pressure, welding time, mounting speed, and etching rate. When parameters exceed the standard range, the business platform automatically pops up a warning and records the abnormal time point. The defective product traceability analysis module calculates the percentage of various defects daily, weekly, and monthly and automatically generates defect distribution reports to locate workstations and equipment models that generate high-frequency defects. The cloud data synchronization module adopts a breakpoint resume mechanism and caches all data locally when the production line is offline, automatically uploading it to the cloud in batches after the network is restored. The business platform also features a hierarchical permission system, where operators only have permission to view workstation data, process engineers have permission to modify parameters and export reports, and administrators have the highest permissions for issuing work orders, deleting cloud data, and configuring the system.

[0019] The segmented conveyor unit production line consists of twelve independent conveyor units, covering all twelve processes including PCB incoming material coding, surface mount technology (SMT) placement, reflow soldering, component insertion, wave soldering, lead trimming, initial visual inspection, statistical process control, functional testing, final inspection, packaging, and outgoing sampling inspection. Each conveyor unit is 1.2 meters long and 0.6 meters wide, accommodating PCBs ranging from 500mm x 400mm to 50mm x 30mm. The main frame of each conveyor unit is constructed of aluminum alloy profiles and equipped with four anti-slip support feet at the bottom. The height of the support feet is adjustable by ±100mm to accommodate different workstation heights. Light-load workstations such as SMT placement and visual inspection use food-grade polyurethane belt conveyors with stepless speed regulation from 0.1 to 0.5 meters per second. Heavy-load workstations such as wave soldering and component insertion use stainless steel roller conveyors with coated surfaces. High-temperature resistant silicone anti-slip sleeves are used. Each conveyor unit is independently equipped with a servo drive motor and a photoelectric sensor. The photoelectric sensor is installed at the inlet and outlet of the conveyor unit. When the printed circuit board enters the conveying area, the sensor triggers a signal and simultaneously wakes up the vision acquisition unit and process parameter acquisition unit of this station to start data acquisition. When the printed circuit board completely leaves the conveyor unit, the sensor sends a completion signal and the business platform marks the end of the acquisition process of this station. A 150 mm operating gap is reserved between two adjacent conveyor units for manual loading and unloading of boards and diversion of defective products. A liftable diversion baffle is set at the gap. The baffle is driven by a small cylinder and automatically lifts after receiving a signal from the real-time error prevention and reporting unit. Defective printed circuit boards are introduced into the partition processing channel, while qualified printed circuit boards are normally transferred to the next process. All servo motor operating parameters, start and stop signals and conveying speeds of all conveyor units are uploaded to the business platform in real time and bound to the traceability code of the printed circuit board flowing through the unit.

[0020] Each conveyor unit in the vision acquisition unit is independently equipped with a complete vision acquisition device. Each device includes a fixed upper industrial camera, a movable lower industrial camera, a dual-axis moving guide rail module, adjustable brightness station fill lights, and an image transmission gateway. The upper industrial camera is a 12-megapixel global shutter industrial camera with a 16mm lens focal length, fixedly mounted on an 800mm high bracket directly above the conveyor unit. The upper industrial camera continuously acquires a complete image of the top surface of the circuit board, covering all areas including components, solder joints, traces, and silkscreen printing. The lower industrial camera is an 8-megapixel global shutter camera, mounted below the conveyor unit's belt or roller. The bottom of the lower industrial camera is equipped with a dual-axis moving guide rail module with X-axis and Y-axis adjustable travel. The guide rail's effective travel range is 0 to 500mm on the X-axis and 0 to 400mm on the Y-axis. The guide rail is driven by a small stepper motor, and the business platform automatically sends adjustment coordinates based on the current work order's printed circuit board dimensions. The camera automatically moves to the center of the bottom surface of the printed circuit board, ensuring that the bottom image completely covers the entire area of ​​the circuit board. Each vision unit is equipped with two sets of LED ring lights, one set installed around the lens of the upper industrial camera and the other set installed at the front of the lower industrial camera. The brightness of the lights can be steplessly adjusted from 0% to 100%, and the brightness parameters are automatically matched for dark printed circuit boards and highly reflective pads. The upper and lower industrial cameras are triggered to shoot synchronously, and the shooting time difference is controlled within one millisecond. The two images generate a set of image data packets, which embed the current printed circuit board traceability code, workstation number, shooting timestamp, and camera coordinate parameters, and are transmitted to the business platform in real time. The image transmission gateway adopts a gigabit industrial switch, and the image transmission rate of a single workstation can reach 50 megabytes per second. For irregularly shaped circuit boards and ultra-thin flexible circuit boards, the guide rail module is equipped with a fine adjustment knob, and the operator can manually compensate for the automatic adjustment error.

[0021] The process parameter acquisition unit is configured differently for each workstation, and collects corresponding process data for different processes. All acquisition terminals are connected to the production equipment at each workstation via RS485 bus to achieve automatic reading of equipment parameters. For surface mount technology, the chip placement workstation collects the chip mounter nozzle pressure, placement coordinate offset, component feeding speed, and placement temperature. The reflow soldering workstation collects the real-time temperature of the eight temperature zones in the furnace, the conveyor belt speed, and the nitrogen concentration in the furnace. The through-hole and wave soldering workstations collect the solder pot temperature, wave height, flux spraying flow rate, and soldering travel speed. The lead cutting workstation collects the shearing stroke and cutting pressure. The functional testing workstation collects the output voltage, loop current, continuity resistance, and insulation resistance. The coding station collects laser power, coding speed, and traceability code printing clarity values. Each parameter acquisition terminal has a built-in local storage chip that caches 24 hours of parameter data when the device is disconnected and automatically synchronizes it to the business platform after communication is restored. All collected parameters are bound to the traceability code of the printed circuit board that the current station passes through, and dozens of sets of process parameters for the complete production process of a single printed circuit board are collected and stored in the cloud traceability archive. When the process parameters exceed the standard range of the process standard document, the process parameter acquisition unit actively sends an early warning signal to the business platform. The business platform records the abnormal time, abnormal value, and corresponding printed circuit board traceability code and simultaneously pushes the early warning prompt to the station operation display screen.

[0022] The real-time error prevention and reporting unit is equipped with an independent edge computing processor and performs image recognition and judgment locally. The unit runs a multi-level weighted judgment algorithm with three layers of judgment logic. The weight thresholds for each layer can be customized for different product types through the business platform. The first layer is pixel matching weight judgment. The system pre-stores standard complete images of various printed circuit boards and compares the real-time images acquired by the visual acquisition unit with the standard images pixel by pixel, calculating the percentage of pixel differences. If the percentage of differences exceeds the first threshold, it is determined that there are defects such as missing components, large areas of exposed copper, or missing silkscreen printing. The second layer is component position offset weight judgment. The algorithm identifies the outlines of resistors, capacitors, chips, and connectors on the printed circuit board and calculates the offset distance between the component center coordinates and the standard coordinates. If the offset distance exceeds the second threshold, it is determined that there is misplacement or component misalignment. The third layer is solder joint grayscale feature weighting. The algorithm extracts the grayscale histogram of the solder joint area and compares it with the grayscale range of the standard solder joint. If the grayscale deviation exceeds the third threshold, it is judged as a defect of cold solder joint, insufficient solder, solder bridging, or cold solder joint. If any of the three layers of judgment triggers the threshold, the unit will immediately output a triple error signal. The first layer is the activation of the red sound and light alarm at the workstation and a high-decibel buzzer to remind the on-site operator. The second layer is the lifting of the diversion baffle of the conveyor unit and the abnormal printed circuit board is introduced into the manual review channel. The third layer is the push of the abnormal work order to the display screen of the manual review unit. The work order includes images of the top and bottom sides of the printed circuit board, the excess weight data, and the corresponding workstation process parameters. The unit has a built-in false detection filtering mechanism and sets a tolerance range for minor pixel differences that do not affect the product function, such as minor scratches on the printed circuit board and minor color differences in materials. All error records, trigger times, and judgment weight values ​​are permanently bound to the traceability code of the printed circuit board and stored in the cloud.

[0023] The manual inspection unit is equipped with a 27-inch 4K high-definition display screen, a traceability code scanner, a mouse operating terminal, and a defect classification input keyboard. After the abnormal printed circuit board is conveyed to the inspection station via a diversion baffle, the operator uses the scanner to scan the traceability code on the surface of the printed circuit board. The display screen automatically retrieves the top and bottom double-sided images of the printed circuit board throughout the entire process, real-time error-proofing unit judgment data, and corresponding station process parameters. The operator reviews and judges the defects in three categories. The first category is genuine defects, and the operator clicks the input keyboard according to the defect type and distinguishes between five categories: placement misalignment, missing parts, poor soldering, surface scratches, and electrical defects. The system automatically marks the defect type and binds it to the traceability code. The source code is pushed to the corresponding defective product partition channel. The second type is visual misdetection, where the operator selects the misdetection option and fills in the reason for the misdetection. The misdetected printed circuit board is returned to the main line for normal circulation, and the system records the misdetected sample and automatically updates the local judgment weight threshold. The third type is boundary ambiguity defects, where the operator submits an expert review work order. The work order is pushed to the back-end process engineer terminal, and the engineer remotely retrieves the complete image and process parameters to give a judgment conclusion. The conclusion is sent back to the review station and the diversion operation is performed. All manual review operations, operator employee numbers, review time and judgment results are all stored in the printed circuit board traceability file.

[0024] The defective product zoning unit is equipped with four independent diversion channels, corresponding to the channels for appearance defects, component misalignment, soldering defects, and electrical malfunctions, respectively. Each channel ends with an independent material storage box, and an electronic display on the outside of the box continuously counts the number of defective products in the current channel and updates the data to the business platform report. Printed circuit boards that are manually confirmed as truly defective are automatically guided to their corresponding channels via a cylinder-driven diversion baffle. A secondary barcode scanner is installed at the channel entrance to reread the traceability code and confirm the defect type matches the channel. Each channel is assigned a corresponding process engineer's workstation. Appearance defects and component misalignment defects are disassembled, analyzed, and the defects caused by materials, equipment, and operator error are recorded by industrial engineering engineers. Because welding defects and electrical malfunctions are addressed by on-site application engineers who conduct in-depth circuit and solder joint inspections to pinpoint welding temperature, equipment parameters, and component quality issues. After completing the defect analysis, the engineer enters a complete defect analysis report into the system, which includes magnified images of the defects, cause determination, and rectification suggestions. The report is archived and linked to the printed circuit board traceability code. Defective products that can be reworked generate a dedicated rework order, which carries the original traceability code. The reworked and repaired printed circuit boards re-enter the production line and undergo a complete inspection process from the corresponding process. Images and parameters collected throughout the rework process are overlaid onto the original traceability file. Printed circuit boards that cannot be reworked and scrapped are marked as scrapped by the system, and the scrapping time, scrapping quantity, and defect cause are recorded.

[0025] The functional testing unit is located after the full appearance inspection process and performs electrical performance testing on semi-finished products without appearance defects. The unit integrates multiple DC regulated power supplies, a four-wire resistance tester, a continuity and insulation tester, and a data acquisition terminal, supporting simultaneous testing of eight printed circuit boards (PCBs). Before testing begins, the PCB traceability code is scanned, and the business platform automatically retrieves the standard electrical parameter thresholds for that PCB model. The standard voltage range is 0 to 48 volts, the loop current is 0 to 10 amperes, the continuity resistance is less than or equal to one ohm, and the insulation impedance is greater than or equal to one hundred megohms. The testing equipment automatically connects to the PCB test contacts and sequentially completes four tests: open-circuit voltage, on-load current, loop continuity, and insulation withstand voltage. The test involves the data acquisition terminal collecting various electrical values ​​in real time and simultaneously uploading them to the business platform to bind the traceability code. If all test data meet the standards, the function is deemed qualified and the printed circuit board is transferred to the comprehensive final inspection station. If any electrical indicator exceeds the standard range, the function is deemed faulty and automatically diverted to the electrical fault handling channel. The unit is equipped with a testing device with display function. The testing device is equipped with an automated servo lifting module that automatically raises and lowers the test probes and is compatible with printed circuit boards of different thicknesses. The front touch screen displays the real-time voltage and current curves of a single printed circuit board, and values ​​exceeding the standard are highlighted in red. All original data of functional tests, test duration, and probe contact status are permanently stored in the cloud traceability archive.

[0026] The cloud-based traceability repository adopts a distributed cloud server architecture with a storage capacity of 100TB, supporting five years of data storage for the entire printed circuit board manufacturing process. Data is encrypted using the AES256 encryption algorithm. The repository archives data in a five-layer hierarchical structure, with each layer independently indexed. The first layer is the work order basic information layer, storing work order number, bill of materials, production date, production line, and basic operator information. The second layer is the process image layer, storing high-definition images of all workstations (top and bottom, double-sided), shooting time, and camera coordinates. The third layer is the process parameter layer, storing operating parameter curves such as temperature, speed, and pressure for each workstation's equipment. The fourth layer is the defective product handling layer, storing error records, manual review and judgment, defect analysis reports, and rework / scrapping records. The fifth layer is the finished product testing layer, storing functional test electrical... The data, final inspection records, and outgoing sampling inspection records are all supported by a built-in multi-dimensional search engine in the business platform, which supports four traceability query methods. The first method is a precise traceability code query, where the unique 24-digit traceability code of the printed circuit board is entered, and all complete data from coding to shipment of the board is retrieved with one click, and a visual production timeline is automatically generated. The second method is a batch query of work orders, where the work order number is entered, and a defect statistical report of the entire batch of printed circuit boards is exported. The third method is a defect type filtering query, where defect types such as cold solder joints or mounting misalignment are selected, and the number of defects and corresponding workstations within a specified time period are counted. The fourth method is a time period interval query, where the start and end dates are selected, and all production, defect, and test summary data within that period are exported. The images, parameter curves, and defect reports generated by the search support export and printing in portable document formats.

[0027] The system is equipped with seven invention toolings, all of which are integrated with various workstations on the production line. The workbench surface of the irregularly shaped component marking device has a 5mm thick, highly elastic, flexible rubber positioning block with a hardness of 40HA, which clamps and fixes the irregularly shaped printed circuit boards. The left side of the workbench has a sealed cover connected to a negative pressure fan, and the dust and ink debris generated during marking are collected by negative pressure adsorption. The bottom of the workbench has a horizontal lead screw adjustment mechanism with an adjustment stroke of 0 to 400 mm, adaptable to marking and positioning various irregularly shaped circuit boards. The circuit board repair workbench is equipped with a multi-directional elastic clamping mechanism, and the clamping arm has three sets of angle-adjustable rotating shafts, achieving rotation from 0 to 180 degrees. Anti-slip silicone pads are pasted on the inner side of the clamps, and the clamping force is adjusted by spring buffer. The circuit board welding fixture has symmetrically arranged spring-pressing components on both sides, and arrows indicate that the width of the pressure block can be changed according to the side length of the data interface. During welding, the pressure block continuously presses down on the circuit board's data interface, and the fixture side integrates an RS485 data interface to collect welding temperature data in real time and upload it synchronously to the business platform. The blower module housing for circuit board welding has a circular permanent magnet embedded inside. The bottom of the windproof cover has a magnetic metal plate for quick disassembly and replacement. The inside of the windproof cover features a uniform airflow guide grille that confines hot air to a localized area of ​​the solder joints. The top of the circuit board positioning and detection device has a transparent polycarbonate pressure plate with adjustable bolts at the four corners controlling the total pressure height from 0 to 80 mm. The bottom black positioning plate is equipped with an electric lifting rod with a lifting stroke of 0 to 50 mm. Elastic positioning pins are distributed on the bottom surface of the positioning plate, corresponding to the positioning holes on the printed circuit board. The lifting module of the circuit board detection device with display function is driven by a servo motor and ball screw, with an automated lifting height of 0 to 120 mm. A 10-inch industrial touch screen counts the number of inspected, qualified, and defective printed circuit boards on the production line in real time, and simultaneously displays real-time voltage and current detection values, with a pop-up highlighting any out-of-tolerance data. The industrial camera system features a synchronously mounted upper camera that captures images from the top surface, while a lower rail-mounted mobile industrial camera with a strip light adapts to and moves with printed circuit boards of different sizes. The upper and lower cameras simultaneously trigger shooting and capture complete images of both sides of the circuit board in one go.

[0028] Example 2: This embodiment uses an actual printed circuit board (PCB) mass production line and a PCB work order for a consumer electronics motherboard as an example. The work order number is HX-20260701-003, and the single work order produces 5,000 double-layer PCBs. The process engineer imports the bill of materials, engineering change documents, PCB standard process documents, and appearance standard drawings for this work order into the work order management module of the business platform in advance. The business platform automatically integrates all the data and generates a unique production execution standard for this work order, which is then distributed to the coding station and surface treatment station. The entire process of surface mount technology (SMT), soldering, and inspection is handled by a terminal. Operators enter the work order number at the coding station terminal and click the "Generate Traceability Code" button. The business platform's encryption algorithm generates 5,000 unique 24-digit traceability codes in batches. An example of the coding format is 07010032607120001-00001. The code segments mean: 0701 represents the production date July 1st, 003 represents the work order number, 260712 represents the production line date July 12th, 2026, and 0001 represents the production line number. The serial number of a single printed circuit board (PCB) is determined by the last five digits, from 00001 to 05000, and a check digit is used for scanning and error correction. Operators switch the printing format and select QR code printing mode in the coding unit. The laser coding machine is set to 16 kW and the coding speed is 2000 mm / min. The traceability code is printed on the blank silkscreen area at the edge of the PCB and does not occupy the component wiring area. After coding, the photoelectric sensor detects the PCB's arrival, and the segmented conveyor belt starts, sending the PCB into the first surface mount technology (SMT) process. The coding unit simultaneously uploads the current PCB traceability code, coding equipment parameters, and coding time to the business platform and establishes the first traceability data record for a single PCB. When coding irregularly shaped PCBs, the irregular part coding device is switched, and flexible rubber blocks fix the board while a negative pressure cover simultaneously absorbs laser coding dust. After all 5,000 PCBs have been coded, the business platform automatically counts the number of completed codings, generates a coding process report, and simultaneously stores it in the cloud work order basic information layer.

[0029] The printed circuit board (PCB) flows into the segmented conveyor unit corresponding to the surface mount technology (SMT) placement station, and the photoelectric sensor at the inlet triggers a signal. The upper and lower cameras of the station's vision acquisition unit simultaneously start capturing images. The upper camera captures images of the SMT components on the top surface of the PCB. The lower moving guide automatically adjusts its X-axis coordinates to 160 mm and Y-axis coordinates to 90 mm based on the PCB's dimensions (320 mm x 180 mm). The supplementary lighting is adjusted to 60% brightness and captures images of the circuitry on the bottom surface of the PCB. These two images are packaged, embedded with the traceability code 00001, and uploaded to the business platform in real-time via a gigabit gateway. Simultaneously, the process parameter acquisition unit is activated, reading the surface mount technology (SMT) placement machine nozzle pressure as 0.3 MPa, placement coordinate offset as ±0.02 mm, feeding speed as 0.2 m / s, and oven temperature as 240 degrees Celsius. All parameter arrays are bound to the traceability code and stored in the business platform's local cache. After placement, the PCB flows to the reflow soldering station, and the corresponding conveyor unit triggers vision acquisition and parameter acquisition again, capturing eight segments. The temperature and nitrogen concentration in the temperature zone are 99.99% in real time, and the conveyor belt speed is 0.25 meters per second. Each subsequent process, including plug-in, wave soldering, lead cutting, initial appearance inspection, statistical process control, functional testing, final inspection, packaging, and outgoing sampling inspection, repeatedly executes the process of printed circuit board placement, sensor triggering, simultaneous image capture by upper and lower cameras, process parameter acquisition, and data binding and traceability code upload. The lower movable industrial camera has adaptive adjustment logic, and when the work order is switched to a small printed circuit board, the guide rail automatically retracts to the center coordinate and the brightness of the supplementary light is increased to 80%. When switching to a large-size industrial control printed circuit board, the guide rail extends to its maximum stroke. The timestamps of all workstation images are accurate to the millisecond level, and the upper and lower images correspond one-to-one. A single printed circuit board flows through twelve processes and generates twelve sets of double-sided image data packets and twelve sets of workstation process parameter data packets, all of which are collected into the dedicated traceability file of the printed circuit board. When the production line temporarily loses network access, the local industrial control computer caches all images and parameters, and the background automatically uploads them to the cloud in batches after the network is restored.

[0030] The real-time error prevention and reporting unit's edge processor receives the double-sided image of the printed circuit board uploaded from the current workstation in real time and retrieves the preset three-level weight thresholds for this work order. The standard thresholds for consumer electronics printed circuit boards are set to a pixel difference threshold of 3%, a component offset threshold of 0.15 mm, and a solder joint grayscale deviation threshold of 12%. The real-time error prevention and reporting unit simultaneously starts the three-layer judgment model calculation. The first layer, pixel matching judgment, compares the real-time image with the standard printed circuit board image of the work order pixel by pixel and calculates the percentage of difference pixels. If a component is missing or there is a large area of ​​exposed copper and the difference percentage exceeds 3%, an error is immediately triggered. The second layer, component offset judgment, extracts the coordinates of resistors, chips, and connectors through artificial intelligence contour recognition and compares them with the standard coordinate offset distance. If the offset exceeds 0.15 mm, it is judged as misplacement. The third layer, solder joint grayscale judgment, segments the grayscale histogram of the solder joint area. The standard solder joint grayscale range is 120 to 200. A value below 120 indicates a cold solder joint, and a value above 200 indicates solder bridging. The average grayscale value of the solder joints of the printed circuit board No. 1248 in this work order was 98 when it was inspected at the wave soldering station. Furthermore, if the third-level threshold is exceeded, the unit immediately initiates a triple error reporting system and triggers a continuous red audible and visual alarm at the workstation. Simultaneously, the conveyor unit's diversion cylinder raises the diversion baffle and guides the printed circuit board into the manual inspection channel. An abnormal work order is simultaneously sent to the manual inspection unit, including double-sided images of the printed circuit board, solder joint grayscale values, wave soldering oven temperature of 268 degrees Celsius, and traceability code HX-20260701-003-01248. Since the standard temperature range is 250 to 260 degrees Celsius and this parameter exceeds the limit, the main conveyor line is suspended. From delivery until the operator completes manual review and resets the alarm signal, the system automatically records the error trigger time, the excess weight, and the corresponding workstation equipment parameters, and binds the traceability code to the defective product processing layer database. For images with pixel differences within 2% caused by dust and slight color differences, the system judges them as normal interference and does not trigger an error. For industrial precision printed circuit board work orders, the business platform automatically switches to strict thresholds, sets the pixel difference threshold to 1%, the component offset threshold to 0.08 mm, and the solder joint grayscale deviation threshold to 5%.

[0031] The abnormal printed circuit board (PCB) traceability code 01248 is transported to the manual review unit via a diversion baffle. The operator scans the QR code on the PCB surface with a handheld barcode scanner. The 4K display automatically loads images of both sides of the PCB throughout the entire process, excessive temperature parameters at the wave soldering station, and real-time error-proofing unit's three-level weighting judgment data. The operator zooms in on the solder joint area image and observes insufficient solder and voids, confirming that these are genuine soldering defects. The operator clicks the soldering defect button on the defect classification keyboard. The system automatically records the operator's employee number, review time, and judgment conclusion, and simultaneously pushes the PCB to the dedicated soldering defect diversion channel. The system also simultaneously enters the preliminary judgment of the defect cause. If the wave soldering oven temperature is deemed to be too high and the flux application is insufficient, and a false inspection work order is subsequently encountered due to slight color difference in the image, the operator will check the visual false inspection and fill in the reason for the false inspection as the background color deviation of the printed circuit board material. The falsely inspected printed circuit board will be diverted back to the main line for normal flow, and the background will automatically fine-tune the grayscale tolerance threshold of the solder joint at that station. If a defect with blurred boundaries is encountered, the operator will click on expert review, and the work order will be pushed to the process engineer's terminal in real time. The engineer will remotely retrieve the complete image and the entire process parameters and give a judgment within five minutes. The judgment will be sent back to the review station to perform the corresponding diversion operation. All manual review operations are fully traceable, and the review record is permanently bound to the printed circuit board traceability code.

[0032] A PCB with poor soldering (traceability code 01248) entered the soldering defect diversion channel, and the display screen at the end of the channel continuously updated the number of soldering defects to twelve. The on-site application engineer retrieved the defective PCB and conducted in-depth disassembly and analysis. The engineer retrieved the process parameters for the wave soldering station on the PCB: a solder pot temperature of 268 degrees Celsius and a soldering travel speed of 0.3 meters per second. Comparing these parameters with the process standard document (solder pot temperature 250-260 degrees Celsius), the engineer confirmed that a malfunction in the solder pot heating module's temperature control caused excessive temperature and reduced solder flow, resulting in a cold solder joint defect. The engineer entered a complete defect analysis report into the system, uploaded magnified images of the solder joints, and recommended repairing the solder pot temperature control module and recording the solder pot temperature every two hours. The PCB could not be reworked and was marked as scrapped by the system, which recorded the scrapping time, defect type, and analysis report, and stored it in the cloud. In the good product processing layer, three printed circuit boards (PCBs) with misaligned mounting were found on the same production line and automatically entered into the component misalignment channel. The industrial engineering engineer analyzed the issue and determined that wear on the pick-and-place machine nozzle caused the positioning misalignment, generating a nozzle replacement rectification work order. The three PCBs were recalibrated and reworked, with the rework work order carrying the original traceability code. After rework, the PCBs were fully inspected again from the surface mount technology (SMT) process, and the images and equipment parameters collected during rework were superimposed onto the original traceability file, forming a complete traceability chain of original defects, rework repair, and secondary inspection. At the end of each day's production line, the business platform automatically summarizes the defect data from the four diversion channels and generates a defect percentage report. This work order had a total of 76 defective PCBs out of 5,000, including 42 with soldering defects, 21 with mounting misalignment, 9 with cosmetic scratches, and 4 with electrical defects. The report was simultaneously pushed to the process department to optimize equipment parameters and workflow.

[0033] Printed circuit boards (PCBs) that pass initial visual inspection and statistical process control are transferred to the functional testing unit. Operators scan the PCB traceability code, and the business platform automatically retrieves the electrical standard thresholds for consumer electronics motherboards. The standard values ​​are: no-load voltage of 5 volts, on-load current of 0.8 amps, on-resistance less than or equal to 0.5 ohms, and insulation resistance greater than or equal to 200 megohms. Test probes automatically press down to connect to the PCB test contacts and sequentially complete the four electrical indicator tests. The data acquisition terminal transmits the values ​​to the business platform in real time and binds them to the traceability code. This work order is No. 3692. During the printed circuit board (PCB) inspection, the open-circuit voltage was only 3.2 volts, which was below the standard threshold. The system determined that the electrical function was faulty and automatically diverted the power to the faulty electrical channel. The on-site application engineer disassembled and inspected the PCB, found that the power line was broken, and recorded the cause of the defect in the traceability file. The remaining 4,920 PCBs with all electrical indicators met the standards were marked as qualified by the system and transferred to the comprehensive final inspection station. The display screen of the testing device with display function showed the real-time voltage and current curves of the current PCB, and the voltage value was highlighted in red to indicate the faulty board. All original electrical test data were automatically archived in the cloud finished product test layer.

[0034] Printed circuit boards that pass functional testing enter the comprehensive final inspection station, where manual inspections are conducted on appearance, components, and silkscreen printing. After the final inspection confirms no defects, they undergo anti-static vacuum packaging. Once packaged, they are transferred to the outgoing sampling inspection station. Quality inspectors randomly select 3% of the printed circuit boards according to industry standards, scan the codes to retrieve traceability files, and verify the entire process record. If the sampling inspection is successful, the entire batch is transferred to the pre-shipment warehouse. If defects are found during sampling, the entire batch of printed circuit boards is suspended from shipment, and the corresponding printed circuit board's entire process data is retrieved using the traceability code to pinpoint the batch defect process. The work order contains 5,000 printed circuit boards. After the production, testing, packaging, and outgoing sampling inspection of the printed circuit board are completed, the business platform initiates the cloud archiving process and stores the basic information of all printed circuit board work orders, double-sided images of twelve processes, complete set of process parameters, seventy-six defective product review and analysis reports, functional test electrical data, final inspection records, and outgoing sampling inspection records of the work order in a unified layer into the cloud traceability repository. Each printed circuit board generates an independent and complete traceability file, which is permanently stored for a period of five years. The local industrial control computer cache data is synchronously cleared and the local storage space is released, and the cloud data is encrypted and backed up twice.

[0035] Three months after the shipment of this batch of printed circuit boards, the end customer reported a fault where one of the motherboards had no voltage upon power-on. The customer provided the traceability code HX-20260701-003-03692 on the printed circuit board surface. The factory quality engineer logged into the traceability search interface of the business platform and entered the 24-digit traceability code to initiate a precise search. The system instantly retrieved the complete traceability file for the printed circuit board and automatically generated a visual production timeline. The timeline was displayed in the order of coding, surface mount, reflow soldering, through-hole assembly, wave soldering, functional testing, and final inspection, showing all double-sided images, process parameters for each station, records of the original 3.2-volt voltage exceeding the standard during functional testing, electrical fault handling reports, and information from the on-site application engineer. The system includes circuit breakage analysis records. Engineers can zoom in online on solder joints and circuit images at each process stage and export wave soldering temperature curves and functional test electrical data in portable document format. A complete traceability report can be printed with one click and submitted to the customer. Through the traceability archive, it is confirmed that the printed circuit board's factory functional test detected a voltage fault and that it was a missed inspection. Quality engineers can quickly locate human error at the final inspection station and simultaneously optimize the final inspection review process, adding a second verification step for functional test data. If the customer reports similar defects in batches, the system can input the work order number to retrieve and export a batch defect statistical report for 5,000 printed circuit boards. Quality engineers can then locate high-frequency defect stations and equipment and carry out systematic process rectification.

[0036] Example 3: This embodiment quantitatively verifies the technical advantages of the circuit board inspection system and control method based on traceability codes of the present invention compared with traditional printed circuit board production inspection solutions through a mass production comparative test of printed circuit boards of the same specifications. The test conditions are 5,000 printed circuit boards of the same model of consumer electronics motherboards and a production line running time of eight hours. The comparison indicators include five dimensions: traceability coverage integrity, defect detection rate, manual review workload, after-sales defect traceability time, and closed-loop management rate of defective products. In the comparison of traceability coverage integrity, the traditional hole array code traceability can only cover the printed circuit board printing process, and the raw materials, surface mount technology, welding and functional test data are not bound, and the traceability coverage rate is only 32%. The system of the present invention marks the printed circuit board with codes. This invention integrates images, process parameters, and inspection records for all twelve processes, ensuring complete archiving of data for each printed circuit board throughout the entire process. It also achieves 100% traceability coverage, allowing for the tracking of raw material suppliers, the operational status of each processing device, operators, and every inspection and judgment record. In comparison, traditional single-camera top-view visual inspection had a 18.7% missed detection rate for bottom solder joint defects and an overall defect detection rate of 79.2%. This invention, however, uses dual cameras simultaneously to acquire double-sided images and employs a three-level weighted hierarchical judgment system, achieving zero missed detections for bottom defects and an overall defect detection rate of 98.6%, while reducing the batch defect rate by over 90%. Furthermore, in comparison, traditional methods lack hierarchical weighted filtering and are less efficient in terms of manual review workload. Color difference and dust all triggered errors, and the manual review of work orders on the eight-hour production line resulted in 1,260 work orders. This invention sets a tolerance threshold to filter out invalid interference and only pushes genuine defects for review, with 172 work orders reviewed concurrently. Manual workload is reduced by 86.3%, and labor costs at workstations are saved. In comparison of after-sales defect tracing time, traditional methods lack unified digital traceability archives, and after-sales fault location requires reviewing paper work orders and equipment ledgers. The average tracing time for a single printed circuit board is 4.2 hours. This invention allows one-click retrieval of full-process data by inputting the traceability code and automatically generates a visual timeline. The complete traceability report is generated in less than ten seconds, improving tracing efficiency by 1,500 times. In comparison of the closed-loop control rate of defective products, traditional production lines have no binding mechanism for defective product diversion. The existing printed circuit board (PCB) production system suffers from issues such as incomplete records, missing scrap and rework ledgers, and a 61.5% closed-loop management rate for defective products. This invention, however, addresses these problems by binding a traceability code to each defective product throughout the entire process—from error reporting, review, triage, analysis, and rework / scrapping—achieving a 100% closed-loop management rate with no defective product loss and no missing defect records. Comprehensive experimental data demonstrates that this invention, through six core technical solutions—segmented conveyor unit full-area scanning, simultaneous visual acquisition by dual cameras, multi-level weighted real-time error prevention, standardized manual review, zoned closed-loop processing of defective products, and cloud-based layered traceability storage—completely solves the technical deficiencies of existing PCB production systems, including traceability gaps, missed bottom defects, excessive false detections, inability to quickly locate defects, chaotic defective product management, and low after-sales traceability efficiency.Simultaneously, seven specialized invention tooling designs optimize the positioning stability of marking, welding, repair, and testing stations, significantly reducing defects caused by PCB deformation and positioning misalignment during processing. This enables digital and visual management of the entire PCB lifecycle, allowing for reversible and precise quality control, and is applicable to mass production lines for various PCBs in consumer electronics, industrial control, automotive electronics, and aerospace industries.

[0037] In the description of this invention, unless otherwise explicitly defined, terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0038] This specific embodiment is merely an explanation of the present invention and is not intended to limit the invention. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they are within the scope of the claims of the present invention.

Claims

1. A circuit board inspection system based on traceability codes, characterized in that: The system includes a business middle platform, a coding unit, a segmented conveyor unit, a multi-station visual acquisition unit, a process parameter acquisition unit, a real-time error prevention and reporting unit, a manual review unit, a defective product zoning processing unit, a functional testing unit, and a cloud-based traceability repository. The segmented conveyor unit consists of multiple independently controllable belt conveyors or roller conveyors connected in series. Single conveyor units are arranged at the junctions of adjacent processes. Each conveyor unit is independently equipped with the visual acquisition unit. The business middle platform is connected to the coding unit, segmented conveyor unit, visual acquisition unit, process parameter acquisition unit, and real-time error prevention and reporting unit. The unit, manual review unit, functional testing unit, and cloud traceability storage unit are electrically connected. The coding unit receives production work orders from the business platform, generates unique traceability codes in batches, and prints them onto the surface of the circuit board. The traceability codes are barcodes / QR codes. The vision acquisition unit includes an upper industrial camera, a lower mobile industrial camera, a guide rail adjustment module, and a station supplementary light. The upper industrial camera is fixed to capture images of the top surface of the circuit board, while the lower industrial camera, equipped with a mobile guide rail, achieves adaptive adjustment of its horizontal and vertical positions and, in conjunction with the supplementary light, synchronously captures images of the bottom surface of the circuit board. The process parameters... The data acquisition unit is independently configured for each workstation and is used to collect the operating parameters and processing parameters of the equipment in each process, and is bound to the corresponding circuit board traceability code. The real-time error prevention and reporting unit has a built-in image weight judgment algorithm and compares the circuit board image uploaded by the visual acquisition unit with the standard sample in real time. If the judgment threshold is exceeded, an error signal is output. The manual review unit receives the abnormal work orders pushed by the real-time error prevention and reporting unit, completes the manual review of visual mis-inspection, and synchronously sends the review results back to the business platform. The defective product zoning processing unit sets up independent channels for appearance defects, misplacement, welding defects, and electrical function defects. The system includes a functional testing unit that collects electrical data on circuit board voltage, current, and conductivity, and uploads this data to a cloud-based traceability repository linked to a traceability code. The cloud-based traceability repository establishes independent data archives for each circuit board traceability code and archives images of the entire process, equipment parameters, testing records, review records, and functional test data. It also supports full-process reverse traceability queries. The system further includes a circuit board repair workbench equipped with a multi-directional elastic clamping mechanism. This mechanism is adaptable to clamping and fixing multi-layer circuit boards and circuit boards with irregular edges. The clamping mechanism includes anti-slip silicone pads and an angle adjustment shaft.

2. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The coding unit includes a coding workbench for irregularly shaped parts. Flexible rubber positioning blocks are attached to the surface of the workbench. A dust removal cover is set on the left side of the workbench. The dust removal cover is connected to a negative pressure dust extraction fan and is used to absorb coding dust. A horizontal adjustment screw is set at the bottom of the workbench to adapt to the positioning of irregularly shaped circuit boards of different sizes.

3. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The system also includes a circuit board soldering fixture. The fixture is equipped with a liftable data interface pressing component. The pressing component includes two sets of symmetrical pressing blocks. The pressing blocks use a spring buffer structure to press down and fix the circuit board data interface and prevent the interface from deforming and warping under the high temperature of soldering. The side of the fixture integrates a data transmission interface and collects soldering temperature data in real time and synchronizes it to the business middle platform.

4. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The system also includes a circuit board soldering blower module. The blower module is equipped with a magnetic windproof cover. The module housing has a built-in permanent magnet. The windproof cover can be quickly installed and removed by magnetic attraction. The windproof cover has a guide grid inside to constrain the hot air to act in the direction of the solder joint area to avoid the surrounding components from being overheated and having poor soldering.

5. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The system also includes a circuit board positioning and detection device. The top of the detection device is equipped with a transparent pressing plate. The total pressing stroke of the pressing plate is controlled by adjusting bolts. The bottom of the device is equipped with a liftable black positioning plate. The lower end of the positioning plate is distributed with positioning pins and is adapted to support the positioning of circuit boards of different thicknesses.

6. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The system is also equipped with a circuit board inspection device with a display function. The inspection device integrates an automated lifting module, which is driven by a servo motor. The front of the device is equipped with a touch screen display, which counts the number of circuit board inspections in real time and visually displays the voltage and current test values ​​of each circuit board. Data exceeding the standard is highlighted in a pop-up window.

7. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The cloud-based traceability repository stores data in layers, including a work order basic information layer, a process image layer, a process parameter layer, a defective product handling layer, and a finished product testing layer. The business platform is equipped with a search engine and supports multi-dimensional retrieval of full-process data by traceability code, work order number, defect type, and production time period.

8. The circuit board inspection system based on traceability codes according to claim 1, characterized in that: The real-time error prevention and reporting unit has a built-in multi-level weight judgment model. The first level is the standard image pixel matching weight, the second level is the component position offset weight, and the third level is the solder joint grayscale feature weight. If any of the three weights exceeds the preset threshold, an error signal will be triggered.

9. A detection method for a circuit board detection system based on traceability codes according to any one of claims 1-8, characterized in that... Includes the following steps: S1. Batch generation of work order traceability codes and circuit board coding: The business platform reads the standardized production work order and generates an independent traceability code for each circuit board based on the unique work order number and sends it to the coding unit. The coding machine prints the traceability code on the blank area of ​​the circuit board. After coding is completed, the segmented conveying unit sends the circuit board into the first production process. S2. Full-process images and process parameters are collected synchronously. When the circuit board flows through each segmented conveying unit, the vision acquisition unit simultaneously collects high-definition images of the top and bottom surfaces of the circuit board. The process parameter acquisition unit collects the current station equipment speed, temperature, pressure, and welding time parameters. All data are bound to the circuit board traceability code and uploaded to the business platform. S3. Real-time visual weight judgment and error prevention warning: The real-time error prevention and error reporting unit retrieves the circuit board image bound with the traceability code and calls the multi-level weight judgment model to compare with the standard sample. If it is determined that there is a placement offset, missing parts or solder joint defects, it will immediately output an audio-visual error and the conveying unit will stop conveying. S4. Abnormal work orders are manually reviewed and classified. The manual review unit receives the error work orders. Operators retrieve images and process parameters of the entire process to manually check and distinguish between real defects and visual mis-detections. Mis-detected products are marked and returned to the production line, while real defective products are pushed to the defective product zoning processing unit. S5. Defective products are sorted and handled in separate channels. Four types of defective products, namely appearance defects, component misalignment, welding defects and electrical faults, enter independent sorting channels. Process engineers and field application engineers respectively conduct defect analysis and record the defect causes, bind traceability codes and store them in the cloud. S6. Electrical function testing of semi-finished products: Circuit boards that pass the appearance re-inspection flow into the functional testing unit and test various electrical indicators of the circuit boards. Voltage, current and continuity data are bound to the traceability code. Defective products are returned to the repair station and re-tested after repair. S7. Final inspection and traceability archive of finished products: Circuit boards that pass the final inspection are packaged with anti-static material and sampled for shipment. All process images, equipment parameters, test records and defect handling records are archived in the cloud traceability repository and a complete traceability archive is formed for each circuit board. S8. Reverse traceability query: When a finished product has after-sales quality problems, enter the circuit board traceability code to retrieve the entire process of production, testing and defect handling data, and locate the defect-causing process, equipment and operator.

10. The detection method of the circuit board detection system based on traceability code according to claim 9, characterized in that: In step S2, the lower movable industrial camera automatically adjusts the guide rail position according to the length and width of the circuit board. The supplementary light adjusts its brightness synchronously with the camera. The timestamps of the bottom image and the top image are completely synchronized, and the upper and lower images are bound to the same traceability code. In step S3, the multi-level weight judgment model presets adjustable thresholds and sets three sets of differentiated weight threshold schemes for precision circuit boards, ordinary consumer electronics circuit boards, and industrial control circuit boards. The business platform automatically switches the judgment standard according to the product type of the work order. In step S5, after the defective product is disposed of, the system automatically generates a rework work order. The rework work order carries the traceability code of the original circuit board, and the rework process data is superimposed on the original traceability file to form a secondary complete traceability record for defect repair. In step S8, during reverse traceability, the system automatically generates a visualized full-process timeline and displays each step's image, equipment operating parameters, and detection judgment results in the order of the process. It also supports image magnification, export of process parameter curves, and one-click printing of defect reports.

Images