A printed circuit board aging test method based on intelligent sensing technology

Abstract

The present application relates to printed circuit board detection technical field, specifically to a kind of printed circuit board aging test method based on intelligent sensing technology, comprising: installing miniature temperature and humidity sensor and impedance measurement probe at printed circuit board key test point, applying preset electrical stress and thermal stress load makes circuit board enter simulation working state, the temperature distribution of synchronous acquisition board surface, environmental humidity distribution and the continuous change data of loop impedance between key test points, the multi-physical field synchronous monitoring data set is generated by aligning with the fusion of multiple data according to uniform timestamp, the abnormal hot zone position in aging process and impedance mutation time are identified by data set, and the preliminary potential failure area of circuit board is positioned.The method can be complete associated with temperature and humidity and impedance change characteristics in the aging process of circuit board by multi-physical field data synchronous acquisition and timestamp fusion alignment, and accurately capture aging abnormal state.

Description

Technical Field

[0001] This invention relates to the field of printed circuit board testing technology, and in particular to an aging test method for printed circuit boards based on intelligent sensing technology. Background Technology

[0002] Conventional printed circuit board aging tests typically involve placing a single type of monitoring element at the test point to independently collect circuit board temperature and humidity parameters or circuit impedance parameters. After applying electrical and thermal stress loads to the circuit board, discrete data on temperature and humidity or impedance change data are obtained step by step. Some test schemes rely solely on impedance probes to collect circuit electrical parameters or solely on temperature and humidity sensors to monitor environmental and board surface parameters. The entire stress loading process uses only a single physical quantity data as the core of monitoring, and the collection of different types of parameters is carried out independently.

[0003] Conventional testing methods make it impossible to unify the timing of temperature and humidity data acquisition with impedance data. The two types of data are disconnected in the time dimension. The linkage changes of temperature, humidity and impedance during the aging process of circuit boards cannot be fully recorded. A single physical quantity data cannot fully reflect the aging state of the circuit board. Abnormal hot areas and impedance changes that occur during the aging process cannot be correlated. When relying on a single data to determine the potential failure area of ​​the circuit board, the judgment result is easily affected by the single nature of the monitoring data and cannot match the actual aging state of the circuit board.

[0004] Temperature and humidity sensors and impedance measurement elements need to be deployed simultaneously at key test points to collect corresponding physical quantity data. Multiple types of monitoring data are integrated according to a unified timestamp. Based on the integrated multi-dimensional data, abnormal aging hot areas and impedance change moments can be identified, thereby locating potential failure areas of the circuit board. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an aging test method for printed circuit boards based on intelligent sensing technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a printed circuit board aging test method based on intelligent sensing technology, comprising: Select multiple key test points on the printed circuit board to be tested, and install a miniature temperature and humidity sensor and an impedance measurement probe at each key test point. A pre-set electrical stress load and thermal stress load are applied to the printed circuit board to bring it into a simulated working state; The miniature temperature and humidity sensor continuously collects surface temperature distribution data and ambient humidity distribution data of the printed circuit board under simulated working conditions. The impedance measurement probe is used to synchronously collect data on the continuous change of loop impedance between the key test points at fixed sampling intervals while stress load is applied. The collected surface temperature distribution data, ambient humidity distribution data and loop impedance continuous change data are fused and aligned according to a unified timestamp to generate a multi-physics field synchronous monitoring dataset. Based on the multi-physics synchronous monitoring dataset, the locations of abnormal hot spots and the moments of impedance abrupt changes in the printed circuit board during the aging process are identified. Based on the location of the abnormal hot zone and the time of impedance change, the initial potential failure area of ​​the printed circuit board is located.

[0007] As a further aspect of the present invention, the installation of miniature temperature and humidity sensors and impedance measurement probes at each key test point specifically includes: Analyze the circuit schematic and layout diagram of the printed circuit board to identify nodes with high current density, concentrated heat sources, or intersecting lines as the basis for selecting key test points. At each selected key test point, the sensing unit of the miniature temperature and humidity sensor was attached to the surface of the printed circuit board using a non-conductive adhesive. Next to the sensing unit of the miniature temperature and humidity sensor, the two contact points of the impedance measurement probe are soldered to the start and end pads of the electrical circuit where the key test point is located, respectively. Calibrate the temperature and humidity readings of each of the miniature temperature and humidity sensors to ensure that their measurement errors are within the allowable range; The zero-point drift of each impedance measurement probe is calibrated to ensure that the impedance measurement reading is close to the theoretical value under no-load conditions.

[0008] As a further aspect of the present invention, the application of a pre-set electrical stress load and a thermal stress load to bring it into a simulated working state specifically includes: The printed circuit board is connected to a programmable power supply and electronic load system, and the voltage amplitude and current waveform of the electrical stress load are set according to its rated operating parameters. Turn on the programmable power supply and electronic load system, and apply the set voltage amplitude and current waveform to the power supply input port of the printed circuit board; While applying the electrical stress load, the entire printed circuit board is placed inside the environmental chamber of a programmable temperature chamber. According to the preset aging test profile, the ambient temperature of the programmable temperature chamber is controlled to cycle according to the set heating, holding and cooling curves to form the thermal stress load; Record the time from the start of applying the electrical stress load and thermal stress load as the time reference origin for the entire aging test.

[0009] As a further aspect of the present invention, the surface temperature distribution data and ambient humidity distribution data of the printed circuit board under simulated working conditions are continuously collected, specifically including: After the time reference origin of the aging test, all miniature temperature and humidity sensors are activated to collect data at the set sampling frequency. Each of the miniature temperature and humidity sensors independently collects the temperature reading on the surface of the printed circuit board at its attachment location and the relative humidity reading of the surrounding microenvironment; The temperature readings from all the miniature temperature and humidity sensors are arranged according to the spatial layout of the sensors to form an instantaneous surface temperature distribution matrix at each sampling moment. The humidity readings from all the miniature temperature and humidity sensors are arranged in the same spatial layout to form an instantaneous environmental humidity distribution matrix at each sampling time. According to the sampling time sequence, the continuous instantaneous surface temperature distribution matrix and instantaneous ambient humidity distribution matrix are stored to form surface temperature distribution data and ambient humidity distribution data that change over time.

[0010] As a further aspect of the present invention, the step of synchronously acquiring continuous change data of loop impedance between the key test points at a fixed sampling interval specifically includes: After the time reference origin of the aging test, start the impedance analyzer corresponding to all impedance measurement probes and set a fixed sampling interval that matches the sampling period of the miniature temperature and humidity sensor. At each sampling interval, the impedance analyzer sends an excitation signal to the impedance measurement probe and measures the response signal of the electrical circuit where the key test point is located; Based on the excitation signal and response signal, the complex impedance value of the electrical circuit at the current moment is calculated, and its impedance amplitude is extracted as the main recording parameter. The impedance amplitude measured at each sampling time and each key test point loop is bound and stored with the loop identification information and sampling timestamp; According to the sampling time sequence, the impedance amplitude sequence of all key test point loops is organized to form a dataset of continuous loop impedance changes over time.

[0011] As a further aspect of the present invention, the step of fusing and aligning the collected surface temperature distribution data, ambient humidity distribution data, and continuous loop impedance change data according to a unified timestamp to generate a multi-physics field synchronous monitoring dataset specifically includes: For each temperature and humidity reading from the miniature temperature and humidity sensor, append a timestamp corresponding to the sampling time; For each impedance amplitude reading from the impedance analyzer, append a timestamp corresponding to the sampling time; Starting from the time reference origin of the aging test, a unified time axis is established, with the smallest scale of the time axis being the period of the highest sampling frequency among all sensors. The time-stamped surface temperature distribution data, ambient humidity distribution data, and continuous change data of loop impedance are respectively mapped onto the unified time axis; For each aligned time point on the time axis, the surface temperature distribution matrix, the ambient humidity distribution matrix, and the impedance amplitude of all loops at that time point are extracted and encapsulated into a data frame. The data frames of all time points are arranged in sequence to form a multi-physics synchronous monitoring dataset.

[0012] As a further aspect of the present invention, based on the multi-physics synchronous monitoring dataset, the location of abnormal hot spots and the moment of impedance change during the aging process of the printed circuit board are identified, specifically including: Extract a continuous surface temperature distribution matrix sequence from the multiphysics synchronous monitoring dataset; Calculate the average temperature of the entire printed circuit board at every moment, and use the average temperature as a reference to identify all sensor locations where the temperature exceeds the reference value by a preset threshold. During continuous sampling times, the sensor positions above the reference value are tracked, and areas where the temperature continuously exceeds the limit and are spatially continuous are marked as candidate abnormal hot areas; From the multi-physics synchronous monitoring dataset, extract the time sequence of impedance amplitude changes for each electrical circuit; A change point detection algorithm is applied to each impedance amplitude sequence to find the sampling time point when the impedance amplitude experiences a step increase or decrease. The sampling time point is the impedance change moment.

[0013] As a further aspect of the present invention, based on the location of the abnormal hot zone and the time of impedance abrupt change, the preliminary potential failure area of ​​the printed circuit board is located, specifically including: Obtain all labeled candidate anomalous hot zones, as well as all impedance abrupt change moments detected from all loops; On the time axis of the multiphysics synchronous monitoring dataset, the impedance abrupt change time is compared with the time range of the candidate abnormal hot zone. Candidate anomalous hot regions that coincide with or are adjacent to the impedance abrupt change time in time are selected as highly correlated anomalous hot regions. Obtain the electrical circuits that are spatially overlapping or adjacent to the anomalous thermal areas that are highly correlated with the thermal anomaly; these electrical circuits are those that are related to both thermal anomalies and impedance mutations. The physical area covered by the highly correlated abnormal hot zone, and the wiring area occupied by the related electrical circuits on the printed circuit board, are collectively defined as the initial potential failure area.

[0014] As a further aspect of the present invention, the step of applying a change point detection algorithm to each impedance amplitude sequence to find the sampling time point where the impedance amplitude experiences a step increase or decrease specifically includes: For an impedance magnitude sequence of an electrical circuit, a sliding window is used to calculate the local statistical characteristics of the sequence, including the local mean and local standard deviation. Compare the local statistical characteristics of adjacent sliding windows, calculate their difference, and when the difference exceeds the set sensitivity threshold, record the time at the boundary of adjacent sliding windows as a potential mutation point. Verify all recorded potential abrupt changes and check whether the impedance amplitude sequence before and after the potential abrupt change satisfies the pattern of a step change rather than a slow drift. Eliminate potential abrupt change points that do not conform to the step change pattern, retain the points that are confirmed to be step changes, and the corresponding sampling time is the impedance change time of the electrical circuit. Repeat the steps for the impedance magnitude sequence of all electrical circuits to obtain the set of impedance change moments for each circuit.

[0015] As a further aspect of the present invention, the physical area covered by the highly correlated abnormal hot zone, and the wiring area occupied by the related electrical circuit on the printed circuit board, are jointly defined as the preliminary potential failure area, specifically including: On the layout diagram of the printed circuit board, mark the sensor installation locations corresponding to all highly correlated abnormal hot areas, and estimate the physical range of their influence based on the thermal diffusion model, with each location as the center. By superimposing all the estimated affected physical areas, a set of failure areas caused by thermal anomalies is obtained; On the layout diagram of the printed circuit board, mark the wiring paths of all relevant electrical circuits, including wires, vias and pads. Extend all marked wiring paths to both sides by a certain width to form a set of failure areas related to the electrical circuit; The set of failure areas caused by thermal anomalies and the set of failure areas related to the electrical circuits are combined by taking the union operation. The physical area on the printed circuit board covered by the union is then defined as the preliminary potential failure area.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: Miniature temperature and humidity sensors and impedance measurement probes are installed at key test points on the printed circuit board to be tested. After applying a pre-set electrical stress load and thermal stress load to the circuit board, the continuous change data of the loop impedance between the key test points are collected synchronously at a fixed sampling interval. The surface temperature distribution data and ambient humidity distribution data of the circuit board are collected synchronously. The acquisition of temperature and humidity parameters and impedance parameters are carried out synchronously. The acquisition nodes of various physical quantity data are completely matched in the time dimension. The real-time changes of environmental parameters and electrical parameters during the aging process of the circuit board can be synchronously retained. There will be no time sequence misalignment of different types of monitoring data, and the linkage relationship between parameters can be fully presented.

[0017] The collected surface temperature distribution data, environmental humidity distribution data, and continuous change data of circuit impedance are fused and aligned according to a unified timestamp to form a multi-physics synchronous monitoring dataset. Based on this dataset, the location of abnormal hot areas and the time of impedance change during the aging process of printed circuit boards are identified. The multi-physics data correspond to each other, and the spatial location information of abnormal hot areas can be matched with the time node of impedance change. The preliminary potential failure area can be located based on multi-dimensional monitoring information. The fluctuation of a single physical quantity data will not directly interfere with the identification results of abnormal features. All kinds of abnormal states during the aging process of circuit boards can be fully captured. Attached Figure Description

[0018] Figure 1 This is a flowchart of a printed circuit board aging test method based on intelligent sensing technology according to the present invention. Figure 2 A flowchart for installing a miniature temperature and humidity sensor and an impedance measurement probe; Figure 3 A flowchart for continuously collecting surface temperature distribution data and ambient humidity distribution data; Figure 4 A bar chart comparing temperature, humidity, and impedance characteristics at different stages of PCB aging test; Figure 5 This is a time-series graph showing the changes in surface temperature and ambient humidity throughout the entire PCB aging test. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0020] See Figure 1Multiple key test points are selected on the printed circuit board (PCB) to be tested, and a miniature temperature and humidity sensor and an impedance measurement probe are installed at each key test point. Pre-set electrical and thermal stress loads are applied to the PCB to simulate its operating state. During the stress load application, the miniature temperature and humidity sensors continuously collect surface temperature distribution data and ambient humidity distribution data of the PCB under simulated operating conditions. The impedance measurement probes collect continuous loop impedance change data between the key test points at fixed sampling intervals. The collected surface temperature distribution data, ambient humidity distribution data, and continuous loop impedance change data are fused and aligned according to a unified timestamp to generate a multi-physics synchronous monitoring dataset. Based on this dataset, the locations of abnormal hot zones and impedance abrupt changes during the aging process of the PCB are identified. Based on the identified abnormal hot zone locations and impedance abrupt change times, the initial potential failure areas of the PCB are located.

[0021] In one embodiment of the present invention, see [reference] Figure 2 Using a printed circuit board (PCB) for power management as an example scenario, the PCB schematic shows multiple MOSFETs and energy storage inductors in the main power conversion area, while the layout diagram further reveals the width and routing of the copper traces powering these components. The criteria for determining critical test points include high-current-density areas identified in the schematic, such as the positive and negative pads of the input filter capacitors and the pins of the power inductors in the main power loop; areas of concentrated heat sources, such as directly below the heat sink pads of the MOSFET devices; and nodes where lines intersect, such as the common connection point of the DC voltage feedback voltage divider network. In some embodiments, the analysis process compares the potential risks of different nodes; for example, high-current-density nodes are preferentially selected as critical test points compared to low-current-density nodes. At each of the selected critical test points, the operator uses a non-conductive adhesive, such as epoxy resin, to attach the sensing unit of a miniature temperature and humidity sensor to the surface of the PCB, ensuring close contact between the sensing unit and the board surface without causing an electrical short circuit. Next to the sensing unit of the miniature temperature and humidity sensor, the operator solders the two contact points of the impedance measurement probe to the start and end pads of the electrical circuit where the critical test point is located, for example, to the two end pads of the same power inductor. The temperature and humidity readings of each miniature temperature and humidity sensor are calibrated by placing the sensor in a standard environment with known precise temperature and humidity, reading its output and comparing it with the standard values. Deviations are corrected using the built-in calibration program to ensure that the measurement error is within acceptable limits, such as a temperature error within ±0.5 degrees Celsius. The zero-point drift of each impedance measurement probe is calibrated by measuring the impedance of a reference circuit known to be close to an ideal short circuit without applying any electrical stress load, and adjusting the measurement reading to match the theoretical impedance value of the reference circuit.

[0022] In practice, pre-set electrical and thermal stress loads are applied to the power management printed circuit board (PCB) with the sensors installed. The PCB is connected to a programmable power supply and electronic load system. The voltage amplitude and current waveform of the electrical stress load are set according to its rated operating parameters; for example, the input voltage is set to 12V DC, and the electronic load is programmed to simulate a dynamic current waveform with a peak current of 5A. The programmable power supply and electronic load system is turned on, and the set voltage amplitude and current waveform are applied to the power input port of the PCB. Simultaneously with the application of the electrical stress load, the entire PCB is placed inside the environmental chamber of a programmable temperature chamber. According to a pre-set aging test profile, the temperature of the environmental chamber of the programmable temperature chamber is controlled to cycle through a set heating, holding, and cooling curve to form a thermal stress load. An exemplary profile is a temperature increase from 25 degrees Celsius to 85 degrees Celsius at a rate of 10 degrees Celsius per minute, a holding time of 120 minutes, and then a decrease to -10 degrees Celsius at a rate of 5 degrees Celsius per minute, repeated cyclically. It is understandable that the specific parameters of the aging test profile can be adjusted according to product specifications, and there is a comparison between high-stress and low-stress profiles in terms of accelerated aging effects. The time from the start of applying electrical and thermal stress loads is recorded as the time reference origin for the entire aging test. Optionally, during the application of the load, the Joule heat power density of the main heat source areas on the printed circuit board can be estimated in real time using a formula:

[0023] in: This represents the Joule heat power density. This indicates the applied input voltage amplitude. This represents the instantaneous value of the load current. This represents the power loss coefficient for that region, which can be obtained from circuit principle analysis and is used to help understand the source of thermal stress. In some embodiments, the applied electrical stress load can be set to different modes such as constant load and cyclic impact load to compare the response of the printed circuit board under different electrical stress conditions.

[0024] In one embodiment of the present invention, taking the aging test of a power management printed circuit board as an example, see [reference needed]. Figure 3After recording the time reference point for the aging test, all installed miniature temperature and humidity sensors were immediately activated. A total of eight miniature temperature and humidity sensors were used, and they acquired data at a set sampling frequency of once per second. Each miniature temperature and humidity sensor independently collected the temperature reading on the surface of the printed circuit board at its attachment location and the relative humidity reading of the surrounding microenvironment. For example, sensor number one, located below the power MOSFET, collected an initial temperature reading of 28.5 degrees Celsius and a humidity reading of 45.2%RH. It is understandable that the readings collected by miniature temperature and humidity sensors in different spatial locations will differ. Sensor number two, located near the inductor, might collect an initial temperature of 30.1 degrees Celsius, while sensor number three, located near the input filter capacitor, might collect an initial humidity of 43.8%RH. This difference in readings reflects the non-uniformity of the initial state of the printed circuit board surface.

[0025] In a specific implementation, temperature readings from all eight miniature temperature and humidity sensors are arranged according to their spatial layout, forming an instantaneous surface temperature distribution matrix at each sampling moment. The spatial layout information of the miniature temperature and humidity sensors is pre-stored in a configuration file; for example, the positions of sensors one through eight are mapped to a 2x4 grid matrix. At a sampling moment, the eight collected temperature readings are filled into the corresponding elements of the matrix according to their positional mapping, forming a 2×4 numerical matrix, which is the instantaneous surface temperature distribution matrix for that sampling moment. In another implementation, humidity readings from all miniature temperature and humidity sensors are arranged according to the exact same spatial layout mapping as the temperature readings, forming an instantaneous environmental humidity distribution matrix at each sampling moment. Similarly, the eight humidity readings are filled into a corresponding 2×4 matrix, forming the instantaneous environmental humidity distribution matrix for that moment. In some embodiments, the layout of the miniature temperature and humidity sensors can adopt other array forms, such as a 3x3 layout, resulting in a 3×3 instantaneous distribution matrix, which contrasts with the aforementioned 2x4 layout in terms of data dimension.

[0026] Following the sampling time sequence, continuous instantaneous surface temperature distribution matrices and instantaneous ambient humidity distribution matrices are stored to form surface temperature distribution data and ambient humidity distribution data that vary over time. Optionally, the storage format is a time-series database, with each record containing a timestamp, a temperature matrix, and a humidity matrix. During continuous data acquisition, the average value of all sensor temperature readings at any given time can be calculated for subsequent analysis. The calculation formula is as follows:

[0027] in: This represents the estimated average temperature on the surface of the printed circuit board at time t, where N represents the total number of deployed miniature temperature and humidity sensors. This represents the temperature reading collected by the i-th miniature temperature and humidity sensor at time t. In some embodiments, the sampling frequency of the miniature temperature and humidity sensor can be adjusted, for example, using a low-frequency sampling of once per second and a high-frequency sampling of ten times per second. The temperature and humidity data sequences obtained at different sampling frequencies show a contrast in temporal resolution. Optionally, the raw readings can be filtered to suppress noise before storage, but the raw data will be saved synchronously. As the aging test proceeds, the surface temperature distribution data and ambient humidity distribution data continuously record the changes in the surface state of the printed circuit board. It is understood that during the heating phase of thermal stress load, most elements in the instantaneous surface temperature distribution matrix will show an upward trend, while in the humidity-controlled environment chamber, the elements in the instantaneous ambient humidity distribution matrix may remain relatively stable or show regular changes.

[0028] In one embodiment of the present invention, taking the aging test of a power management printed circuit board as an example, after recording the time reference origin of the aging test, the impedance analyzers corresponding to all impedance measurement probes are immediately activated. The number of impedance analyzers is consistent with the number of key test point loops, which is five loops in this example. The fixed sampling interval of the impedance analyzers is set to match the sampling period of the miniature temperature and humidity sensor, both being once per second. At each sampling interval, the impedance analyzer sends a sinusoidal excitation signal with a specific frequency and amplitude to the connected impedance measurement probe, and simultaneously measures the response signal of the electrical circuit where the key test point is located to the excitation signal. Based on the excitation signal and the response signal, the complex impedance value of the electrical circuit at the current moment is calculated. Optionally, the complex impedance value is:

[0029] in: Represents complex impedance, This represents a complex voltage vector representing the excitation signal. The complex current vector representing the response signal is used to extract the magnitude of the complex impedance value as the primary recording parameter, i.e., the impedance amplitude. It is understandable that for different electrical circuits, such as the power main circuit and the signal feedback circuit, the measured impedance amplitudes can differ significantly in magnitude. The impedance amplitude of the power circuit is typically in the milliohm range, while the impedance amplitude of the signal circuit may be in the hundreds of ohms range. The impedance amplitude measured at each sampling time and each key test point circuit is bound and stored with the circuit identification information and a precise sampling timestamp. Following the sampling time sequence, the impedance amplitude sequences of all five circuits are organized to form a dataset of continuous changes in circuit impedance over time. In some embodiments, the data can be stored in a table format indexed by time and with circuit identification as the column name.

[0030] In the specific implementation, multi-physics data fusion and alignment are performed. A timestamp corresponding to the sampling time is appended to each temperature and humidity reading from the miniature temperature and humidity sensor, with millisecond-level accuracy. Similarly, a timestamp corresponding to the sampling time is appended to each impedance amplitude reading from the impedance analyzer, also with millisecond-level accuracy. A unified time axis is established starting from the time reference origin of the aging test. The smallest scale of the time axis is the period of the highest sampling frequency among all sensors. In this example, the sampling frequency of the miniature temperature and humidity sensor and the impedance analyzer is the same, so the smallest scale of the time axis is one second. The time-stamped surface temperature distribution data, ambient humidity distribution data, and continuous loop impedance change data are mapped onto the unified time axis. This mapping process involves assigning each data point to a corresponding discrete time slot on the time axis based on its timestamp. For each aligned point on the time axis, the surface temperature distribution matrix, ambient humidity distribution matrix, and impedance amplitude of all loops at that point are extracted and encapsulated into a structured data frame. Optionally, a data frame may contain the following fields: timestamp, temperature matrix, humidity matrix, and an array containing five impedance amplitudes. All data frames arranged in chronological order constitute a multiphysics synchronous monitoring dataset. In some embodiments, data fusion can be performed in real time during the acquisition process or in batches after the acquisition is completed. There is a comparison between the two methods in terms of implementation complexity and real-time performance.

[0031] In one embodiment of the present invention, anomaly feature identification is performed on the aging process of a power management printed circuit board based on a pre-constructed multi-physics synchronous monitoring dataset. A continuous surface temperature distribution matrix sequence ordered by time is extracted from the multi-physics synchronous monitoring dataset. The average temperature of the entire printed circuit board at each sampling moment is calculated. The average temperature is calculated based on the readings of all miniature temperature and humidity sensors at that moment. Using the calculated average temperature as a benchmark, all sensor locations whose temperatures exceed the benchmark value by a preset threshold are identified. The preset threshold is set to be 10 degrees Celsius above the average temperature. In continuous sampling moments, sensor locations whose temperatures exceed the preset threshold are tracked. Areas with continuously exceeding the temperature and spatial continuity are marked as candidate abnormal hot zones. Spatial continuity is determined based on the adjacency of sensor installation locations. The impedance amplitude sequence of each electrical circuit is extracted from the multi-physics synchronous monitoring dataset as a sequence of changes over time. A change point detection algorithm is applied to each impedance amplitude sequence to identify sampling points where the impedance amplitude experiences a step increase or decrease; these sampling points are the impedance abrupt change points. In some embodiments, the preset threshold can be set as a dynamic threshold, such as the average temperature plus twice the standard deviation, which contrasts with a fixed 10-degree Celsius threshold in application. It is understood that the calculation of the average temperature reflects the overall thermal state of the printed circuit board, while sensor locations with temperatures exceeding the average threshold indicate localized overheating.

[0032] For an impedance amplitude sequence of an electrical circuit, a sliding window is used to calculate the local statistical characteristics of the sequence, including the local mean and local standard deviation. The local statistical characteristics of adjacent sliding windows are compared, and their differences are calculated. When the difference exceeds a set sensitivity threshold, the time at the boundary between adjacent sliding windows is recorded as a potential abrupt change point. All recorded potential abrupt change points are verified to check whether the impedance amplitude sequence before and after the potential abrupt change point conforms to a step change pattern, rather than a slow drift. Potential abrupt change points that do not conform to the step change pattern are excluded, and the points confirmed as step changes are retained; their corresponding sampling times are the impedance abrupt change times of the electrical circuit. The above steps are repeated for the impedance amplitude sequences of all electrical circuits to obtain the set of impedance abrupt change times for each circuit. Optionally, the change point detection can use the cumulative sum (CUSUM) algorithm, and one calculation step in the algorithm can be expressed as:

[0033] in: This represents the cumulative sum statistic at time t. This represents the cumulative total and statistical value from the previous moment. This represents the observed impedance amplitude at time t. This represents the mean of the target process (such as the local mean before mutation). The standard deviation of the target process is represented by... It is an allowed offset parameter. This indicates that the larger of the two values ​​is selected as the result. When When the preset control limit H is exceeded, a signal indicating a change point is emitted. In some embodiments, the width of the sliding window can be adjusted; a wider sliding window is more sensitive to slow trends, while a narrower sliding window is more sensitive to rapid steps, and there is a difference in detection performance between the two. Referring to Table 1, a simplified example is shown illustrating the process of identifying the location of the high-temperature sensor at a certain moment.

[0034] Table 1: Example of comparing the temperature of each sensor with the average temperature at a certain sampling time

[0035] In practical implementation, according to the example in Table 1, if the temperature readings of sensors 2, 3, and 5 exceed the average temperature and reach a preset threshold, they are identified as high-temperature points. In a continuous time series, if sensors 2, 3, and 5 are spatially adjacent and their temperatures consistently exceed the limit, the area they cover is marked as a candidate abnormal hot zone. It is understood that candidate abnormal hot zones are dynamically changing; their location, size, and temperature levels may vary at different stages of the aging test. Optionally, after detecting changes in the impedance amplitude sequence, results similar to "loop A detected an impedance step increase at timestamp t=3024 seconds" can be obtained. Impedance abrupt changes in multiple loops may occur simultaneously or dispersedly.

[0036] See Figure 4 This is a bar chart comparing temperature, humidity, and impedance characteristics at different stages of PCB aging testing. The average temperature gradually increases with the aging process, reflecting the overall thermal accumulation effect of the PCB under continuous stress. The average humidity gradually decreases with the aging process, related to the chamber circulation or evaporation of moisture on the PCB surface. The average impedance gradually increases with the aging process, reflecting the performance degradation of electrical circuits due to aging. From the initial to the later stages, temperature and impedance increase synchronously, while humidity decreases synchronously, forming a clear aging process characteristic. Starting from the middle stage (when anomalies appear), the rate of increase in temperature and impedance, as well as the rate of decrease in humidity, both accelerate significantly, consistent with the description in the patent of "abnormal hot zones and concentrated impedance abrupt changes." There is a strong positive correlation between temperature increase and impedance increase, while humidity decrease serves as supporting evidence and can be used to locate potential failure areas on the PCB.

[0037] In one embodiment of the present invention, after identifying abnormal features based on a multi-physics synchronous monitoring dataset, all marked candidate abnormal hot zones and all impedance abrupt change times detected from all loops are obtained. In specific implementations, candidate abnormal hot zones may include multiple regions marked at different stages of aging testing, each region having a range of start and end times. Impedance abrupt change times are a discrete list of time points. On the timeline of the multi-physics synchronous monitoring dataset, the impedance abrupt change times are compared with the time range of candidate abnormal hot zones to determine whether the impedance abrupt change time falls within the duration of a candidate abnormal hot zone, or whether it is sufficiently close to the start or end time of that duration. Candidate abnormal hot zones that coincide with or are adjacent to the impedance abrupt change times in time are selected as highly correlated abnormal hot zones. It can be understood that temporal coincidence means that the impedance abrupt change and the thermal anomaly may be caused by the same physical failure process, while temporal proximity may indicate a causal or co-occurring relationship. Identify electrical loops that are spatially overlapping or adjacent to the anomalous hot areas that are highly correlated with the thermal anomaly. These electrical loops are those that are related to both thermal anomalies and impedance mutations. In some embodiments, the spatial relationship is determined by comparing the location of the anomalous hot area on the printed circuit board layout with the wiring path of the electrical loop. Loop paths that pass through the hot area are considered overlapping, and loop paths that are close to the boundary of the hot area are considered adjacent.

[0038] In practical implementation, the physical area covered by highly correlated abnormal hot zones, along with the wiring area occupied by related electrical circuits on the printed circuit board, are collectively defined as the initial potential failure area. On the layout diagram of the printed circuit board, the sensor installation locations corresponding to all highly correlated abnormal hot zones are marked, and the physical extent of their impact is estimated using a thermal diffusion model, with each location as the center. Optionally, a simplified thermal diffusion model can estimate the impact range centered on the sensor location with a radius of... A circular region, wherein:

[0039] in: This represents the estimated heat-affected radius. A coefficient representing the relationship between the thermal conductivity of the printed circuit board substrate. This represents the heat power dissipation near the sensor location, estimated based on the electrical stress load. This indicates the peak temperature recorded by the sensor. This represents a set temperature threshold. All estimated affected physical areas are superimposed to obtain a set of failure regions caused by thermal anomalies. On the printed circuit board layout, all relevant electrical circuit routing paths are marked, including conductors, vias, and pads. All marked routing paths are extended to both sides by a certain width to form a set of failure regions related to electrical circuits. The extension width can be set according to process capabilities and typical defect sizes. The set of failure regions caused by thermal anomalies and the set of failure regions related to electrical circuits are then combined. The physical area on the printed circuit board covered by the union is jointly defined as the initial potential failure region. In some embodiments, the union operation can be performed using Boolean operations on the graphical representations of the two region sets using image processing software. It is understood that the set of failure regions caused by thermal anomalies may be one or more irregular shapes, while the set of failure regions related to electrical circuits is typically a strip-shaped area along the routing path; the union result of the two is contrasting in shape and coverage.

[0040] See Figure 5 This is a time-series graph showing the surface temperature and ambient humidity changes throughout the 100-hour PCB aging test, visually illustrating the multiphysics characteristics. Surface temperature rises slowly from approximately 35°C to around 40°C from 0–30 hours, representing the initial stable phase. From 30–40 hours, an abnormal hot zone (red shading) emerges, with temperatures surging to 50–55°C, a typical sign of localized overheating. From 40–100 hours, temperatures gradually decrease and stabilize in the 30–38°C range, indicating the later deterioration phase after the abnormal hot zone subsides. Ambient humidity fluctuates dramatically, oscillating between 30–75%RH, highly correlated with chamber circulation or environmental disturbances. From 60–70 hours, a high humidity zone (blue shading) emerges, consistently maintaining above 50%RH, peaking near 65%RH. A weak negative correlation exists between humidity and temperature; humidity is significantly lower during the high-temperature phase (30–40 hours), while temperature remains low during the high-humidity phase (60–70 hours).

[0041] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A printed circuit board aging test method based on intelligent sensing technology, characterized in that, Includes the following steps: Select multiple key test points on the printed circuit board to be tested, and install a miniature temperature and humidity sensor and an impedance measurement probe at each key test point. A pre-set electrical stress load and thermal stress load are applied to the printed circuit board to bring it into a simulated working state; The miniature temperature and humidity sensor continuously collects surface temperature distribution data and ambient humidity distribution data of the printed circuit board under simulated working conditions. The impedance measurement probe is used to synchronously collect data on the continuous change of loop impedance between the key test points at fixed sampling intervals while stress load is applied. The collected surface temperature distribution data, ambient humidity distribution data and loop impedance continuous change data are fused and aligned according to a unified timestamp to generate a multi-physics field synchronous monitoring dataset. Based on the multi-physics synchronous monitoring dataset, the locations of abnormal hot spots and the moments of impedance abrupt changes in the printed circuit board during the aging process are identified. Based on the location of the abnormal hot zone and the time of impedance change, the initial potential failure area of ​​the printed circuit board is located.

2. The printed circuit board aging test method based on intelligent sensing technology according to claim 1, characterized in that, The installation of miniature temperature and humidity sensors and impedance measurement probes at each key test point specifically includes: Analyze the circuit schematic and layout diagram of the printed circuit board to identify nodes with high current density, concentrated heat sources, or intersecting lines as the basis for selecting key test points. At each selected key test point, the sensing unit of the miniature temperature and humidity sensor was attached to the surface of the printed circuit board using a non-conductive adhesive. Next to the sensing unit of the miniature temperature and humidity sensor, the two contact points of the impedance measurement probe are soldered to the start and end pads of the electrical circuit where the key test point is located, respectively. Calibrate the temperature and humidity readings of each of the miniature temperature and humidity sensors to ensure that their measurement errors are within the allowable range; The zero-point drift of each impedance measurement probe is calibrated to ensure that the impedance measurement reading is close to the theoretical value under no-load conditions.

3. The printed circuit board aging test method based on intelligent sensing technology according to claim 2, characterized in that, The application of pre-set electrical and thermal stress loads to bring it into a simulated working state specifically includes: The printed circuit board is connected to a programmable power supply and electronic load system, and the voltage amplitude and current waveform of the electrical stress load are set according to its rated operating parameters. Turn on the programmable power supply and electronic load system, and apply the set voltage amplitude and current waveform to the power supply input port of the printed circuit board; While applying the electrical stress load, the entire printed circuit board is placed inside the environmental chamber of a programmable temperature chamber. According to the preset aging test profile, the ambient temperature of the programmable temperature chamber is controlled to cycle according to the set heating, holding and cooling curves to form the thermal stress load; Record the time from the start of applying the electrical stress load and thermal stress load as the time reference origin for the entire aging test.

4. The printed circuit board aging test method based on intelligent sensing technology according to claim 3, characterized in that, Continuously collect surface temperature distribution data and ambient humidity distribution data of the printed circuit board under simulated working conditions, specifically including: After the time reference origin of the aging test, all miniature temperature and humidity sensors are activated to collect data at the set sampling frequency. Each of the miniature temperature and humidity sensors independently collects the temperature reading on the surface of the printed circuit board at its attachment location and the relative humidity reading of the surrounding microenvironment; The temperature readings from all the miniature temperature and humidity sensors are arranged according to the spatial layout of the sensors to form an instantaneous surface temperature distribution matrix at each sampling moment. The humidity readings from all the miniature temperature and humidity sensors are arranged in the same spatial layout to form an instantaneous environmental humidity distribution matrix at each sampling time. According to the sampling time sequence, the continuous instantaneous surface temperature distribution matrix and instantaneous ambient humidity distribution matrix are stored to form surface temperature distribution data and ambient humidity distribution data that change over time.

5. The printed circuit board aging test method based on intelligent sensing technology according to claim 4, characterized in that, The method of synchronously acquiring continuous loop impedance change data between the key test points at fixed sampling intervals specifically includes: After the time reference origin of the aging test, start the impedance analyzer corresponding to all impedance measurement probes and set a fixed sampling interval that matches the sampling period of the miniature temperature and humidity sensor. At each sampling interval, the impedance analyzer sends an excitation signal to the impedance measurement probe and measures the response signal of the electrical circuit where the key test point is located; Based on the excitation signal and response signal, the complex impedance value of the electrical circuit at the current moment is calculated, and its impedance amplitude is extracted as the main recording parameter. The impedance amplitude measured at each sampling time and each key test point loop is bound and stored with the loop identification information and sampling timestamp; According to the sampling time sequence, the impedance amplitude sequence of all key test point loops is organized to form a dataset of continuous loop impedance changes over time.

6. The printed circuit board aging test method based on intelligent sensing technology according to claim 5, characterized in that, The collected surface temperature distribution data, ambient humidity distribution data, and continuous loop impedance change data are fused and aligned according to a unified timestamp to generate a multi-physics synchronous monitoring dataset, which specifically includes: For each temperature and humidity reading from the miniature temperature and humidity sensor, append a timestamp corresponding to the sampling time; For each impedance amplitude reading from the impedance analyzer, append a timestamp corresponding to the sampling time; Starting from the time reference origin of the aging test, a unified time axis is established, with the smallest scale of the time axis being the period of the highest sampling frequency among all sensors. The time-stamped surface temperature distribution data, ambient humidity distribution data, and continuous change data of loop impedance are respectively mapped onto the unified time axis; For each aligned time point on the time axis, the surface temperature distribution matrix, the ambient humidity distribution matrix, and the impedance amplitude of all loops at that time point are extracted and encapsulated into a data frame. The data frames of all time points are arranged in sequence to form a multi-physics synchronous monitoring dataset.

7. The printed circuit board aging test method based on intelligent sensing technology according to claim 6, characterized in that, Based on the aforementioned multi-physics synchronous monitoring dataset, the locations of abnormal hot spots and the moments of impedance abrupt changes during the aging process of printed circuit boards are identified, specifically including: Extract a continuous surface temperature distribution matrix sequence from the multiphysics synchronous monitoring dataset; Calculate the average temperature of the entire printed circuit board at every moment, and use the average temperature as a reference to identify all sensor locations where the temperature exceeds the reference value by a preset threshold. During continuous sampling times, the sensor positions above the reference value are tracked, and areas where the temperature continuously exceeds the limit and are spatially continuous are marked as candidate abnormal hot areas; From the multi-physics synchronous monitoring dataset, extract the time sequence of impedance amplitude changes for each electrical circuit; A change point detection algorithm is applied to each impedance amplitude sequence to find the sampling time point when the impedance amplitude experiences a step increase or decrease. The sampling time point is the impedance change moment.

8. The aging test method for printed circuit boards based on intelligent sensing technology according to claim 7, characterized in that, Based on the location of the abnormal hot zone and the time of impedance abrupt change, the initial potential failure area of ​​the printed circuit board is located, specifically including: Obtain all labeled candidate anomalous hot zones, as well as all impedance abrupt change moments detected from all loops; On the time axis of the multiphysics synchronous monitoring dataset, the impedance abrupt change time is compared with the time range of the candidate abnormal hot zone. Candidate anomalous hot regions that coincide with or are adjacent to the impedance abrupt change time in time are selected as highly correlated anomalous hot regions. Obtain the electrical circuits that are spatially overlapping or adjacent to the anomalous thermal areas that are highly correlated with the thermal anomaly; these electrical circuits are those that are related to both thermal anomalies and impedance mutations. The physical area covered by the highly correlated abnormal hot zone, and the wiring area occupied by the related electrical circuits on the printed circuit board, are collectively defined as the initial potential failure area.

9. The printed circuit board aging test method based on intelligent sensing technology according to claim 8, characterized in that, The step of applying a change point detection algorithm to each impedance amplitude sequence to identify the sampling point where the impedance amplitude experiences a step increase or decrease specifically includes: For an impedance magnitude sequence of an electrical circuit, a sliding window is used to calculate the local statistical characteristics of the sequence, including the local mean and local standard deviation. Compare the local statistical characteristics of adjacent sliding windows, calculate their difference, and when the difference exceeds the set sensitivity threshold, record the time at the boundary of adjacent sliding windows as a potential mutation point. Verify all recorded potential abrupt changes and check whether the impedance amplitude sequence before and after the potential abrupt change satisfies the pattern of a step change rather than a slow drift. Eliminate potential abrupt change points that do not conform to the step change pattern, retain the points that are confirmed to be step changes, and the corresponding sampling time is the impedance change time of the electrical circuit. Repeat the steps for the impedance magnitude sequence of all electrical circuits to obtain the set of impedance change moments for each circuit.

10. The aging test method for printed circuit boards based on intelligent sensing technology according to claim 9, characterized in that, The physical area covered by the highly correlated abnormal hot zone, and the wiring area occupied by the related electrical circuits on the printed circuit board, are jointly defined as the preliminary potential failure area, specifically including: On the layout diagram of the printed circuit board, mark the sensor installation locations corresponding to all highly correlated abnormal hot areas, and estimate the physical range of their influence based on the thermal diffusion model, with each location as the center. By superimposing all the estimated affected physical areas, a set of failure areas caused by thermal anomalies is obtained; On the layout diagram of the printed circuit board, mark the wiring paths of all relevant electrical circuits, including wires, vias and pads. Extend all marked wiring paths to both sides by a certain width to form a set of failure areas related to the electrical circuit; The set of failure areas caused by thermal anomalies and the set of failure areas related to the electrical circuits are combined by taking the union operation. The physical area on the printed circuit board covered by the union is then defined as the preliminary potential failure area.

Images