PCB hot-pressing cooling drilling cross-process stress coordination control method and system

Abstract

The application discloses a PCB hot-pressing cooling drilling cross-process stress coordination control method and system, and belongs to the technical field of printed circuit board manufacturing. The method comprises the following steps: obtaining the solidification temperature distribution data of the board surface in the hot-pressing stage; monitoring the warping amount change and temperature change in real time in the cooling process, calculating the cooling rate, warping change rate, stress release degree and dispersion, and judging the stress abnormality source according to the correlation between the stress release degree, the cooling rate and the hot-pressing solidification temperature distribution; dynamically adjusting the cooling parameters according to the monitoring results and the abnormality source judgment results to form real-time closed-loop control; taking the board ID as an index to jointly record the hot-pressing and cooling stage data and transfer the data to the drilling process; the drilling process generates differentiated drilling compensation parameters according to the stress release state and the hot-pressing solidification state of each board and executes drilling compensation; after the drilling is completed, the precision deviation is attributed to different layers and then fed back to the hot-pressing process and the cooling process. The application realizes hot-pressing cooling drilling whole-process stress coordination control and cross-process closed-loop optimization.

Description

Technical Field

[0001] This invention relates to the field of printed circuit board (PCB) manufacturing technology, specifically to a method and system for cross-process stress coordination control between the three stages of hot pressing, cooling, and drilling in the PCB lamination process. Background Technology

[0002] With the development of technologies such as 5G communication and artificial intelligence, the number of PCB layers is increasing and the material systems are becoming more complex, leading to ever-increasing demands on manufacturing precision. The lamination and drilling processes are key steps that determine product yield. The internal stress distribution during the lamination stage determines the dimensional stability of the board, while the drilling precision directly affects the reliability of interlayer connections.

[0003] In existing processes, hot pressing, cooling, and drilling are typically managed as independent steps. During hot pressing, the heating rate, curing temperature, and holding time are controlled by a program. Some solutions use multi-point thermocouples to monitor the temperature distribution on the board surface, but the data is only used for internal process judgment. The cooling stage generally employs open-loop control with a preset cooling curve, executing a fixed program without adjustment based on the actual condition of the board. The drilling process, however, involves X-ray inspection of the first piece to measure expansion and contraction values, followed by uniform compensation of the drilling program; the same compensation value is used for the same batch.

[0004] The aforementioned "each section manages one part" model has significant drawbacks. In the hot-pressing stage, uneven temperature distribution within the press leads to variations in the cured areas, which are not specifically compensated for during the cooling stage. If insufficiently cured areas are cooled at the same rate, stress can easily be "frozen," causing stress release during subsequent drilling and resulting in hole position displacement. The cooling stage operates in an open-loop manner according to a preset curve, unable to adaptively adjust based on actual stress release. Drilling compensation relies on static expansion and contraction values ​​after cooling, failing to utilize dynamic data from the hot-pressing and cooling processes to predict deformation risks in advance. Boards located at different positions within the same batch exhibit significant differences in heating and cooling histories, but existing compensation strategies employ a "same batch, same compensation" approach, which cannot address inter-board differences. When drilling accuracy deviates, it's impossible to trace the root cause to uneven hot-pressing curing, insufficient cooling stress release, or a problem with the drilling itself. Therefore, this invention provides a cross-process stress collaborative control method and system for PCB hot-pressing cooling drilling, aiming to break down data barriers between the three processes and achieve full-process stress collaborative control and cross-process closed-loop optimization. Summary of the Invention

[0005] To overcome the technical defects of existing technologies where hot pressing, cooling and drilling processes operate independently and lack data linkage and collaborative control, this invention provides a cross-process stress collaborative control method and system for PCB hot pressing, cooling and drilling.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In one aspect, the present invention provides a method for cross-process stress collaborative control of PCB hot-pressing cooling drilling, comprising the following steps:

[0008] S1: Obtain the temperature distribution data of the board surface during the hot pressing stage and the baseline state data before the start of cooling.

[0009] S2: During the cooling process, monitor the warpage and temperature changes of each area of ​​the board in real time, calculate the cooling rate, warpage rate, stress release degree and stress release degree dispersion of each area, and determine the source of stress anomalies based on the correlation between stress release degree, cooling rate and hot pressing curing temperature distribution.

[0010] S3: Based on the stress release monitoring results and the judgment results of the abnormality source, dynamically adjust the cooling parameters to form a real-time closed-loop control until cooling is completed.

[0011] S4: Using the plate ID as an index, jointly record the hot pressing stage data and cooling stage data of each plate, and transfer the above data to the drilling process.

[0012] S5: The drilling process receives data from the hot pressing stage and the cooling stage. Based on the stress release state and hot pressing curing state of each plate, it generates plate-level and region-level differentiated drilling compensation parameters and performs drilling compensation. After drilling is completed, the hole position accuracy is checked, and the accuracy deviation is attributed to different layers and fed back to the hot pressing process and the cooling process respectively, forming a closed loop across the entire process.

[0013] Preferably, step S1 specifically includes:

[0014] S11: Before the end of the heat preservation stage during hot pressing, acquire curing temperature distribution data for each area of ​​the PCB board using a multi-point thermocouple array, and generate a curing temperature distribution matrix. And calculate the temperature difference on the plate surface:

[0015]

[0016] in This is the highest curing temperature of the board surface. This is the lowest curing temperature for the board surface; at the same time, the position information of each board in the press is recorded.

[0017] S12: Obtain PCB design information, including stack-up parameters and drilling design data, and based on the differences in the thermal expansion coefficients of materials in different regions of the stack-up parameters, combine this with the curing temperature distribution recorded in S11. The theoretical cooling curves required for each region to cool from the curing temperature to room temperature were pre-calculated, and the board surface was divided into sections. For the grid, a lower upper limit for the cooling rate is set for areas with lower curing temperatures.

[0018] S13: Before hot pressing is completed and cooling begins, the initial warpage distribution on the board surface is obtained using a laser displacement sensor or a structured light 3D scanner, and an initial warpage distribution matrix is ​​generated. .

[0019] S14: Acquire initial temperature distribution data at multiple points on the plate surface using a thermocouple array or infrared thermal imager, and generate an initial temperature distribution matrix. .

[0020] Furthermore, step S2 specifically includes:

[0021] S21: Continuously sample the temperature data of each monitoring point at a set frequency and calculate the real-time cooling rate of each monitoring point:

[0022]

[0023] in For the first Each monitoring point is Cooling rate at any time The current temperature. The sampling interval is denoted as .

[0024] S22: Continuously sample warpage data at each monitoring point using the same frequency, and calculate the warpage rate of change at each monitoring point:

[0025]

[0026] in For the first Each monitoring point is The rate of change of warping at any given moment. This represents the current warpage.

[0027] S23: Based on warpage data, calculate the stress release degree of each region:

[0028]

[0029] in For the first Each monitoring point is Stress relief at any given time The initial rate of warpage at the start of cooling; A higher value indicates more complete release of internal stress.

[0030] S24: Calculate the dispersion of stress release degree at each monitoring point from the average stress release degree on the plate surface:

[0031]

[0032] in for The average stress relief value at all monitoring points on the plate at any given time:

[0033]

[0034] in This represents the total number of monitoring points.

[0035] S25: The system determines whether to trigger intervention when the stress release level in a certain area is lower than the preset target value. And the cooling rate is higher than the preset threshold. When this occurs, the intervention mechanism is triggered:

[0036]

[0037] After intervention is triggered, the source of the anomaly is determined by a combined attribution of hot pressing and cooling:

[0038]

[0039] in The discreteness threshold, This is the temperature difference threshold.

[0040] Furthermore, step S3 specifically includes:

[0041] S31: When intervention is triggered, the system generates a cooling rate adjustment command, and the adjustment amount is calculated using a PID control algorithm.

[0042]

[0043] in For stress relief deviation, , , These are control parameters.

[0044] S32: Determine the final cooling rate adjustment based on the anomaly source determination result: If the anomaly source is determined to be uneven hot-pressing curing, then add a compensation coefficient to the adjustment amount calculated by the PID controller. This further reduces the cooling rate of the area; The calculation method is as follows ,in This refers to the curing temperature during the hot pressing stage in this area. This is the average curing temperature of the board surface; if the source of the anomaly is determined to be improper cooling parameters, the adjustment output by the PID calculation will be used directly.

[0045] S33: For areas where stress relief has reached the target value, maintain or appropriately accelerate the cooling rate.

[0046] S34: During the cooling process, continuously monitor the changes in stress release to form a real-time closed loop of "monitoring, evaluating, adjusting, and re-monitoring" until cooling is complete and the stress release in each area reaches the target value.

[0047] Furthermore, step S4 specifically includes:

[0048] S41: Using the panel ID as an index, record the hot pressing stage data and cooling stage data for each panel, including the panel surface curing temperature distribution matrix. Temperature difference on the board surface Stress relief degree-time curve Final stress release degree in each region And records of interventions during the cooling process.

[0049] S42: Based on the combined data of hot-pressing curing temperature distribution and cooling warping evolution, predict the residual stress distribution and expected expansion / contraction of each plate after cooling:

[0050]

[0051] in For the first Expected increase or decrease in volume for each region This refers to the curing temperature during the hot pressing stage in this area. For the overall temperature difference of the board surface, For the final stress relief degree, The coefficient of thermal expansion is To cool the total temperature drop, The room temperature is 25℃.

[0052] S43: Transfer the above data to the compensation calculation module of the drilling process using the board ID as the index.

[0053] Furthermore, step S5 specifically includes:

[0054] S51: According to each board Calculate drilling compensation based on hot pressing parameters: for plates with sufficient stress release. And the hot-press curing is uniform Standard borehole compensation value is adopted. For plates with insufficient stress relief Or uneven curing under heat pressing For the sheet metal, calculate the drilling compensation:

[0055]

[0056] in For the first Drilling compensation for plate components. This is the standard borehole compensation value. , , These are the weighting coefficients.

[0057] S52: The compensation amount adopts a differentiated distribution strategy, with different areas of the board distributed according to their respective... and hot pressing curing temperature Generate independent compensation amount :

[0058]

[0059] in For the first Block parts in coordinates The amount of borehole compensation at that location. This represents the final stress relief at that coordinate. This represents the expected expansion or contraction at that coordinate. This represents the hot-pressing curing temperature at that coordinate. This represents the average curing temperature of the board surface.

[0060] Generate a compensation distribution matrix and import it into the CNC system of the drilling equipment to correct the drilling coordinates in real time.

[0061] S53: After drilling is completed, check the actual hole position accuracy, compare the actual offset with the predicted offset, and calculate the prediction deviation:

[0062]

[0063] when When the preset threshold is exceeded, model calibration is automatically triggered. The calibration adopts a hierarchical attribution strategy: if the predicted deviation of the plates at the edge of the press in the same batch is generally large, it is determined to be a temperature uniformity problem in the hot pressing stage, and feedback is given to the hot pressing process to optimize temperature uniformity; if the plates at the center of the press also have deviation, it is determined to be a problem with cooling parameters or model weights, and feedback is given to the cooling process for adjustment. And cooling adjustment strategies; by hierarchical attribution, the drilling accuracy test results are fed back to the hot pressing process and the cooling process respectively, forming a closed-loop optimization across the entire process.

[0064] On the other hand, to achieve the above method, the present invention provides a cross-process stress coordination control system for PCB hot pressing and cooling drilling, comprising the following modules:

[0065] The hot pressing parameter recording module is used to acquire the curing temperature distribution data of each area of ​​the PCB board before the end of the heat preservation stage, calculate the temperature difference of the board surface, and record the position information of each board in the press.

[0066] The cooling stress monitoring module is used to acquire warpage and temperature data of various areas of the PCB board in real time during the cooling process, and continuously sample at a set frequency to output the real-time warpage change rate and real-time cooling rate of each monitoring point.

[0067] The stress release assessment module receives the warpage rate and cooling rate output by the cooling stress monitoring module, as well as the curing temperature distribution data transmitted by the hot pressing parameter recording module, and calculates the stress release degree and stress release degree dispersion at each monitoring point. When the stress release degree is lower than the target value and the cooling rate is higher than the threshold, an intervention trigger signal is output. Based on the comparison results of the dispersion and the dispersion threshold, and the comparison results of the curing temperature deviation and the temperature difference threshold, the abnormality source is determined to be uneven hot pressing curing or improper cooling parameters.

[0068] The cooling parameter dynamic adjustment module is used to receive the intervention trigger signal, stress release deviation and abnormality source judgment result output by the stress release degree assessment module, calculate the cooling rate adjustment amount using PID control algorithm, and superimpose a compensation coefficient when the abnormality originates from uneven hot pressing curing, and output the zoned cooling control command for each area.

[0069] The cross-process data recording and transmission module is used to jointly record the hot pressing stage data and cooling stage data of each plate using the plate ID as an index, and then transmit the above data to the drilling compensation calculation module.

[0070] The drilling compensation calculation module is used to receive the final stress release degree, expected expansion and contraction amount and hot pressing parameters transmitted by the cross-process data recording and transmission module, calculate the differentiated drilling compensation amount for each plate and each area of ​​the plate surface, and output the drilling compensation command.

[0071] The drilling accuracy feedback module is used to obtain the actual hole position accuracy detection results after drilling is completed, calculate the prediction deviation, and output a model calibration trigger signal when the prediction deviation exceeds a preset threshold. The signal is fed back to the hot-pressing parameter recording module and the stress release evaluation module respectively using a hierarchical attribution strategy.

[0072] Compared with the prior art, the present invention has the following beneficial effects:

[0073] (1) Break down the data barriers between the three processes and achieve stress coordination throughout the entire process. This invention establishes a feedforward channel between hot pressing and cooling, a stress state transmission channel between cooling and drilling, and a dual feedback channel between drilling and hot pressing / cooling, so that stress information can form a complete closed loop flow between the three processes, fundamentally solving the problems of control lag and blind decision-making caused by the "information islands" between processes in traditional processes.

[0074] (2) A quantifiable index for stress release is introduced to achieve real-time perception of stress release status and accurate tracing of anomalies. This invention defines the stress release degree. This enables real-time quantitative monitoring of stress release state during the cooling process; by calculating the dispersion... And combined with the temperature deviation of hot pressing curing By performing joint attribution judgment, it is possible to accurately distinguish whether the stress anomaly originates from uneven curing during the hot pressing stage or from improper parameters during the cooling stage, thus avoiding the problems of difficulty in tracing the source of anomalies and the inability to make passive adjustments in traditional methods.

[0075] (3) Achieve feedforward and feedback composite control of cooling parameters to improve stress release adequacy and board surface flatness. This invention uses the curing temperature distribution during the hot pressing stage as the basis for feedforward control. Lower upper limits for cooling rates are pre-set in areas with lower curing temperatures, while simultaneously using real-time stress release... A PID algorithm is used for closed-loop regulation of the feedback signal, and a compensation coefficient is added when the anomaly originates from uneven hot-press curing. Compared to the open-loop preset cooling technology of the prior art, the present invention achieves a composite control of "feedforward prediction + feedback adjustment", which makes stress release more complete and warpage significantly reduced.

[0076] (4) Achieve three-factor differentiated drilling compensation at the plate level and region level, significantly improving drilling accuracy and batch consistency. The drilling compensation amount of this invention... By integrating three factors—hot pressing curing temperature, cooling stress release, and expected expansion and contraction—this invention achieves differentiated compensation between different panels and between different areas of the same panel. Compared to the existing "same batch, same compensation" strategy, this invention significantly improves the batch consistency of drilling accuracy.

[0077] (5) Construct a full-process dual-path closed-loop optimization mechanism to achieve continuous system evolution. This invention predicts drilling accuracy deviation. By employing a hierarchical attribution strategy, feedback is provided to both the hot pressing and cooling processes, forming a closed-loop dual-path system encompassing "hot pressing, cooling, drilling, and hot pressing / cooling." As production batches increase, the control parameters for each process are continuously optimized, making the system more accurate with use, resulting in simultaneous improvements in yield and efficiency. Attached Figure Description

[0078] Figure 1 This is an overall flowchart of the method of the present invention.

[0079] Figure 2 This is a module architecture diagram of the system of the present invention.

[0080] Figure 3 This is a schematic diagram of the combined attribution logic for hot pressing and cooling.

[0081] Figure 4This is a flowchart of the combined feedforward and feedback control for cooling parameters.

[0082] Figure 5 This is a schematic diagram of a dual-path closed loop across all processes. Detailed Implementation

[0083] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0084] Method Implementation Examples

[0085] like Figure 1 As shown, this invention provides a method for cross-process stress coordination control in PCB hot-pressing and cooling drilling, comprising the following steps:

[0086] Step S1: Obtain the surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins.

[0087] In one specific implementation, step S1 includes the following steps:

[0088] S11: Before the end of the heat preservation stage during hot pressing, i.e. when the curing is completed, use a multi-point thermocouple array on the pressing equipment to obtain the curing temperature distribution data of each area on the PCB board.

[0089] Specifically, the thermocouples in this multi-point thermocouple array are arranged at a density of one monitoring point per 100 cm², covering the entire area of ​​the board surface. The curing temperature data of each monitoring point are mapped to two-dimensional coordinates on the board surface to generate a curing temperature distribution matrix. ,in , , and These represent the number of grid divisions along the length and width of the plate, respectively.

[0090] Then, the temperature difference on the plate surface was calculated:

[0091]

[0092] in This is the highest curing temperature among all monitoring points on the board surface. This represents the lowest curing temperature among all monitoring points on the board surface. Simultaneously, the position information of each board within the press, such as whether it's in the center or edge of the press, is recorded to facilitate subsequent tracking of the impact of temperature uniformity on cooling stress release during the hot pressing stage.

[0093] S12: Obtain PCB design information, including stack-up parameters, namely the material type, thickness, coefficient of thermal expansion (CTE) value of each layer, and drilling design data, namely hole coordinates, hole diameter, and number of holes.

[0094] Based on the differences in the thermal expansion coefficients of materials in different regions of the stacked structure, combined with the curing temperature distribution recorded in S11, The theoretical cooling curves required for each region to cool from the curing temperature to room temperature are pre-calculated.

[0095] The specific method is as follows: Divide the plate surface into sections identical to S11. Grid, for each grid node Based on the curing temperature at that node Coefficient of thermal expansion of materials and target stress release degree Calculate the upper limit of the theoretical cooling rate required for this node. This ensures that, under the theoretical cooling curve, the stress release degree in each area of ​​the board surface can reach the target value simultaneously upon completion of cooling. For areas with lower curing temperatures, a lower upper limit for the cooling rate is typically set at the edge of the press to allow for more sufficient stress release time.

[0096] S13: Before hot pressing is completed and cooling begins, the initial warpage distribution of the board surface is obtained using a laser displacement sensor or a structured light 3D scanner.

[0097] Specifically, the PCB board surface is divided into sections identical to S12. The grid is used to measure the Z-axis height of each grid node, generating an initial warping distribution matrix. Warpage is defined as the difference between the maximum and minimum height values ​​within the grid.

[0098]

[0099] S14: Use a thermocouple array or infrared thermal imager to acquire initial temperature distribution data at multiple points on the plate surface, and generate an initial temperature distribution matrix. .

[0100] Step S2: Monitor warping and temperature in real time, calculate various indicators, and determine the source of anomalies.

[0101] In one specific implementation, step S2 specifically includes the following steps:

[0102] S21: During the cooling process, temperature data at each monitoring point is continuously sampled every 3 seconds, and the real-time cooling rate at each monitoring point is calculated.

[0103]

[0104] in For the first Each monitoring point is Cooling rate at any time For the first Each monitoring point is Temperature at any moment min, or 3 seconds, is the sampling interval. Cooling rate. A positive value indicates that the board is cooling down; the larger the value, the faster the cooling.

[0105] S22: Continuously sample the warp data at each monitoring point with the same 3-second sampling interval, and calculate the warp rate of change at each monitoring point:

[0106]

[0107] in For the first Each monitoring point is The rate of change of warping at any given moment. For the first Each monitoring point is The amount of warping over time. min is the sampling interval.

[0108] rate of warpage It reflects the activity level of internal stress release. The higher the rate, the faster the stress is being released. When the rate approaches zero, the stress has been basically released.

[0109] S23: Calculate the stress release degree of each region based on the warp change rate data:

[0110]

[0111] in For the first Each monitoring point is Stress relief at any given time When cooling begins, i.e. Time The initial warp change rate at each monitoring point.

[0112] Stress relief The value ranges from 0% to 100%. At the start of cooling, ,at this time This indicates that the stress has not yet been released; during the cooling process, Gradually decrease, Gradually increase; when When approaching 0, Approaching 100% indicates that the stress has been basically released.

[0113] This indicator enables real-time quantitative monitoring of stress release during the cooling process, providing clear feedback signals for closed-loop control.

[0114] S24: Calculate the dispersion of stress release degree at each monitoring point from the average stress release degree on the plate surface:

[0115]

[0116] in for The average stress relief value at all monitoring points on the plate at any given time:

[0117]

[0118] in This represents the total number of monitoring points.

[0119] Reflects the first The degree to which the stress release behavior in each region deviates from the overall trend of the plate surface. The larger the value, the greater the difference between the stress release state of the region and the average state of the plate surface.

[0120] S25: As Figure 1 As shown, the system determines whether to trigger intervention when the stress release level in a certain area is lower than the target value. And the cooling rate is higher than the threshold. When this occurs, the intervention mechanism is triggered:

[0121]

[0122] Preferably, Take 90%. Specifically, when the stress relief reaches 90%, the influence of residual stress on subsequent drilling accuracy can be ignored, that is, the drilling offset is <15μm. Below this value, residual stress significantly affects drilling accuracy.

[0123] Preferably, Take 2.5℃ / min. Specifically, in the glass transition temperature (Tg) range of approximately 135℃-155℃ for the FR-4 material system, when the cooling rate exceeds 2.5℃ / min, the plate rapidly passes through the Tg range, and the movement of the resin molecular chain segments does not have enough time to fully respond to the temperature change, resulting in the stress being "frozen" within the plate.

[0124] like Figure 1 and Figure 3 As shown, after the intervention is triggered, the system uses a combined attribution method based on thermal pressure and cooling to determine the source of the anomaly. Specifically, it first determines whether there is an abnormal stress release in the area, i.e. .like This indicates that the stress release behavior in this area is consistent with the overall trend of the plate surface and no special treatment is required.

[0125] like Then, it is further determined whether the curing temperature of this area during the hot pressing stage deviates from the average curing temperature of the board surface:

[0126]

[0127] Preferably, the dispersion threshold Take 15%. Specifically, based on the 3σ principle of statistical process control (SPC), when the stress release degree of a certain monitoring point deviates from the average value of the plate surface by more than 15%, that point has exceeded the upper limit of the normal process fluctuation range, and the root cause needs to be traced.

[0128] Preferably, the temperature difference threshold Take 3℃. Specifically, the normal temperature fluctuation range of the hot platen of the press is within ±1.5℃. 3℃ is twice this normal fluctuation range. If it exceeds 3℃, it is judged to be a temperature uniformity problem in the hot pressing stage.

[0129] Step S3: Dynamically adjust cooling parameters to form a real-time closed loop until cooling is complete.

[0130] In one specific implementation, step S3 specifically includes the following steps:

[0131] S31: As Figure 1 As shown, when intervention is triggered, the system generates a cooling rate adjustment command. The adjustment amount is calculated using a PID control algorithm:

[0132]

[0133] in This represents the deviation in stress relief.

[0134] Preferably, , , .

[0135] Specifically, This determines the intensity of the system's response to the deviation. Setting it to 1.0 ensures that the cooling rate adjustment and the deviation value are roughly in a 1:1 ratio, allowing the system response to match the magnitude of the deviation. A value of 0.3 is used to eliminate steady-state error and ensure that the final stress relief converges to the target value; A value of 0.1 is used to suppress overshoot and avoid excessive adjustment of the cooling rate, which could lead to new stress problems. The above parameters are determined by tuning based on the actual response characteristics of the PCB lamination and cooling process.

[0136] S32: As Figure 4As shown, this embodiment employs a combined feedforward and feedback control strategy. When determining the final cooling rate adjustment, the result of the anomaly source determination in step S25 is first processed:

[0137] If the source of the anomaly is determined to be uneven curing under hot pressing, i.e. Then, a compensation coefficient is added to the adjustment amount calculated by the PID controller. This further reduces the cooling rate of the area.

[0138] Compensation coefficient The calculation method is as follows:

[0139]

[0140] Preferably, The value range is 1.10-1.30.

[0141] Specifically, for every 1°C increase in temperature difference, the compensation coefficient increases by 0.1, meaning the cooling rate decreases by an additional 10%. This proportional relationship is based on the linear compensation principle between the degree of curing and the cooling rate. The greater the temperature deviation, the less complete the curing during the hot pressing stage, the lower the resin crosslinking density, and the more difficult the stress release. Therefore, a greater reduction in the cooling rate is needed to compensate for this.

[0142] For example, if the hot-pressing curing temperature in a certain area deviates from the average value by 5°C, then... The cooling rate in this area is further reduced by 50% on top of the PID control. When the temperature difference is 2℃ and does not exceed the 3℃ threshold, no compensation coefficient is added. The final cooling rate adjustment is... .

[0143] If the source of the anomaly is determined to be improper cooling parameters, then the PID control output adjustment should be calculated directly. Execution is performed without adding compensation coefficients.

[0144] S33: For stress relief that has reached the target value In areas where cooling is required, maintain the current cooling rate or appropriately increase the cooling rate to improve production efficiency, but the increased cooling rate must not exceed 2.5℃ / min.

[0145] S34: As Figure 1 As shown, the change in stress release is continuously monitored during the cooling process. After adjusting S3, the monitoring returns to S2 and continues. If... If the target is still not met, the system will re-enter S3 for further adjustment, forming a real-time closed loop of "monitoring, evaluation, adjustment, and re-monitoring" until cooling is complete.

[0146] Cooling completion refers to the point when the stress in each area is fully released. When all reach 90%, the system exits the closed loop and enters S4.

[0147] Step S4: Jointly record the data of the hot pressing and cooling stages, and transfer them to the drilling process using the plate ID as an index.

[0148] In one specific implementation, step S4 includes the following steps:

[0149] S41: As Figure 1 As shown, after cooling is completed, the system uses the board ID as an index to jointly record the hot pressing stage data and cooling stage data of each board.

[0150] The hot-pressing stage data includes: a matrix of curing temperature distribution on the board surface. Temperature difference on the board surface Information on the position of the sheet metal in the press.

[0151] Cooling phase data includes: stress release degree-time curve Temperature-time curve Final stress release degree in each region That is, when cooling is complete The values ​​and intervention records during the cooling process, including the intervention trigger time, the result of the abnormality source judgment, and the adjustment amount.

[0152] S42: Based on the combined data of hot-pressing curing temperature distribution and cooling warping evolution, predict the residual stress distribution and expected expansion / contraction of each plate after cooling:

[0153]

[0154] in For the first Expected increase or decrease in volume for each region This refers to the curing temperature during the hot pressing stage in this area. For the overall temperature difference of the board surface, For the final stress relief degree, The coefficient of thermal expansion is To cool the total temperature drop, i.e. The room temperature is 25℃.

[0155] Preferably, the prediction model A multiple linear regression model was adopted. Specifically, the model was trained using curing temperature during hot pressing, cooling stress release degree, and material CTE as input features, and actual borehole offset as label. The regression coefficients were determined by the least squares method.

[0156] S43: The above data is transmitted to the compensation calculation module of the drilling process via the MES system or industrial Ethernet, indexed by the board ID. Preferably, the data format adopts JSON structured format.

[0157] Step S5: Borehole differential compensation, detection accuracy and hierarchical attribution feedback.

[0158] In one specific implementation, step S5 includes the following steps:

[0159] S51: The drilling process receives data from step S4 and, based on the data from each board... Calculate the borehole compensation amount using hot pressing parameters.

[0160] For plates with sufficient stress relief And the hot-press curing is uniform Standard borehole compensation value is adopted. mm.

[0161] Preferably, The uniformity threshold is set at 5℃. Specifically, when the temperature difference between the board surfaces is within 5℃, the curing uniformity during the hot pressing stage is acceptable, i.e., the curing degree difference is <3%. If the temperature difference exceeds 5℃, compensation is required during the cooling stage. mm represents the historical average expansion and contraction compensation value of this plate type under conditions of full stress release, derived from statistical analysis of production data.

[0162] For plates with insufficient stress relief Or uneven curing under heat pressing For the sheet metal, calculate the drilling compensation:

[0163]

[0164] in For the first Drilling compensation for plate components. , , These are the weighting coefficients.

[0165] Specifically, This means that for every 1% drop in stress relief below the target value, the compensation increases by 0.003 mm. This value is derived from the linear regression slope between stress relief and borehole offset. This means that 80% of the expected increase or decrease in volume is converted into actual compensation, while the remaining 20% ​​is absorbed by the dynamic compensation capability of the CNC system. This indicates that each 1°C temperature difference on the plate surface results in an additional expansion or contraction of approximately 1.2 μm. This value is based on a weighted average of the CTE differences between different materials.

[0166] S52: The compensation amount adopts a differentiated distribution strategy, with different areas of the board distributed according to their respective... and hot pressing curing temperature Generate independent compensation amounts:

[0167]

[0168] in For the first Block parts in coordinates The amount of borehole compensation at that location. This represents the final stress relief at that coordinate. This represents the expected expansion or contraction at that coordinate. This represents the hot-pressing curing temperature at this coordinate.

[0169] After generating the compensation distribution matrix, it is imported into the CNC system of the drilling equipment. During drilling, the CNC system adjusts the compensation distribution matrix according to the current position of the drill bit. The corresponding compensation amount is obtained by looking up the table, and the borehole coordinates are corrected in real time.

[0170] S53: After drilling is completed, the actual hole position accuracy is checked by X-Ray or AOI, and the actual offset is compared with the predicted offset to calculate the prediction deviation.

[0171]

[0172] Preferably, when When the deviation exceeds 15%, model calibration is automatically triggered. Specifically, this threshold is the upper limit of the allowable error for the compensation accuracy of CNC equipment. When the prediction deviation exceeds 15%, it indicates that the prediction model can no longer accurately reflect the actual process state and needs to be recalibrated.

[0173] like Figure 1 and Figure 5 As shown, the calibration employs a hierarchical attribution strategy:

[0174] If the predicted deviation of the plates at the edge of the same batch of medium-pressure presses is generally large, that is... If so, it is determined to be a problem of temperature uniformity during the hot pressing stage, that is... If the temperature difference is too large, feedback is sent to the hot pressing process to optimize temperature uniformity. Specifically, the heating power distribution of the press is adjusted to reduce the temperature difference on the plate surface. Extend the heat preservation time by 5-10 minutes to improve the uniformity of curing.

[0175] If the plate at the center of the press also has a deviation, that is If the problem persists, it is determined to be a cooling parameter or model weight issue, and feedback is sent to the cooling process for adjustment. And cooling adjustment strategy, specifically, if the deviation direction is insufficient compensation will Adjust from 90% to 92%, or if the deviation direction is overcompensated... Adjusted from 90% to 88%, the PID controller was recalibrated. , , Parameters, or update compensation coefficients The reference temperature difference in the calculation formula has been adjusted from 10 to 12.

[0176] The drilling accuracy test results are fed back to the hot pressing and cooling processes through the above-mentioned hierarchical attribution, forming a result as follows: Figure 5 The diagram shows a closed-loop optimization across all processes.

[0177] System Implementation Examples

[0178] like Figure 2 As shown, this invention provides a cross-process stress coordination control system for PCB hot pressing and cooling drilling, comprising the following modules:

[0179] Hot pressing parameter recording module: used to acquire curing temperature distribution data of various areas on the PCB board before the end of the heat preservation stage. Calculate the temperature difference on the plate surface It also records the position information of each plate in the press and stores it using the plate ID as an index.

[0180] Cooling stress monitoring module: Used to acquire warpage and temperature data of various areas of the PCB board in real time during the cooling process, and continuously sample at a frequency of once every 3 seconds, outputting the real-time warpage change rate of each monitoring point. and real-time cooling rate .

[0181] Stress relief assessment module: used to receive the output of the cooling stress monitoring module. and and the hot-pressing parameter recording module transmits Calculate the stress release degree at each monitoring point. and dispersion ;when and Output intervention trigger signal in real time; and according to and If the source of the abnormality is determined to be uneven curing under hot pressing, otherwise it is determined to be due to improper cooling parameters.

[0182] Cooling parameter dynamic adjustment module: used to receive the intervention trigger signal and stress relief deviation output by the stress relief assessment module. Based on the results of the anomaly source determination, a PID control algorithm is used to calculate the cooling rate adjustment. And when the anomaly originates from uneven hot-press curing, a compensation coefficient is added. Output zone cooling control commands for each area.

[0183] Cross-process data recording and transmission module: Used to jointly record the hot pressing stage data and cooling stage data of each board using the board ID as an index, and transmit them to the drilling compensation calculation module through the MES system or industrial Ethernet.

[0184] Drilling compensation calculation module: used to receive data transmitted by the cross-process data recording and transmission module. , and hot pressing parameters, for and The sheet metal uses standard compensation values. mm; for or For the sheet metal parts, calculate the differential compensation amount. and regional compensation amount It also outputs drilling compensation commands.

[0185] Drilling accuracy feedback module: Used to obtain the actual hole position accuracy detection results after drilling is completed and to calculate the predicted deviation. ,when When the temperature exceeds 15%, a model calibration trigger signal is output. A hierarchical attribution strategy is used to feed this signal back to the hot-pressing parameter recording module to optimize hot-pressing temperature uniformity, and also to the stress release assessment module to optimize... .

[0186] like Figure 2 As shown, the data flow between modules is as follows: the hot pressing parameter recording module and the cooling stress monitoring module transmit data to the stress release degree assessment module; the stress release degree assessment module transmits the intervention signal and the abnormality source judgment result to the cooling parameter dynamic adjustment module, and at the same time transmits the stress release status data to the cross-process data recording and transmission module; the cross-process data recording and transmission module transmits the combined data to the drilling compensation calculation module; the drilling compensation calculation module outputs the compensation command to the drilling equipment, and at the same time transmits the drilling compensation data to the drilling accuracy feedback module; the drilling accuracy feedback module feeds back to the hot pressing parameter recording module and the stress release degree assessment module respectively, realizing a closed loop across processes throughout the entire process.

[0187] The method and system of the present invention will be further described below through specific application scenario embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0188] Example 1: Coordinated control of lamination, cooling, and drilling in a certain type of 24-layer high-frequency high-speed mixed-lamination PCB board

[0189] This embodiment uses a 24-layer high-frequency, high-speed hybrid PCB board for communication equipment, with a board size of 510mm × 450mm, a board thickness of 3.2mm, and a material system of FR-4 and PTFE hybrid pressing, as an example of the pressing, cooling, and drilling process with a Tg of approximately 175℃, to illustrate the specific implementation and effect verification of the method of the present invention in actual production. A total of 24 boards were pressed in this batch, arranged in a 4-row × 6-column layout in the press. The curing temperature during the hot pressing stage was set at 190℃, and the holding time was 120 minutes.

[0190] Step S1: Obtain the surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins.

[0191] S11: Before the end of the heat preservation stage during the hot pressing stage, i.e. at the 118th minute, the curing temperature distribution data of each board surface is obtained through an embedded thermocouple array.

[0192] Thermocouples are arranged at a density of one monitoring point per 100 cm². The area of ​​a 510 mm × 450 mm plate is approximately 229,500 mm², or 2295 cm². Dividing this by 100 cm² / point, a total of approximately 23 monitoring points are arranged. These monitoring points are arranged in a 5x5 grid, with 5 points per row at the edges, 9 points in a 3x3 grid in the center, and 1 point at each of the four corners, totaling approximately 23 points. The curing temperature data of each monitoring point is mapped to a two-dimensional coordinate system on the plate surface to generate a curing temperature distribution matrix. The grid is divided into , That is, the plate is divided into 20 units along the length direction and 18 units along the width direction, with a total of 360 grid nodes. The temperature value of each node is obtained by bilinear interpolation of the data from 23 monitoring points.

[0193] Taking plate number #08 located in the center area of ​​the press and plate number #21 located in the edge area of ​​the press as examples, their curing temperature data are shown in Table 1 below:

[0194] Table 1

[0195]

[0196] Temperature values ​​at 23 monitoring points for panel #08 ranged from 188.5℃ to 191.2℃, with a standard deviation of 0.8℃. Temperature values ​​at 23 monitoring points for panel #21 ranged from 183.2℃ to 189.8℃, with a standard deviation of 2.1℃. The temperature difference at the edge area reached 6.6℃, and the curing temperature was significantly lower than that in the center area by approximately 3-5℃, indicating that the resin curing and cross-linking density in the edge area may be insufficient. The surface temperature difference of the 24 panels in this batch ranged from 2.3℃ to 7.1℃, with an average of 4.1℃. Specifically, the distribution was as follows: the four central panels... Eight panels in the middle area 12 edge panels .

[0197] S12: Obtain PCB design information and pre-calculate theoretical cooling curves.

[0198] Board design information: The stack-up is 24 layers, with the inner layer being FR-4 material with a CTE of approximately 14-17 ppm / ℃, and the outer layer being PTFE high-frequency material with a CTE of approximately 20-25 ppm / ℃. The drilling design includes 158 holes, including 86 signal holes, 48 ​​grounding vias, 16 heat dissipation holes, and 8 positioning holes, with a hole diameter range of 0.2 mm to 1.0 mm.

[0199] Based on the differences in CTE and curing temperature distribution in each region, the upper limit of the theoretical cooling rate is pre-calculated. The calculation method is as follows: for each grid node, based on the curing temperature at that node... Material CTE and target stress release degree of 90%, based on stress relaxation kinetic equation Calculate the maximum allowable cooling rate.

[0200] Taking panel #21 as an example, the data of a certain monitoring point in the edge area is as follows: , ppm / ℃, total temperature drop during cooling , min⁻¹, the stress relaxation constant of glass fiber reinforced epoxy resin, substituted into :

[0201]

[0202] A monitoring point in the central area: , ppm / ℃ :

[0203]

[0204] The theoretical upper limit of the cooling rate in the edge region is approximately 1.0℃ / min, and in the center region, it is approximately 1.1℃ / min. Both are lower than the empirical threshold of 2.5℃ / min, indicating that this thick plate requires slower cooling to achieve sufficient stress release. The difference between the center and the edge is small, only about 4%, therefore, in actual control... Setting a uniform upper limit is reasonable; differentiated control should primarily be based on real-time monitoring. Dynamic adjustments are made.

[0205] S13: Before cooling begins, the initial warp distribution of the plate surface is obtained through a laser displacement sensor.

[0206] A laser displacement sensor is positioned 500mm directly above the board surface, with a measurement accuracy of ±5μm. During measurement, the laser beam scans the grid nodes point by point. Each node's measurement value is corrected for systematic errors caused by board tilt and sensor installation deviations before generating a height matrix. Initial warpage of board #08. mm, the maximum height within the grid is 0.28mm minus the minimum height is -0.04mm. Initial warpage of panel #21. mm, maximum height value 0.22mm minus minimum height value -0.06mm.

[0207] S14: Obtain initial temperature distribution data at multiple points on the plate surface through a thermocouple array.

[0208] At the instant cooling begins, defined as the end of hot pressing and the simultaneous activation of the cooling equipment, temperature values ​​at each monitoring point are read through an embedded thermocouple array to generate an initial temperature distribution matrix. The average initial temperature of both plates was approximately 188℃, and the standard deviation of the initial temperature values ​​at each monitoring point was ±1.5℃.

[0209] Step S2: Monitor warping and temperature in real time, calculate various indicators, and determine the source of anomalies.

[0210] The cooling phase is every 3 seconds. Temperature and warpage data from each monitoring point are continuously sampled at a frequency of once per minute. The data acquisition system is driven by a PLC controller, and during each acquisition, the temperature and warpage values ​​of all 23 monitoring points are read synchronously, with a timestamp synchronization accuracy of ±10ms.

[0211] S21-S22: Calculate the real-time cooling rate and warpage rate at each monitoring point.

[0212] by Taking the time 'min' as an example, this time is the 300th sampling, which is 15min ÷ 0.05min / sampling = 300 samplings. The temperature data at the 300th sampling is denoted as... The 299th sampling is The temperature data at min is denoted as .

[0213] Panel #08, central area, uniform curing: Temperature at each monitoring point during the 300th sampling. The temperature ranged from 156℃ to 161℃, a decrease of approximately 0.10℃ to 0.12℃ on average compared to the 299th sampling. Average cooling rate ℃ / min, calculate the average slope of 3 sampling points before and after that time. Warp data Distributed between 0.08mm and 0.22mm, warpage rate mm / min.

[0214] Panel #21, edge area, uneven curing: During the 300th sampling, a monitoring point on the edge was recorded as monitoring point i=7, located at the lower left corner of panel #21, with a temperature of... ,Pick The average cooling rate of the five sampling points before and after min was calculated. Warpage data After taking the moving average mm / min.

[0215] S23: Calculate the stress relief degree of each region.

[0216] Monitoring points of panel #08 ,in This represents the initial warp rate at the start of cooling, i.e., the first sampling. Pick to min, which is the warping change from the first to the second sampling divided by the sampling interval: Part #08 The average value is approximately 0.12 mm / min, therefore Distribution .

[0217] Panel #21 Monitoring Point 7: mm / min, mm / min, Average surface area That is, the arithmetic mean of the 23 monitoring points.

[0218] S24: Calculate the stress release dispersion.

[0219] Part #08: Each monitoring point The maximum value is 8%, and the average value is 5%.

[0220] Part #21: .

[0221] S25: Determine whether intervention is triggered and identify the source of the anomaly.

[0222] Triggering condition is and ,in , .

[0223] Part #08: In most areas, it has reached 90%. No intervention is triggered.

[0224] Panel #21 Monitoring Point 7: and This triggers intervention.

[0225] Anomaly source determination: First, determine... Further judgment ,in That is, the curing temperature recorded at monitoring point 7 in stage S11. This refers to the average curing temperature of panel #21, with a difference of 3.0℃. , According to the judgment rules, If the temperature is within the range of 3.0℃, it is determined to be due to uneven hot-pressing curing; otherwise, it is determined to be due to improper cooling parameters. Since the temperature is not greater than 3.0℃, the source of the anomaly is determined to be improper cooling parameters.

[0226] Of the 24 panels in this batch, five panels, numbered #19, #21, #22, #23, and #24, located at the edge of the press, triggered the intervention mechanism. The anomaly in four of these panels was determined to be due to uneven hot-pressing curing. One panel #21 monitoring point 7 was determined to have improper cooling parameters.

[0227] Step S3: Dynamically adjust cooling parameters to form a real-time closed loop until cooling is complete.

[0228] S31: The PID control algorithm is used to calculate the cooling rate adjustment.

[0229] For monitoring point 7 of panel #21:

[0230]

[0231] The proportional term of a PID controller: ℃ / min.

[0232] Integral term calculation: starting from cooling Up to now The cumulative deviation of min. Due to the sampling interval min, there are 300 sampling points from the first sampling to the 300th sampling. It decreases approximately linearly over time, from about 85% initially to 62%, therefore Approximate from linear growth to The average deviation is approximately 16.5%. The integral term is approximately... ,in The area under the deviation-time curve is approximately the area of ​​a triangle. %·min, which is 0.02475. Multiply by The integral term contributes approximately 0.0074℃ / min.

[0233] Differential term: This represents the rate of change of deviation. exist The rate of change near min is approximately The contribution of the differential term is ℃ / min.

[0234] Total PID output: ℃ / min.

[0235] In reality, the above calculation is a simplified approximation. During actual system execution, the PLC controller's built-in digital PID algorithm performs cyclic calculations with a period of 0.05 minutes, using complete data from all 300 sampling points in each calculation. After complete system calculation, .

[0236] S32: Determine the final adjustment amount based on the results of the anomaly source judgment.

[0237] The anomaly at monitoring point 7 was determined to be caused by improper cooling parameters. No compensation coefficient is added. The output is directly calculated using PID control.

[0238]

[0239] The cooling rate in this area decreased from 3.2 °C / min to approximately 1.56 °C / min.

[0240] For other monitoring points on panel #21, if the source of the anomaly is uneven hot-pressing curing, Then a compensation coefficient needs to be added. For example, monitoring point 8... , The difference is 4.0℃.

[0241]

[0242] If the PID output of the monitoring point is The final adjustment amount .

[0243] S33: Maintain or appropriately accelerate the cooling rate for areas where stress relief has reached the standard.

[0244] Part #08 and other areas Maintain the current cooling rate of 2.1℃ / min. However, if a certain area... If the cooling rate is below 2.0℃ / min, the cooling rate can be appropriately increased to 2.2℃ / min to improve efficiency, but it must not exceed 2.5℃ / min.

[0245] S34: Continuous monitoring forms a closed loop.

[0246] After the intervention is implemented, monitoring continues at a frequency of once every 3 seconds.

[0247] At min, the 400th sampling: monitoring point 7 , , mm / min.

[0248] At time min, the 600th sample: , .

[0249] At min time, the 800th sample: Once the target value is reached, the system confirms that the stress in the area has been fully released and exits the intervention. Afterwards, cooling continues according to the normal cooling curve.

[0250] When cooling is complete, The total cooling time was extended due to intervention, approximately 10 minutes longer than the 65 minutes for panel #08. (Monitoring points for panel #21) The range was 88%-94%, with an average of 91%, of which 7 were original abnormal monitoring points. .

[0251] Step S4: Jointly record the data of the hot pressing and cooling stages, and transfer them to the drilling process using the plate ID as an index.

[0252] S41: Record data for the hot pressing and cooling stages of each plate.

[0253] Part #08: Each monitoring point The range is 91%-95%, with an average of 93%.

[0254] Part #21: Each monitoring point The range was 88%-94%, with an average of 91%. The peripheral areas, namely monitoring points 7, 8, 9, 10, and 15... Approximately 88%-91%, in the central area, i.e., monitoring points 1-6. Approximately 92%-94%.

[0255] Intervention records during the cooling process: Part #21 triggered 2 interventions, with monitoring point 7 at... Triggered once per minute, adjustment rate 1.64℃ / min; monitoring point 8 at... Triggered once per minute, with an adjustment rate of 2.10℃ / min. The trigger time, abnormality source judgment result, and adjustment rate for each intervention were all recorded.

[0256] S42: Predict the residual stress distribution and expected expansion / contraction of each plate.

[0257] The prediction model uses multiple linear regression, and the input features are expressed as deviations from the baseline value:

[0258]

[0259] The regression coefficients were determined by least squares fitting of 500 sets of historical production data. μm, μm / ℃, μm / ℃, μm / %, μm / (ppm / ℃·℃). Goodness of fit .

[0260] A comparative calculation is performed using the central area of ​​panel #08 and the edge area of ​​panel #21 as examples:

[0261] Monitoring point 1 in the central area of ​​panel #08: , , , ppm / ℃ :

[0262] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm. μm.

[0263] Monitoring point 7 in the edge area of ​​panel #21: , , , ppm / ℃ :

[0264] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm. μm.

[0265] Calculation results show that the expected expansion and contraction in the central area of ​​panel #08 is approximately 10.2 μm, and the expected expansion and contraction in the edge area of ​​panel #21 is approximately 14.2 μm. Due to the lower curing temperature, greater temperature difference between the panel surfaces, less sufficient stress release, and higher CTE in the edge area, the expected expansion and contraction is approximately 39% higher.

[0266] Other edge panels in this batch, namely #19, #22, #23, and #24, were due to... and different, The distribution is slightly different, but it is basically on the same order of magnitude as plate #21.

[0267] S43: Transfer the data to the drilling process compensation calculation module.

[0268] S52: Calculate the borehole compensation amount.

[0269] Weighting coefficients: , , Standard compensation value mm.

[0270] Take the center area of ​​panel #08. , μm, :

[0271]

[0272]

[0273] Part #21, take , μm, :

[0274]

[0275] The differential compensation amount for panel #21, taking edge monitoring point 7 as an example:

[0276]

[0277]

[0278]

[0279] S53: Drilling accuracy inspection.

[0280] After drilling was completed, the actual positions of all 158 holes were detected using AOI. For each hole, the offset between its actual coordinates and the theoretical design coordinates was calculated, and the average and maximum offsets of each plate were statistically analyzed, as shown in Table 2 below.

[0281] Table 2

[0282]

[0283] Statistics for all 24 panels in the batch: a total of 158 × 24 = 3792 holes were inspected, with an average offset of 15.5 μm, a standard deviation of 5.6 μm, a CPK of 1.65, and a failure rate of 0%. All hole offsets were less than the 50 μm upper limit of the specification. The average offset of the four panels in the central area, including panel #08, was 12.8 μm, while the average offset of the twelve panels in the edge area, including panels #19-#24, was 17.2 μm. The differences mainly stemmed from uneven curing temperature distribution during the hot-pressing stage and differences in stress release during the cooling stage.

[0284] This embodiment takes a certain type of 24-layer high-frequency high-speed mixed-press PCB as an example to fully demonstrate the specific implementation of the method of the present invention from obtaining the temperature distribution of hot pressing and curing, real-time monitoring and intervention of the cooling process, to the differential compensation of drilling. It verifies the applicability and control accuracy of the present invention to thick plates and mixed-press material systems. By monitoring the stress release degree in real time and triggering dynamic intervention, the stress release degree in the edge area is increased from 62% to 91%, and the drilling CPK reaches 1.65.

[0285] Example 2: Validation of cross-batch effect across three consecutive batches

[0286] To further verify the cross-batch consistency and continuous optimization capabilities of this invention, for the same product model, identical to Example 1, a 24-layer high-frequency high-speed mixed-press PCB board with a board size of 510mm × 450mm and a board thickness of 3.2mm, was produced in three consecutive batches, with 24 boards per batch, totaling 72 boards. The method of this invention was used to control each batch. The production intervals for the three batches were 5 days and 7 days, respectively. The same set of pressing and drilling equipment was used, the same operators were employed, and the raw material suppliers and batches remained consistent to eliminate interference from variables related to equipment, personnel, and materials. To distinguish this from the detailed process of a single board in Example 1, specific board numbers are no longer used in this embodiment. Instead, "central area board" and "edge area board" refer to boards located at corresponding positions on the press, and the statistical average or typical value of these positions within the batch is used for explanation. The execution of each batch is described in detail below, batch by batch and step by step.

[0287] Batch B001

[0288] Step S1: Obtain the surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins.

[0289] S11: Before the end of the heat preservation stage during the hot pressing stage, i.e. at the 118th minute, the curing temperature distribution data of each board surface is obtained through an embedded thermocouple array.

[0290] The 24 panels in this batch were arranged in a 4-row x 6-column layout in the press. Thermocouples were arranged at a density of one monitoring point per 100cm², for a total of 23 monitoring points. The curing temperature was set at 190℃, and the actual holding time was 120 minutes.

[0291] Each panel The temperature ranged from 2.5℃ to 7.2℃, with a batch average of 4.2℃ and a standard deviation of 1.1℃. This included the four central panels. The temperature range is 2.5-3.5℃, with an average of 2.9℃; the middle area has 8 panels. The temperature range is 3.5-5.0℃, with an average of 4.3℃; 12 plates at the edge. The temperature range is 5.0-7.2℃, with an average of 6.1℃. This is the lowest curing temperature for the board surface. It is located at the edge corner of the press. Maximum curing temperature of the board surface. Located in the central area of ​​the press. Average curing temperature for the entire batch. .

[0292] S12: Obtain PCB design information and pre-calculate theoretical cooling curves.

[0293] The panel design information is the same as in Example 1: the stack is 24 layers, FR-4 and PTFE are mixed and pressed, and the drilling design includes 158 holes.

[0294] Based on the differences in CTE and curing temperature distribution in each region, the upper limit of the theoretical cooling rate is pre-calculated. Using typical data from marginal areas, , Substituting ppm / ℃ into the stress relaxation kinetic equation ,in min⁻¹, , :

[0295]

[0296] Typical data for the central area , ppm / ℃ :

[0297]

[0298] Both were below 2.5℃ / min. As a unified upper limit, differentiated control is based on real-time monitoring. Dynamic adjustments are made.

[0299] S13: Before cooling begins, the initial warp distribution of the plate surface is obtained through a laser displacement sensor.

[0300] The initial warpage of each panel ranged from 0.22 mm to 0.35 mm, with an average of 0.28 mm across the entire batch and a standard deviation of 0.04 mm. The average warpage was 0.25 mm in the center area and 0.32 mm in the edge area.

[0301] S14: Obtain initial temperature distribution data at multiple points on the plate surface through a thermocouple array.

[0302] At the moment cooling begins, the average initial temperature of each board surface is approximately 188℃, and the standard deviation of the initial temperature value at each monitoring point is ±1.5℃.

[0303] Step S2: Monitor warping and temperature in real time, calculate various indicators, and determine the source of anomalies.

[0304] S21-S22: Calculate the real-time cooling rate and warpage rate at each monitoring point.

[0305] The cooling phase lasts every 3 seconds, i.e. Min is one continuous sampling. Taking the min time as an example, the monitoring data for each board is as follows:

[0306] A monitoring point on a panel in the edge area: ℃ / min, mm / min.

[0307] A monitoring point on a certain panel in the central area: ℃ / min, mm / min.

[0308] S23: Calculate the stress relief degree of each region.

[0309] This monitoring point is located on the edge area of ​​the panel. mm / min, Average surface area .

[0310] This monitoring point is located on the central area panel: It is approximately 89%, close to the target value of 90%.

[0311] S24: Calculate the stress release dispersion.

[0312] This monitoring point is located on the edge area of ​​the panel. .

[0313] This monitoring point is located on the central area panel: Approximately 5%.

[0314] S25: Determine whether intervention is triggered and identify the source of the anomaly.

[0315] , .

[0316] This batch contains 6 trigger plates, all located at the edge of the press. Each trigger plate... The trigger value ranged from 16% to 22%, all of which were higher than the threshold of 15%. The temperature range is 2.5℃ to 4.2℃. This includes four sheet metal pieces. The source of the anomaly was determined to be uneven hot-pressing curing; 2 panels The source of the anomaly was determined to be improper cooling parameters.

[0317] Step S3: Dynamically adjust cooling parameters to form a real-time closed loop until cooling is complete.

[0318] S31-S32: The PID control algorithm is used to calculate the cooling rate adjustment amount, and the final adjustment amount is determined according to the source of the anomaly.

[0319] Taking one of the edge plates with uneven heat pressing and curing as an example:

[0320] PID calculation . , Final adjustment amount The cooling rate decreased from 3.5℃ / min to approximately 0.87℃ / min.

[0321] Take one of the edge plates with improper cooling parameters as an example:

[0322] PID calculation No superposition Direct execution reduced the cooling rate from 3.2℃ / min to approximately 1.56℃ / min.

[0323] The cooling parameters for the six intervention plates can be adjusted from 1.5℃ / min to 2.8℃ / min. Range 1.10-1.42.

[0324] S33-S34: Continuous monitoring forms a closed loop.

[0325] After intervention, the 6 panels The initial efficiency improved from 55%-68% to 88%-93%, an average increase of 29%. The intervention duration ranged from 25-45 minutes. Upon completion of cooling, each plate... The range is 88%-95%, with an average of 90.5% for the entire batch and a standard deviation of 2.1%.

[0326] Step S4: Jointly record the data of the hot pressing and cooling stages, and transfer them to the drilling process using the board ID as an index.

[0327] S41: Record data for the hot pressing and cooling stages of each plate.

[0328] 6 intervention panels The range was 88%-93%, with an average of 90.2%. The remaining 18 panels had issues in their center and middle areas. The range is 91%-95%, with an average of 92.8%.

[0329] S42: Predict the residual stress distribution and expected expansion / contraction of each plate.

[0330] The prediction model uses multiple linear regression, and the regression coefficients are the same as in Example 1: μm, μm / ℃, μm / ℃, μm / %, μm / (ppm / ℃·℃)).

[0331] Taking a certain plate in the edge area as an example, its typical value is , , , ppm / ℃ :

[0332] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0333]

[0334] Taking a certain panel in the central area as an example, its typical value is , , , ppm / ℃ :

[0335] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0336]

[0337] 6 intervention panels The range is 15.8-17.5μm, and the remaining plates... The range is 10.2-13.5 μm.

[0338] S43: Transfer the data to the drilling process compensation calculation module.

[0339] Step S5: Borehole differential compensation, detection accuracy and hierarchical attribution feedback.

[0340] S51-S52: Calculate the borehole compensation amount.

[0341] Weighting coefficients: , , Standard compensation value mm.

[0342] Taking a certain panel in the edge area as an example, , μm, :

[0343]

[0344]

[0345] Taking a certain panel in the central area as an example, , μm, :

[0346]

[0347]

[0348] The compensation amount for the 6 intervention plates ranges from 0.042 to 0.065 mm, with an average of 0.052 mm. The compensation amount for the remaining plates ranges from 0.022 to 0.038 mm, with an average of 0.028 mm.

[0349] S53: Drilling accuracy inspection and layered attribution feedback.

[0350] After drilling was completed, all 158 holes were inspected using AOI. The drilling accuracy statistics for each plate position are shown in Table 3 below:

[0351] Table 3

[0352]

[0353] Note: Edge areas are divided into two categories: "non-triggered intervention" and "triggered intervention." "Triggered intervention" refers to the cooling process. and The system performed dynamic adjustment of cooling parameters; "No intervention triggered" means that although it is located in the edge area, the intervention triggering conditions are not met, and cooling is completed according to the normal cooling curve.

[0354] Statistics of all 24 plates in the batch: average offset 16.8μm, standard deviation 6.2μm, CPK=1.58, failure rate 0.8%, two hole offsets >50μm, located on two edge plates respectively.

[0355] Predicted deviation of press edge plate The average value is 13.8%, which is close to the 15% threshold, and the system determines that it needs to be fine-tuned.

[0356] Hierarchical attribution feedback:

[0357] Feedback from the hot pressing process: Plates in the edge areas generally... The temperature is too high, which is determined to be a problem with temperature uniformity during the hot pressing stage. The average temperature of 4.2℃ is slightly high. Specific feedback measures: Adjust the heating power distribution of the press, increase the power in the edge areas by 5%, and extend the heat preservation time from 120 minutes to 125 minutes.

[0358] Cooling process feedback: weighting coefficients The value was adjusted from 0.30 to 0.32, meaning the weight of the stress release deviation on the compensation amount was increased. Keep it at 90%.

[0359] Batch B002

[0360] Step S1: Obtain the surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins.

[0361] S11: At the 123rd minute before the end of the heat preservation stage during hot pressing, the curing temperature distribution data of each board surface is obtained through an embedded thermocouple array.

[0362] Based on feedback from B001, the following adjustments were made to the hot pressing process: the heating power distribution was adjusted, the power in the edge area was increased by 5%, the holding time was extended from 120 minutes to 125 minutes, and the other parameters remained unchanged.

[0363] The 24 panels in this batch were arranged in a 4-row x 6-column layout in the press, with the curing temperature set at 190℃ and the actual holding time at 125 minutes.

[0364] Each panel The temperature range is 2.1℃ to 6.2℃, with an average of 3.8℃ for the entire batch, a decrease of 0.4℃ compared to B001, a reduction of 9.5%, and a standard deviation of 0.9℃. This includes the four central panels. The temperature range is 2.1-3.2℃, with an average of 2.6℃; the middle area has 8 panels. The temperature range is 3.2-4.5℃, with an average of 3.9℃; 12 plates at the edge. The temperature range is 4.5-6.2℃, with an average of 5.5℃. This is the minimum curing temperature for the board surface. The maximum curing temperature is 1.7℃ higher than B001. Average curing temperature of the entire batch .

[0365] S12: Obtain PCB design information and pre-calculate theoretical cooling curves.

[0366] Based on the differences in CTE and curing temperature distribution in each region, the upper limit of the theoretical cooling rate was recalculated. Typical data from the edge region were used. , Substitute ppm / ℃ into the equation:

[0367]

[0368] Typical data for the central area , ppm / ℃ :

[0369]

[0370] The two are not very different. It will remain a uniform upper limit.

[0371] S13: Before cooling begins, the initial warp distribution of the plate surface is obtained through a laser displacement sensor.

[0372] The initial warpage of each panel ranged from 0.20mm to 0.32mm, with an average of 0.26mm for the entire batch, a decrease of 0.02mm compared to B001, representing a reduction of 7.1%, with a standard deviation of 0.035mm. The average warpage for panels in the central area was 0.23mm, and for panels in the edge areas it was 0.30mm.

[0373] S14: Obtain initial temperature distribution data at multiple points on the plate surface through a thermocouple array.

[0374] At the moment cooling begins, the average initial temperature of each board surface is approximately 188℃, and the standard deviation of the initial temperature value at each monitoring point is ±1.5℃.

[0375] Step S2: Monitor warping and temperature in real time, calculate various indicators, and determine the source of anomalies.

[0376] S21-S22: Calculate the real-time cooling rate and warpage rate at each monitoring point.

[0377] During the cooling phase, samples were continuously taken every 3 seconds. Taking min time as an example:

[0378] Taking a monitoring point of a certain plate in the edge area as an example, The temperature / min is lower than that of B001 (3.3℃ / min). mm / min.

[0379] A monitoring point on a certain panel in the central area: ℃ / min, mm / min.

[0380] S23: Calculate the stress relief degree of each region.

[0381] This monitoring point is located on the edge area of ​​the panel. mm / min, Average surface area .

[0382] This monitoring point is located on the central area panel: It is approximately 91%, which has reached the target value of 90%.

[0383] S24: Calculate the stress release dispersion.

[0384] This monitoring point is located on the edge area of ​​the panel. .

[0385] This monitoring point is located on the central area panel: Approximately 4%.

[0386] S25: Determine whether intervention is triggered and identify the source of the anomaly.

[0387] , .

[0388] This batch has 4 panels that triggered intervention, accounting for 16.7% of the total batch, a decrease of 33.3% compared to B001. All of them are panels located at the edge of the press. The trigger value range is 15%-19%, which is narrower than the 16%-22% range of B001. This includes three boards. The temperature is 3.2℃-3.5℃, that is... >3℃, the abnormality was determined to be caused by uneven hot-pressing curing; 1 piece of board It is 2.0℃, that is If the temperature is ≤3℃, the source of the anomaly is determined to be improper cooling parameters.

[0389] Step S3: Dynamically adjust cooling parameters to form a real-time closed loop until cooling is complete.

[0390] S31-S32: The PID control algorithm is used to calculate the cooling rate adjustment amount, and the final adjustment amount is determined according to the source of the anomaly.

[0391] Taking one of the edge plates with uneven heat pressing and curing as an example:

[0392] PID calculation . , Final adjustment amount The cooling rate decreased from 3.1℃ / min to approximately 0.87℃ / min.

[0393] Take one of the edge plates with improper cooling parameters as an example:

[0394] PID calculation No superposition Direct execution reduced the cooling rate from 2.9°C / min to approximately 1.42°C / min.

[0395] The cooling parameter adjustment range for the four intervention plates is 1.2-2.3℃ / min, which is lower than the 1.5-2.8℃ / min range for B001. Range 1.10-1.35.

[0396] S33-S34: Continuous monitoring forms a closed loop.

[0397] After intervention, the four panels The initial efficiency improved from 60%-70% to 89%-94%, an average increase of 27%. The intervention duration ranged from 20-35 minutes, shorter than the 25-45 minutes for B001. Upon completion of cooling, each plate... The range was 90%-96%, with an average of 91.8% for the entire batch, an increase of 1.3 percentage points compared to B001, and a standard deviation of 1.8%, which was narrower than the 2.1% of B001.

[0398] Step S4: Jointly record the data of the hot pressing and cooling stages, and transfer them to the drilling process using the plate ID as an index.

[0399] S41: Record data for the hot pressing and cooling stages of each plate.

[0400] 4 intervention panels The range is 90%-94%, with an average of 92.0%. The remaining 20 panels... The range is 92%-96%, with an average of 93.5%.

[0401] S42: Predict the residual stress distribution and expected expansion / contraction of each plate.

[0402] The prediction model uses multiple linear regression with constant regression coefficients.

[0403] Taking a certain plate in the edge area as an example, its typical value is , , , ppm / ℃ :

[0404] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0405]

[0406] Taking a certain panel in the central area as an example, its typical value is , , , ppm / ℃ :

[0407] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0408]

[0409] 4 intervention panels The range is 14.5-16.8μm, and the remaining plates... The range is 10.0-12.8 μm.

[0410] S43: Transfer the data to the drilling process compensation calculation module.

[0411] Step S5: Borehole differential compensation, detection accuracy and hierarchical attribution feedback.

[0412] S51-S52: Calculate the borehole compensation amount.

[0413] The weighting coefficients remain unchanged: , , Standard compensation value mm.

[0414] Taking a certain panel in the edge area as an example, , μm, :

[0415]

[0416]

[0417] Taking a certain panel in the central area as an example, , μm, :

[0418]

[0419]

[0420] The compensation range for the four intervention plates is 0.035-0.055mm, with an average of 0.044mm. The compensation range for the remaining plates is 0.020-0.035mm, with an average of 0.026mm.

[0421] S53: Drilling accuracy inspection and layered attribution feedback.

[0422] After drilling was completed, all 158 holes were inspected using AOI. The drilling accuracy statistics for each plate position are shown in Table 4 below:

[0423] Table 4

[0424]

[0425] Note: Edge areas are divided into two categories: "non-triggered intervention" and "triggered intervention." "Triggered intervention" refers to the cooling process. and The system performed dynamic adjustment of cooling parameters; "No intervention triggered" means that although it is located in the edge area, the intervention triggering conditions are not met, and cooling is completed according to the normal cooling curve.

[0426] Full batch statistics: average offset 15.2μm, a decrease of 1.6μm compared to B001; standard deviation 5.5μm, narrowing compared to 6.2μm of B001; CPK=1.72, an increase of 0.14 compared to B001; failure rate 0.4%; one hole offset >50μm, located on an edge plate.

[0427] Predicted deviation of sheet metal at the edge of press The mean is 10.5%, which is below the 15% threshold, so there is no need to trigger model calibration.

[0428] Hierarchical attribution feedback:

[0429] Feedback on the hot pressing process: Hot pressing uniformity continues to improve, especially in edge areas. The temperature has been reduced from 4.2℃ to 3.8℃. The heating power distribution has been further optimized, and the heat preservation time has been extended to 128 minutes.

[0430] Cooling process feedback: No adjustments required.

[0431] Batch B003

[0432] Step S1: Obtain the surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins.

[0433] S11: Before the end of the heat preservation stage during the hot pressing stage, i.e. at the 126th minute, acquire the curing temperature distribution data of each board surface through an embedded thermocouple array.

[0434] Based on feedback from B002, the heating power distribution in the hot pressing process has been further optimized. The power in the edge area has been increased by 3% compared to B002, and the heat preservation time has been extended to 128 minutes, while the other parameters remain unchanged.

[0435] The 24 panels in this batch were arranged in a 4-row x 6-column layout in the press, with the curing temperature set at 190℃ and the actual holding time at 128 minutes.

[0436] Each panel The temperature range is 1.8℃ to 5.5℃, with an average of 3.2℃ for the entire batch. This is a further decrease of 0.6℃ compared to B002, and a cumulative decrease of 1.0℃ compared to B001, with a standard deviation of 0.7℃. This includes the four central panels. The temperature range is 1.8-2.8℃, with an average of 2.3℃; the middle area has 8 panels. The temperature range is 2.8-4.0℃, with an average of 3.4℃; 12 plates at the edge. The temperature range is 4.0-5.5℃, with an average of 4.8℃. This is the minimum curing temperature for the board surface. The temperature is 1.3℃ higher than B002 and a cumulative increase of 3.0℃ compared to B001, with the highest curing temperature... Average curing temperature of the entire batch .

[0437] S12: Obtain PCB design information and pre-calculate theoretical cooling curves.

[0438] Using typical data from marginal areas, , Substitute ppm / ℃ into the equation:

[0439]

[0440] Typical data for the central area , ppm / ℃ :

[0441]

[0442] Both are still below 2.5℃ / min.

[0443] S13: Before cooling begins, the initial warp distribution of the plate surface is obtained through a laser displacement sensor.

[0444] The initial warpage of each panel ranged from 0.18mm to 0.28mm, with an average of 0.23mm across the entire batch. This represents a cumulative decrease of 0.05mm compared to B001, a reduction of 17.9%, with a standard deviation of 0.03mm. The average warpage for panels in the central area was 0.20mm, while that for panels in the edge areas was 0.26mm.

[0445] S14: Obtain initial temperature distribution data at multiple points on the plate surface through a thermocouple array.

[0446] At the moment cooling begins, the average initial temperature of each board surface is approximately 188℃, and the standard deviation of the initial temperature value at each monitoring point is ±1.5℃.

[0447] Step S2: Monitor warping and temperature in real time, calculate various indicators, and determine the source of anomalies.

[0448] S21-S22: Calculate the real-time cooling rate and warpage rate at each monitoring point.

[0449] Taking a monitoring point of a certain plate in the edge area as an example, The temperature / min is ℃, which is a further decrease compared to 2.9℃ / min for B002. mm / min.

[0450] A monitoring point on a certain panel in the central area: ℃ / min, mm / min.

[0451] S23: Calculate the stress relief degree of each region.

[0452] This monitoring point is located on the edge area of ​​the panel. mm / min, Average surface area .

[0453] This monitoring point is located on the central area panel: It is approximately 92%, exceeding the target value of 90%.

[0454] S24: Calculate the stress release dispersion.

[0455] This monitoring point is located on the edge area of ​​the panel. .

[0456] This monitoring point is located on the central area panel: Approximately 3%.

[0457] S25: Determine whether intervention is triggered and identify the source of the anomaly.

[0458] Only two boards in this batch triggered the intervention, accounting for 8.3% of the entire batch, a decrease of 66.7% compared to B001. Both were extreme cases at the edge corners of the press. The trigger values ​​were 16% and 18%, respectively, which just exceeded the threshold of 15%. The temperatures were 2.8℃ and 3.2℃ respectively. The source of the anomalies in both cases was determined to be uneven hot-pressing curing. This is because 3.2℃ > 3℃, while 2.8℃ ≤ 3℃. However, the temperature of this board... Furthermore, considering its location in an extreme corner, the system still determines that uneven hot-press curing is the dominant factor, while improper cooling parameters are a secondary factor. Therefore, it treats all cases as uneven hot-press curing and superimposes the results. compensate.

[0459] Step S3: Dynamically adjust cooling parameters to form a real-time closed loop until cooling is complete.

[0460] S31-S32: The PID control algorithm is used to calculate the cooling rate adjustment amount, and the final adjustment amount is determined according to the source of the anomaly.

[0461] Taking one of the corner pieces as an example:

[0462] PID calculation . , Final adjustment amount The cooling rate decreased from 2.8℃ / min to approximately 0.93℃ / min.

[0463] Another corner piece:

[0464] , , , Final adjustment amount The cooling rate decreased from 2.6℃ / min to approximately 0.94℃ / min.

[0465] The cooling parameters for the two plates are adjusted to 1.0℃ / min and 1.8℃ / min, respectively. They are 1.28 and 1.32 respectively.

[0466] S33-S34: Continuous monitoring forms a closed loop.

[0467] After intervention, the two plates The rates increased from the initial 65% and 68% to 92% and 93%, respectively. The intervention durations were 22 minutes and 28 minutes, respectively.

[0468] When cooling is complete, each plate The range was 91%-97%, with an average of 93.2% for the entire batch, an increase of 2.7 percentage points compared to B001, and a standard deviation of 1.5%, which was significantly narrower than B001's 2.1%.

[0469] Step S4: Jointly record the data of the hot pressing and cooling stages, and transfer them to the drilling process using the board ID as an index.

[0470] S41: Record data for the hot pressing and cooling stages of each plate.

[0471] 2 intervention plates The percentages are 92% and 93% respectively. The remaining 22 panels... The range is 92%-97%, with an average of 94.0%.

[0472] S42: Predict the residual stress distribution and expected expansion / contraction of each plate.

[0473] Taking a certain plate in the edge area as an example, its typical value is , , , ppm / ℃ :

[0474] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0475]

[0476] Taking a certain panel in the central area as an example, its typical value is , , , ppm / ℃ :

[0477] Temperature deviation contribution: μm; Contribution of plate surface temperature difference: μm; Stress relief deviation contribution: μm; Material CTE × temperature drop contribution: μm.

[0478]

[0479] 2 intervention panels The thicknesses are 13.5 μm and 14.8 μm respectively, and the remaining plates... The range is 9.5-12.0 μm.

[0480] S43: Transfer the data to the drilling process compensation calculation module.

[0481] Step S5: Borehole differential compensation, detection accuracy and hierarchical attribution feedback.

[0482] S51-S52: Calculate the borehole compensation amount.

[0483] Taking a certain panel in the edge area as an example, , μm, :

[0484]

[0485]

[0486] Taking a certain panel in the central area as an example, , μm, :

[0487]

[0488]

[0489] The compensation amounts for the two intervention plates are 0.038 mm and 0.048 mm, respectively, while the compensation amounts for the remaining plates range from 0.018 to 0.032 mm, with an average of 0.024 mm.

[0490] S53: Drilling accuracy inspection and layered attribution feedback.

[0491] After drilling was completed, all 158 holes were inspected using AOI. The drilling accuracy statistics for each plate position are shown in Table 5 below:

[0492] Table 5

[0493]

[0494] Note: Edge areas are divided into two categories: "non-triggered intervention" and "triggered intervention." "Triggered intervention" refers to the cooling process. and The system performed dynamic adjustment of cooling parameters; "No intervention triggered" means that although it is located in the edge area, the intervention triggering conditions are not met, and cooling is completed according to the normal cooling curve.

[0495] Full batch statistics: average offset 14.1μm, a decrease of 2.7μm compared to B001, standard deviation 5.0μm, significantly narrower than 6.2μm of B001, CPK=1.85, an increase of 0.27 compared to B001, failure rate 0%, and offset of all 3792 holes less than the upper limit of 50μm specification.

[0496] Full batch prediction deviation The mean is 8.2%, which is below the 15% threshold, so there is no need to trigger model calibration.

[0497] A comprehensive comparison of the three batches is shown in Table 6 below:

[0498] Table 6

[0499]

[0500] This embodiment demonstrates the system self-evolution capability of the present invention's end-to-end dual-loop closed-loop mechanism through cross-batch effect verification across three consecutive batches. As the number of production batches increases, the hot pressing process continuously optimizes temperature uniformity through hierarchical attribution feedback. The temperature dropped from 4.2℃ to 3.2℃, the need for intervention in the cooling process continued to decrease, the intervention rate of the boards decreased from 25% to 8.3%, the control parameters of each process were continuously optimized, and CPK increased from 1.58 to 1.85, which verified the continuous evolutionary characteristic of the system that "becomes more accurate with use".

[0501] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the above embodiments, or equivalent substitutions can be made to some or all of the technical features, and such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention. The scope of protection claimed by the present invention is determined by the claims.

Claims

1. A method for cross-process stress coordination control in PCB hot pressing and cooling drilling, characterized in that, Includes the following steps: S1: Obtain the plate surface curing temperature distribution data during the hot pressing stage and the baseline state data before cooling begins; S2: During the cooling process, monitor the warpage and temperature changes of each area of ​​the board in real time, calculate the cooling rate, warpage rate, stress release degree and stress release degree dispersion of each area, and determine the source of stress anomalies based on the correlation between stress release degree, cooling rate and hot pressing curing temperature distribution. S3: Based on the stress release monitoring results and the judgment results of the abnormality source, dynamically adjust the cooling parameters to form a real-time closed-loop control until cooling is completed; S4: Using the plate ID as an index, jointly record the hot pressing stage data and cooling stage data of each plate, and transfer the above data to the drilling process; S5: The drilling process receives data from the hot pressing stage and the cooling stage. Based on the stress release state and hot pressing curing state of each plate, it generates plate-level and region-level differentiated drilling compensation parameters and performs drilling compensation. After drilling is completed, the hole position accuracy is checked, and the accuracy deviation is attributed to different layers and fed back to the hot pressing process and the cooling process respectively, forming a closed loop across the entire process.

2. The method according to claim 1, characterized in that, Step S1 specifically includes: S11: Before the end of the heat preservation stage during hot pressing, acquire curing temperature distribution data for each area of ​​the PCB board using a multi-point thermocouple array, and generate a curing temperature distribution matrix. And calculate the temperature difference on the plate surface. ,in This is the highest curing temperature of the board surface. This is the lowest curing temperature for the board surface; simultaneously, the position information of each board component in the press is recorded. S12: Obtain PCB design information, including stack-up parameters and drilling design data, and based on the differences in the thermal expansion coefficients of materials in different regions of the stack-up parameters, combine this with the curing temperature distribution recorded in S11. The theoretical cooling curves required for each region to cool from the curing temperature to room temperature were pre-calculated, and the board surface was divided into sections. For the grid, a lower upper limit for the cooling rate is set for areas with lower curing temperatures; S13: Before hot pressing is completed and cooling begins, the initial warpage distribution on the board surface is obtained using a laser displacement sensor or a structured light 3D scanner, and an initial warpage distribution matrix is ​​generated. ; S14: Acquire initial temperature distribution data at multiple points on the plate surface using a thermocouple array or infrared thermal imager, and generate an initial temperature distribution matrix. .

3. The method according to claim 1, characterized in that, Step S2 specifically includes: S21: Continuously sample the temperature data of each monitoring point at a set frequency and calculate the real-time cooling rate of each monitoring point: ， in For the first Each monitoring point is Cooling rate at any time The current temperature. The sampling interval; S22: Continuously sample warpage data at each monitoring point using the same frequency, and calculate the warpage rate of change at each monitoring point: ， in For the first Each monitoring point is The rate of change of warping at any given moment. This represents the current warp amount; S23: Based on warpage data, calculate the stress release degree of each region: ， in For the first Each monitoring point is Stress relief at any given time The initial rate of warpage at the start of cooling; S24: Calculate the dispersion of stress release degree at each monitoring point from the average stress release degree on the plate surface: ， in for The average stress relief value at all monitoring points on the plate at any given time: ， in This represents the total number of monitoring points. S25: The system determines whether to trigger intervention when the stress release level in a certain area is lower than the preset target value. And the cooling rate is higher than the preset threshold. When this occurs, the intervention mechanism is triggered: ， After intervention is triggered, the source of the anomaly is determined by a combined attribution of hot pressing and cooling: ， in The discreteness threshold, This is the temperature difference threshold.

4. The method according to claim 1, characterized in that, Step S3 specifically includes: S31: When intervention is triggered, the system generates a cooling rate adjustment command, and the adjustment amount is calculated using a PID control algorithm. ， in For stress relief deviation, , , For control parameters; S32: Determine the final cooling rate adjustment based on the anomaly source determination result: If the anomaly source is determined to be uneven hot-pressing curing, then add a compensation coefficient to the adjustment amount calculated by the PID controller. This further reduces the cooling rate of the area; The calculation method is as follows ,in This refers to the curing temperature during the hot pressing stage in this area. This is the average curing temperature of the board surface; if the source of the anomaly is determined to be improper cooling parameters, the adjustment amount calculated and output by the PID will be used directly. S33: For areas where stress relief has reached the target value, maintain or appropriately accelerate the cooling rate; S34: During the cooling process, continuously monitor the changes in stress release to form a real-time closed loop of "monitoring, evaluating, adjusting, and re-monitoring" until cooling is complete and the stress release in each area reaches the target value.

5. The method according to claim 1, characterized in that, Step S4 specifically includes: S41: Using the panel ID as an index, record the hot pressing stage data and cooling stage data for each panel, including the panel surface curing temperature distribution matrix. Temperature difference on the board surface Stress relief degree-time curve Final stress release degree in each region And records of interventions during the cooling process; S42: Based on the combined data of hot-pressing curing temperature distribution and cooling warping evolution, predict the residual stress distribution and expected expansion / contraction of each plate after cooling: ， in For the first Expected increase or decrease in volume for each region This refers to the curing temperature during the hot pressing stage in this area. For the overall temperature difference of the board surface, For the final stress relief degree, The coefficient of thermal expansion is To cool the total temperature drop; S43: Transfer the above data to the compensation calculation module of the drilling process using the board ID as the index.

6. The method according to claim 1, characterized in that, Step S5 specifically includes: S51: According to each board Calculate drilling compensation based on hot pressing parameters: for plates with sufficient stress release. And the hot-press curing is uniform Standard borehole compensation value is adopted. For plates with insufficient stress relief Or uneven curing under heat pressing For the sheet metal, calculate the drilling compensation: ， in For the first Drilling compensation for plate components. This is the standard borehole compensation value. , , These are the weighting coefficients; S52: The compensation amount adopts a differentiated distribution strategy, with different areas of the board distributed according to their respective... and hot pressing curing temperature Generate independent compensation amount : ， in For the first Block parts in coordinates The amount of borehole compensation at that location. This represents the final stress relief at that coordinate. This represents the expected expansion or contraction at that coordinate. This represents the hot-pressing curing temperature at that coordinate. The average curing temperature of the board surface is determined; a compensation distribution matrix is ​​generated and imported into the CNC system of the drilling equipment to correct the drilling coordinates in real time. S53: After drilling is completed, check the actual hole position accuracy, compare the actual offset with the predicted offset, and calculate the prediction deviation: ， when When the preset threshold is exceeded, model calibration is automatically triggered. The calibration adopts a hierarchical attribution strategy: if the predicted deviation of the plates at the edge of the press in the same batch is generally large, it is determined to be a temperature uniformity problem in the hot pressing stage, and feedback is given to the hot pressing process to optimize temperature uniformity; if the plates at the center of the press also have deviation, it is determined to be a problem with cooling parameters or model weights, and feedback is given to the cooling process for adjustment. And cooling adjustment strategies; by hierarchical attribution, the drilling accuracy test results are fed back to the hot pressing process and the cooling process respectively, forming a closed-loop optimization across the entire process.

7. A PCB hot-pressing cooling drilling cross-process stress coordination control system, used to implement the method as described in any one of claims 1-6, characterized in that, Includes the following modules: The hot pressing parameter recording module is used to acquire the curing temperature distribution data of each area of ​​the PCB board before the end of the heat preservation stage, calculate the temperature difference of the board surface, and record the position information of each board in the press. The cooling stress monitoring module is used to acquire warpage data and temperature data of each area of ​​the PCB board in real time during the cooling process, and continuously sample at a set frequency to output the real-time warpage change rate and real-time cooling rate of each monitoring point. The stress release assessment module is used to receive the warpage rate and cooling rate output by the cooling stress monitoring module, as well as the curing temperature distribution data transmitted by the hot pressing parameter recording module, and to calculate the stress release degree and stress release degree dispersion at each monitoring point. When the stress release degree is lower than the target value and the cooling rate is higher than the threshold, an intervention trigger signal is output. Based on the comparison results of the dispersion and the dispersion threshold, and the comparison results of the curing temperature deviation and the temperature difference threshold, the source of the abnormality is determined to be uneven hot-press curing or improper cooling parameters. The cooling parameter dynamic adjustment module is used to receive the intervention trigger signal, stress release deviation and abnormality source judgment result output by the stress release degree assessment module, calculate the cooling rate adjustment amount using PID control algorithm, and superimpose a compensation coefficient when the abnormality originates from uneven hot pressing curing, and output the zoned cooling control command for each area. The cross-process data recording and transmission module is used to jointly record the hot pressing stage data and cooling stage data of each plate using the plate ID as an index, and transmit the above data to the drilling compensation calculation module. The drilling compensation calculation module is used to receive the final stress release degree, expected expansion and contraction amount and hot pressing parameters transmitted by the cross-process data recording and transmission module, calculate the differentiated drilling compensation amount for each plate and each area of ​​the plate surface, and output the drilling compensation command. The drilling accuracy feedback module is used to obtain the actual hole position accuracy detection results after drilling is completed, calculate the prediction deviation, and output a model calibration trigger signal when the prediction deviation exceeds a preset threshold. The signal is fed back to the hot-pressing parameter recording module and the stress release evaluation module respectively using a hierarchical attribution strategy.

Images