Diode packaging quality analysis and detection method

Abstract

The invention discloses a diode packaging quality analysis and detection method, and relates to the technical field of semiconductor detection, and the method comprises the steps: building a data set; collecting production raw materials, process steps, process parameters and product performance parameters of different types of diodes; processing the data set; establishing a relational expression among the production raw materials, the process steps, the process parameters and the product performance parameters; establishing and training a model; selecting features with large correlation degree with the performance parameters to perform cross training on the model until the performance of the model meets a preset requirement; verifying the model; generating a prediction performance parameter; comparing the obtained actual performance parameters with the predicted performance parameters to generate an optimization scheme; and optimizing the production process by using an optimization scheme until the production requirements are met. The identification accuracy of the model is determined according to the detection and verification result, the large model is utilized to reduce the cost of subsequent detection and inspection, and meanwhile, the processing technology can be optimized to improve the final product quality.

Description

Technical Field

[0001] The invention relates to the technical field of semiconductor detection, in particular to a diode packaging quality analysis and detection method. Background Art

[0002] A diode is a semiconductor device with unidirectional conductivity and is widely used in electronic circuits. Its basic function is to allow current to flow in one direction and prevent it from flowing in the opposite direction. Since the functions of diodes in different circuits are different, their packaging forms and production processes are also different; the quality of diode packaging has an important impact on the service life, working performance and reliability of the diode. The common method of detecting the packaging quality of diodes on the market is to detect the defects of diode packaging through machine vision. Visual inspection can only locate specific defects and cannot predict the causes of defects; there are many reasons for a specific defect, and the superposition of different factors may produce the same defect; for example, under transparent packaging conditions, bubbles are generated at the packaging; it is generally believed that the temperature and the purity of the packaging material and the atmosphere mixed in the packaging process will cause bubbles to appear, and in the existing technology, it can only rely on manual experience to optimize the production process one by one to improve the defects; therefore, how to systematically analyze and detect the quality of diode packaging is conducive to determining the problems of defects in order to improve the yield rate of products in the production process. Summary of the invention

[0003] The present invention is proposed in view of the above-mentioned background problems.

[0004] Therefore, the problem to be solved by the present invention is how to determine the defective process steps according to the diode packaging quality, so as to optimize the processing method and improve the product yield.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: A first aspect of the present invention provides a diode packaging quality analysis and detection method, comprising the following specific steps: S1, establish a data set; collect production materials, process steps, process parameters and product performance parameters of different types of diodes; In step S1, different data sets are divided according to different types of diodes; Including plug-in package A and SMD package B; Among them, plug-in package A is divided into terminal type A1 and direct plug-in type A2 according to different types; According to the packaging material, it is divided into glass packaging and metal packaging; the terminal type plug-in diode of glass packaging is A11, and the terminal type plug-in diode of metal packaging is A12; The glass-encapsulated through-hole plug-in diode is A21, and the metal-encapsulated through-hole plug-in diode is A22; SMD package B is divided into SMA type B1, SMB type B2, SMC type B3 and SOD type B4.

[0006] In step S1, production raw materials, process steps, and process parameter data are collected for each data set established; Production materials include semiconductor materials, packaging materials and welding materials; The process steps include semiconductor material preparation steps and packaging steps; Process parameter data include the type, purity, and physical properties of production raw materials; process steps include operation time, temperature, pressure, and environmental conditions.

[0007] Specifically, semiconductor materials are divided into many types according to the raw materials and dopants, such as silicon, germanium, and gallium nitride; dopants: phosphorus, boron, factors affecting performance parameters also include the purity of semiconductor materials and doping concentration; in addition, it can be further subdivided according to supplier categories; common packaging materials are DPAK and SOT-23, for example, the thermal conductivity and dielectric constant of packaging materials will also affect performance parameters; welding materials and auxiliary materials have a greater impact on the performance parameters of diodes, and the main influencing factors include the type of solder, component ratio, and physical properties such as thermal conductivity and dielectric constant; The steps in the packaging process, including cutting, welding, molding, and the type and model of packaging equipment, will affect the performance of the final product.

[0008] Various parameter data in the process steps, such as temperature parameters, time parameters, pressure parameters, current and voltage parameters, temperature and atmosphere conditions, will have an impact on the product; therefore, detailed analysis is required after the above data is obtained.

[0009] S2, processing the data set; establishing the relationship between production raw materials, process steps, process parameters and product performance parameters; and sorting from low to high using correlation as the screening condition; In step S2, the public database, the internal database of the enterprise or the experience data are manually annotated. The data elements in the production raw materials are all connected with the performance parameters to establish a first-order relationship function, which is written into the control conditions of the model; Compare each data element in the process steps and process parameters with the standard production steps and parameters to determine the specific indicator type and gain nature of the data element affecting the performance parameters, whether it is a positive gain or a negative gain; Correlation analysis is used to characterize the correlation between different parameters in step S2 for the same performance indicator: in, r is the correlation value; X i , Y i are the observed values ​​of the two parameters, , are the means of the two parameters respectively; Randomly combine any two parameters, determine the correlation between them, and establish a correlation database.

[0010] S3, model establishment and training; establish a decision tree model, use production materials, process steps, and process parameters as model input; product performance parameters as model output; select features with a high correlation with performance parameters to cross-train the model until the performance of the model meets the preset requirements; Information gain is used as the criterion for feature selection in step S3: IG(M,a)=H(M)-H(M|a) Among them, IG(M,a) is the information gain of feature a in data set M; H(M) is the entropy of data set M, and H(M|a) is the entropy of data set M under the condition of feature a.

[0011] S4, verification of the model; inputting the production materials, process steps and process parameter data of the diode package into the input end of the model; the model generates predicted performance parameters; performing performance parameter tests on diodes produced in the same batch, comparing the actual performance parameters obtained with the predicted performance parameters, and generating an optimization plan; In step S4, different test equipment is used to perform multiple tests on the same indicator to reduce the error generated during the test process; in step S4, the deviation between the predicted performance parameter and the actual performance parameter is calculated, and the performance parameter with the largest deviation is taken, and the input parameter with the largest correlation with the performance parameter is calculated to generate an optimization plan; the performance parameters of common diodes include current-voltage characteristics: forward current (IF), reverse current (IR), forward voltage (VF); thermal performance: maximum operating temperature, thermal resistance; reliability parameters: failure rate, environmental adaptability (high temperature resistance, moisture resistance).

[0012] S5, optimizing the production process using the optimization solution of step S4 until the production requirements are met.

[0013] In step S5, the optimized solution is used to optimize the production process and carry out production, and the method of step S4 is used to perform testing again until the production requirements are met.

[0014] The optimized production method that meets the production requirements obtained in step S5 is saved; and the predicted performance parameters obtained under this parameter are used as standard values ​​to evaluate the performance parameters of the subsequently produced products and determine the quality coefficient of the batch of products.

[0015] A second aspect of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned diode package quality analysis and detection method when executing the computer program.

[0016] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-mentioned diode package quality analysis and detection method.

[0017] The beneficial effects of the present invention are as follows: the present invention formulates decision rules by establishing a decision tree model to classify data; determines the type and gain properties of the performance parameters affected by each input variable parameter; determines the correlation between different variable data by calculating them, and characterizes the degree of influence on specific performance parameters by the correlation; thereby predicting the quality of the production of the batch of products, and testing and verifying the products through existing testing methods. On the one hand, the recognition accuracy of the model can be determined according to the test and verification results, and the cost of subsequent testing and inspection can be reduced by using a large model; on the other hand, the deficiencies in the production process can be traced back according to the defects of subsequent products, and the processing technology can be optimized to improve the final product quality; specifically, taking the TO-18 diode packaging process and the corresponding testing method as an example, the temperature and welding time in the welding process are analyzed; the influence of different welding temperatures and times on the welding quality is analyzed, and the optimal welding temperature and time are optimized and generated. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 The figure is a flow chart of the diode packaging quality analysis and detection method. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1 A diode packaging quality analysis and detection method comprises the following specific steps: S1, establish a data set; collect production materials, process steps, process parameters and product performance parameters of different types of diodes; In step S1, different data sets are divided according to different types of diodes; Including plug-in package A and SMD package B; Among them, plug-in package A is divided into terminal type A1 and direct plug-in type A2 according to different types; According to the packaging material, it is divided into glass packaging and metal packaging; the terminal type plug-in diode of glass packaging is A11, and the terminal type plug-in diode of metal packaging is A12; The glass-encapsulated through-hole plug-in diode is A21, and the metal-encapsulated through-hole plug-in diode is A22; SMD package B is divided into SMA type B1, SMB type B2, SMC type B3 and SOD type B4.

[0024] In step S1, production raw materials, process steps, and process parameter data are collected for each data set established; Production materials include semiconductor materials, packaging materials and welding materials; The process steps include semiconductor material preparation steps and packaging steps; Process parameter data include the type, purity, and physical properties of production raw materials; process steps include operation time, temperature, pressure, and environmental conditions.

[0025] Specifically, semiconductor materials are divided into many types according to the raw materials and dopants, such as silicon, germanium, and gallium nitride; dopants: phosphorus, boron, factors affecting performance parameters also include the purity of semiconductor materials and doping concentration; in addition, it can be further subdivided according to supplier categories; common packaging materials are DPAK and SOT-23, for example, the thermal conductivity and dielectric constant of packaging materials will also affect performance parameters; welding materials and auxiliary materials have a greater impact on the performance parameters of diodes, and the main influencing factors include the type of solder, component ratio, and physical properties such as thermal conductivity and dielectric constant; The steps in the packaging process, including cutting, welding, molding, and the type and model of packaging equipment, will affect the performance of the final product.

[0026] Various parameter data in the process steps, such as temperature parameters, time parameters, pressure parameters, current and voltage parameters, temperature and atmosphere conditions, will have an impact on the product; therefore, detailed analysis is required after the above data is obtained.

[0027] S2, processing the data set; establishing the relationship between production raw materials, process steps, process parameters and product performance parameters; and sorting from low to high using correlation as the screening condition; In step S2, a first-order relationship function is established between the data elements in the production raw materials and the performance parameters through a public database, an internal enterprise database or empirical data by manual annotation, and written into the control conditions of the model; Compare each data element in the process steps and process parameters with the standard production steps and parameters to determine the specific indicator type and gain nature of the data element affecting the performance parameters, whether it is a positive gain or a negative gain; Correlation analysis is used to characterize the correlation between different parameters in step S2 for the same performance indicator: in, r is the correlation value; X i , Y i are the observed values ​​of the two parameters, , are the means of the two parameters respectively; Randomly combine any two parameters, determine the correlation between them, and establish a correlation database.

[0028] S3, model establishment and training; establish a decision tree model, use production materials, process steps, and process parameters as model input; product performance parameters as model output; select features with a high correlation with performance parameters to cross-train the model until the performance of the model meets the preset requirements; Information gain is used as the criterion for feature selection in step S3: IG(M,a)=H(M)-H(M|a) Among them, IG(M,a) is the information gain of feature a in data set M; H(M) is the entropy of data set M, and H(M|a) is the entropy of data set M under the condition of feature a.

[0029] S4, verification of the model; inputting the production materials, process steps and process parameter data of the diode package into the input end of the model; the model generates predicted performance parameters; performing performance parameter tests on diodes produced in the same batch, comparing the actual performance parameters obtained with the predicted performance parameters, and generating an optimization plan; In step S4, the same indicator is tested multiple times using different test equipment to reduce the error generated during the test process; in step S4, the deviation between the predicted performance parameter and the actual performance parameter is calculated, and the performance parameter with the largest deviation is taken to calculate the input parameter with the largest correlation with the performance parameter, so as to generate an optimization plan. Common diode performance parameters include current-voltage characteristics: forward current (IF), reverse current (IR), forward voltage (VF); thermal performance: maximum operating temperature, thermal resistance; reliability parameters: failure rate, environmental adaptability (high temperature resistance, moisture resistance).

[0030] S5, optimizing the production process using the optimization solution of step S4 until the production requirements are met.

[0031] In step S5, the optimized solution is used to optimize the production process and carry out production, and the method of step S4 is used to perform testing again until the production requirements are met.

[0032] The optimized production method that meets the production requirements obtained in step S5 is saved; and the predicted performance parameters obtained under this parameter are used as standard values ​​to evaluate the performance parameters of the subsequently produced products and determine the quality coefficient of the batch of products.

[0033] The following is a detailed description of the solution of the present invention using a specific example: TO-18 is a common diode packaging form; it refers to the process of packaging diode components in a TO-18 shell according to a specific size, shape and pin configuration; TO package is a cylindrical package, usually made of metal (such as iron or copper) to provide good thermal conductivity and electrical insulation; TO package consists of a TO tube socket and a TO tube cap, which together form a sealed package to protect sensitive semiconductor components; by consulting public data, it can be seen that the TO-18 packaging process mainly includes the following steps: 1. Preparation Material preparation: Prepare TO-18 package shell, leads, adhesives (such as epoxy resin) and other related materials.

[0034] Component preparation: The diodes to be packaged are cleaned to remove surface dirt and oxides to ensure good contact.

[0035] 2. Welding leads Lead welding: Solder the leads of the diode to the leads of the TO-18 package. Usually manual welding or automatic welding equipment is used.

[0036] Welding parameter control: Control welding temperature and time to ensure the reliability and strength of the solder joint.

[0037] 3. Package body assembly Place the diode: Place the soldered diode into the metal housing of the TO-18 package.

[0038] Secure position: Use adhesive to secure the diode in place within the housing to ensure it remains in place.

[0039] 4. Sealing Sealant Application: Apply a sealant (such as epoxy) to the top of the package to protect it from moisture and contamination.

[0040] Curing: Curing of sealants usually requires specific temperature and humidity conditions.

[0041] 5. Testing Electrical performance test: Conduct electrical performance test on the packaged diode to ensure it meets the specification requirements.

[0042] Visual inspection: Check the appearance of the package to ensure there are no defects such as cracks or bad soldering.

[0043] 6. Labeling and packaging Identification: Mark relevant information on the package, such as model number, batch number, etc.

[0044] Packaging: The packaged diodes should be properly packaged for transportation and storage.

[0045] When using this solution based on the above public data, you need to follow the following steps: S1, establish a data set; collect data for the TO-18 type diode to be packaged and processed; the data includes; The production materials, process steps, process parameters and product performance parameters of the diode; this data set is recorded as A1.

[0046] The data set contains data on production materials, process steps, and process parameters; records them by category; Specific: Production materials include semiconductor materials, packaging materials and welding materials; The process steps include semiconductor material preparation steps and packaging steps; Process parameter data include the type, purity, and physical properties of production raw materials; process steps include operation time, temperature, pressure, and environmental conditions.

[0047] The number of packages in the semiconductor industry is relatively large. In the actual packaging process, raw materials have a high impact on the final quality of the product. Taking the basic semiconductor raw materials that meet the needs of the factory as an example, it is assumed that the TO-18 package shell, leads, adhesives (such as epoxy resin) and other auxiliary materials in the raw materials meet the processing needs; the data of the standard raw materials are used for processing and the subsequent steps are executed.

[0048] S2, processing the data set; establishing the relationship between production raw materials, process steps, process parameters and product performance parameters; and sorting from low to high using correlation as the screening condition; In step S2, a first-order relationship function is established between the data elements in the production raw materials and the performance parameters through a public database, an internal enterprise database or empirical data by manual annotation, and written into the control conditions of the model; Compare each data element in the process steps and process parameters with the standard production steps and parameters to determine the specific indicator type and gain nature of the data element affecting the performance parameters, whether it is a positive gain or a negative gain; Specifically: in the process of welding leads, automatic welding equipment is usually used for processing; in automatic welding, the welding temperature and time are fixed parameters, and the yield rates of different types of automatic equipment during the welding process are also different; by actually processing different equipment, obtaining processing data, and systematically analyzing the obtained data, the yield rate data, processing parameter data and optimal processing parameters of different equipment can be obtained; for example, when using equipment T1 and using 63% solder as solder for welding packaging, the temperature needs to be controlled at 190-200 degrees. Experiments have found that when the temperature is controlled at 193 degrees, the yield rate is the highest, so 193 degrees is the optimal temperature under current conditions; when using equipment T2 for packaging, 63% solder is also used as solder, and the temperature needs to be controlled at 200-215 degrees. Experiments have found that when the temperature is controlled at 204 degrees, the yield rate is the highest, so 204 degrees is the optimal temperature under current conditions. The temperature can only ensure that the solder wire melts quickly, which is beneficial to the purpose of packaging; the time during the welding process is also an important factor affecting the welding quality; when the temperature is high, the required welding time is correspondingly shortened, and when the welding temperature is low, the required welding time is correspondingly increased. The most intuitive reflection of the welding quality is the fullness of the solder joint, which affects the conductive performance of the pin on the one hand, and the stability and sealing of the diode on the other hand; Therefore, for more than two variables, correlation analysis is used to characterize the correlation between different parameters in the steps and different performance indicators: in, r is the correlation value; X i , Y i are the observed values ​​of the two parameters, , are the means of the two parameters respectively; Randomly combine any two parameters, determine the correlation between them, and establish a correlation database; prioritize putting variables that affect the same performance parameter index together for calculation to obtain correlation values, such as the welding temperature and welding time in the previous section; characterize the degree of association between different variable parameters and different performance parameters by calculating the obtained correlation values; obtain the optimal temperature-time data set and establish a corresponding database; the data element that meets the data set is a positive gain, otherwise it is a negative gain; for example, in the above case, when using T1 equipment for processing, the temperature is controlled at 193 degrees Celsius and the welding time is controlled at 5 seconds to achieve the best welding quality; when the temperature is 198 degrees Celsius, due to the slightly higher temperature, when the welding is still maintained for 5 seconds, the amount of solder will increase and the size of the solder joint will be slightly larger; at this time, its conductivity with the pin and the stability and sealing of the diode remain consistent with the optimal temperature-time, but the increase in the solder joint may cause solder joint interference and there is a risk of short circuit; the correlation value of its performance parameters related to short circuit will increase.

[0049] In this embodiment, the welding method of the welding equipment will also affect the firmness of the diode after packaging and the quality of the pins; for example, the energy storage capping welding machine, whose power supply part adopts capacitor energy storage discharge welding, has the functions of adjusting energy storage voltage, charging and discharging, pneumatic pressurization, pneumatic unloading, etc., stable and reliable operation, perfect protection function, convenient adjustment of process parameters, safe and reliable operation; it is suitable for the packaging welding of high-power devices such as transistor laser diodes; and the medium frequency inverter capping welding machine adopts medium frequency inverter technology, with higher welding efficiency, suitable for scenes with high requirements on welding quality, stable welding process, and good consistency of solder joint quality; therefore, in the process of producing TO-18 diodes, the medium frequency inverter capping welding machine is used for better performance, so in the type of welding equipment, the medium frequency inverter capping welding machine is used for positive gain, and the energy storage capping welding machine is used for negative gain; the above formula is used to calculate the obtained temperature-time data element and welding equipment data element again, and when a new correlation representation is obtained, the correlation degree of multiple variables to welding performance parameters is calculated.

[0050] In this embodiment, during the packaging of the TO-18 diode, the adhesive and sealant used in the packaging body will affect the stability of the product during use; the adhesive material will affect the firmness of the diode bonding and the drying time; and the sealant material will affect the sealing effect; the failure of the sealant will cause moisture and contamination inside the diode, which will inevitably affect the performance of the final product. Therefore, the above formula can be used to calculate the correlation between different data elements to determine the degree of correlation between the variable parameters and the performance of the final product.

[0051] S3, model establishment and training; establish a decision tree model, use production materials, process steps, and process parameters as model input; product performance parameters as model output; select highly relevant features to cross-train the model until the model performance meets the preset requirements; Information gain is used as the criterion for feature selection in step S3: IG(M,a)=H(M)-H(M|a) Among them, IG(M,a) is the information gain of feature a in data set M; H(M) is the entropy of data set M, and H(M|a) is the entropy of data set M under the condition of feature a. The screening criteria are set manually, and the information gain that meets the set value is used as the feature for training.

[0052] In this step, a decision tree model is established, taking the aforementioned welding temperature and time as an example; when the model input is temperature and time, it is first compared with the data in the established data set to determine whether the input data element is a positive gain or a negative gain; for example, in the weld evaluation of welding quality, the influence of temperature is higher than that of time; and the influence of the welding equipment model on the welding quality is relatively small. When calculating the information gain, the information gain value of temperature is greater than the information gain value of time and welding equipment model; therefore, data with a larger temperature information gain is preferentially used as feature data to train the model, thereby reducing the amount of calculation of the model and enabling it to converge quickly; and then the time information gain data and the information gain data of the welding equipment model are used for verification to provide the accuracy of the model.

[0053] S4, verification of the model; inputting the production materials, process steps and process parameter data of the diode package into the input end of the model; the model generates predicted performance parameters; performing performance parameter tests on diodes produced in the same batch, comparing the actual performance parameters obtained with the predicted performance parameters, and generating an optimization plan; In step S4, the same indicator is tested multiple times using different test equipment to reduce the error generated during the test process; in step S4, the deviation between the predicted performance parameter and the actual performance parameter is calculated, and the performance parameter with the largest deviation is taken to calculate the input parameter with the largest correlation with the performance parameter, so as to generate an optimization plan. Common diode performance parameters include current-voltage characteristics: forward current (IF), reverse current (IR), forward voltage (VF); thermal performance: maximum operating temperature, thermal resistance; reliability parameters: failure rate, environmental adaptability (high temperature resistance, moisture resistance).

[0054] S5, optimizing the production process using the optimization solution of step S4 until the production requirements are met.

[0055] In step S5, the optimized solution is used to optimize the production process and carry out production, and the method of step S4 is used to perform testing again until the production requirements are met.

[0056] The optimized production method that meets the production requirements obtained in step S5 is saved; and the predicted performance parameters obtained under this parameter are used as standard values ​​to evaluate the performance parameters of the subsequently produced products and determine the quality coefficient of the batch of products.

[0057] For easier understanding of the solution, please refer to the following simplified example.

[0058] Table 1 Diode processing data parameter table Suppose that the correlation between each pair of features is calculated. For example: Correlation between raw material type and performance parameters r=0.65 Correlation between process steps and performance parameters r=0.70 The correlation between process parameters (temperature) and performance parameters is r=0.85 Select features with high correlation (such as process parameters) to establish and train the model; use the decision tree model, input features as "raw material type", "process steps" and "process parameters", train and calculate information gain, and select features with high information gain as the validation set to train the model.

[0059] Assume that after training, the mean square error of the model is MSE=0.03, which meets the preset requirements.

[0060] Enter the production data of a new batch: Raw material type: A Process steps: P1 Process parameters: 190°C The model predicted performance parameter was 20.9mA, and the actual test result was 21.8mA.

[0061] Compare the results and generate an optimization plan: Calculate the error between actual and predicted: Error = | 20.9 − 21.8 | = 0.9 mA According to the difference between the prediction and the actual, the process parameters are optimized, such as adjusting the temperature, and the process parameters can be adjusted to 195°C.

[0062] A new round of testing and verification is conducted, and step S4 is repeated for verification until the production requirements are met. The above is only an exemplary introduction to this solution. The specific optimization solution is subject to the joint determination of multiple variables. In the process of generating the optimization solution, the variable parameters that have a great influence on the performance parameters are adjusted first, and then the parameters (correlation) that have a small influence on the performance parameters are adjusted, and finally a suitable optimization solution is selected to optimize the processing steps.

[0063] The present invention utilizes a large model to reduce the cost of subsequent testing and inspection, can trace the deficiencies in the production process according to the defects that occur in subsequent products, and optimize the processing technology to improve the final product quality.

[0064] Example 2 This embodiment also provides a computer device, which is suitable for the steps of the diode package quality analysis and detection method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the steps of the diode package quality analysis and detection method proposed in the above embodiment. The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0065] Example 3 This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the diode package quality analysis and detection method described in Embodiment 1 are implemented.

[0066] The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination of them, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0067] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A diode packaging quality analysis and detection method, characterized in that: The specific steps include: S1, establish a data set; collect production materials, process steps, process parameters and product performance parameters of different types of diodes; S2, processing the data set; establishing the relationship between production raw materials, process steps, process parameters and product performance parameters; S3, model establishment and training; establish a decision tree model, use production materials, process steps, and process parameters as model input; product performance parameters as model output; select features with a high correlation with performance parameters to cross-train the model until the performance of the model meets the preset requirements; S4, Validation of the model; Input the production materials, process steps and process parameter data of the diode package into the input end of the model; the model generates predicted performance parameters; perform performance parameter tests on diodes produced in the same batch, compare the actual performance parameters obtained with the predicted performance parameters, and generate an optimization plan; S5, optimizing the production process using the optimization solution of step S4 until the production requirements are met.

2. A diode packaging quality analysis and detection method according to claim 1, characterized in that: In step S1, different data sets are divided according to different types of diodes; Including plug-in package A and SMD package B; Among them, plug-in package A is divided into terminal type A1 and direct plug-in type A2 according to different types; According to the packaging material, it is divided into glass packaging and metal packaging; the terminal type plug-in diode of glass packaging is A11, and the terminal type plug-in diode of metal packaging is A12; The glass-encapsulated through-hole plug-in diode is A21, and the metal-encapsulated through-hole plug-in diode is A22; SMD package B is divided into SMA type B1, SMB type B2, SMC type B3 and SOD type B4.

3. A diode packaging quality analysis and detection method according to claim 2, characterized in that: In step S1, production raw materials, process steps, and process parameter data are collected for each data set established; Production materials include semiconductor materials, packaging materials and welding materials; The process steps include semiconductor material preparation steps and packaging steps; Process parameter data include the type, purity, and physical properties of production raw materials; process steps include operation time, temperature, pressure, and environmental conditions.

4. The diode packaging quality analysis and detection method according to claim 1, characterized in that: In step S2, the public database, the internal database of the enterprise or the experience data are manually annotated. The data elements in the production raw materials are all established with the first-order relationship function of the performance parameters and written into the control conditions of the model; Compare each data element in the process steps and process parameters with the standard production steps and parameters to determine the specific indicator type and gain nature of the data element affecting the performance parameters, whether it is a positive gain or a negative gain.

5. The diode packaging quality analysis and detection method according to claim 1, characterized in that: Correlation analysis is used to characterize the correlation between different parameters in step S2 for the same performance indicator: in, r is the correlation value; X i , Y i are the observed values ​​of the two parameters, , are the means of the two parameters respectively; Any two input parameters that affect the same performance parameter are randomly combined, the correlation between them is determined, and a correlation database is established.

6. The diode packaging quality analysis and detection method according to claim 1, characterized in that: Information gain is used as the criterion for feature selection in step S3: IG(M,a)=H(M)-H(M|a) Among them, IG(M,a) is the information gain of feature a in data set M; H(M) is the entropy of data set M, and H(M|a) is the entropy of data set M under the condition of feature a.

7. The diode packaging quality analysis and detection method according to claim 1, characterized in that: In step S4, different testing equipment is used to perform multiple tests on the same indicator to reduce the error generated during the test process.

8. The diode packaging quality analysis and detection method according to claim 1, characterized in that: In step S4, the deviation between the predicted performance parameter and the actual performance parameter is calculated, a performance parameter with the largest deviation is taken, and the input parameter with the largest correlation with the performance parameter is calculated, so as to generate an optimization plan.

9. The diode packaging quality analysis and detection method according to claim 1, characterized in that: In step S5, the optimized solution is used to optimize the production process and carry out production, and the method of step S4 is used to perform testing again until the production requirements are met.

10. The diode packaging quality analysis and detection method according to claim 1, characterized in that: The optimized production method that meets the production requirements obtained in step S5 is saved; and the predicted performance parameters obtained under this parameter are used as standard values ​​to evaluate the performance parameters of the subsequently produced products and determine the quality coefficient of the batch of products.

Images