A sorting die bonding precision compensation method

Abstract

The application discloses a kind of sorting die-bonding precision compensation methods, the method includes the following steps: using sorting equipment to sort die; collect the historical data of last time processing of different models substrate, and according to the historical data of classification according to substrate model of acquisition, data storage is simultaneously established database;The substrate to be processed is matched with the data in database, and the processing parameters of historical data are obtained, the deviation value in historical data is used to establish the compensation relationship model between suction nozzle and die, and the initial processing parameters in historical data are used as preset processing parameters, before die-bonding, the preset processing parameters are adjusted using compensation model, and accurate initial processing parameters are obtained.The application can improve the accuracy of initial setting parameters by collecting historical data and compensating and adjusting preset parameters according to historical deviation data, and also has the function of real-time adjusting die-bonding position, to improve the accuracy of each die-bonding.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing analysis, in particular to a sorting die bonding precision compensation method. BACKGROUND

[0002] Die bonding is a crucial link in the semiconductor manufacturing process, and the precision of die bonding directly affects the performance and quality of semiconductor devices. Currently, due to various factors such as equipment precision, process parameters, material properties, etc., there are often certain precision deviations in the die bonding process. These deviations may cause problems such as inaccurate position and angle deviation of the die during sorting and die bonding, thereby affecting the yield and performance of the product.

[0003] To solve these problems, in traditional die bonding operations, manual searching and calculation of data are usually required to determine the best processing parameters. The parameter determination method does not determine the parameters according to historical data, and the manual parameter determination method not only has low efficiency, but also reduces the accuracy of the first parameter setting. In addition, since there are deviations in each die bonding operation, the traditional die bonding operation does not calculate the deviation by image shooting for each die bonding operation, so it cannot timely detect the position deviation and angle deviation problems in the die bonding process, thereby affecting the die bonding precision.

[0004] Therefore, there is an urgent need for a new sorting die bonding precision compensation method that can automatically and accurately compensate for the precision deviation in the die bonding process to improve the yield and performance of the product to meet the high precision requirements of modern semiconductor manufacturing.

[0005] At present, there is no effective solution to the problems in the related art. SUMMARY

[0006] To overcome the above technical problems existing in the prior art, the present application proposes a sorting die bonding precision compensation method.

[0007] To this end, the specific technical solutions adopted by the present application are as follows:

[0008] A sorting die bonding precision compensation method, the method comprising the following steps:

[0009] S1, using a sorting device to sort the die;

[0010] S2, collecting historical data of the last time the different types of substrates were processed, and classifying the collected historical data according to the substrate type, and establishing a database to store the data;

[0011] S3, match the substrate to be processed with the data in the database, obtain the processing parameters of the historical data, establish a compensation relationship model between the suction nozzle and the die based on the deviation value in the historical data, use the initial processing parameters in the historical data as the preset processing parameters, adjust the preset processing parameters using the compensation model before die bonding, and obtain accurate initial processing parameters;

[0012] S4, use the obtained initial parameters to perform die bonding operation, take a photo of the grabbed die after each time the die is grabbed, record the image, and perform defect detection on the external surface of the die according to the photographed image;

[0013] S5, compare the position of the current grabbed die with the position of the last die, calculate the relative deviation value, and calculate the compensation value according to the deviation value;

[0014] S6, output the compensation value and perform die bonding operation according to the compensated target position to ensure that the die is accurately solidified on the substrate;

[0015] S7, after the processing is completed, integrate the processing data, input the database, and replace the historical data of the current category substrate processing in the database.

[0016] As a preferred embodiment, the historical data of the last time the different models of substrates are processed is collected, and the collected historical data is classified according to the substrate model, and a database is established to store the data, including the following steps:

[0017] S21, collect processing parameters, quality parameters, equipment status, and environmental parameters, the processing parameters include die bonding pressure, temperature, and time, the quality parameters include error parameters and compensation amount, the equipment status includes equipment number and calibration record, and the environmental parameters include temperature and humidity and vibration level;

[0018] S22, classify the collected data according to the substrate model;

[0019] S23, clean and format the data, and remove error records;

[0020] S24, create a database and import the sorted data.

[0021] The last time processing data can most accurately reflect the state of the current production equipment, process and environment, and the decisions made based on these data are more targeted and timely, which can timely adjust the production parameters and optimize the process flow to adapt to the changes of actual production situation.

[0022] As a preferred embodiment, the matching of the substrate to be processed with the data in the database, the acquisition of the processing parameters in the historical data, the establishment of the compensation relationship model between the nozzle and the die, the use of the initial processing parameters in the historical data as the preset processing parameters, the adjustment of the preset processing parameters by using the compensation model before die bonding, and the acquisition of the accurate initial processing parameters include the following steps:

[0023] S31, input the substrate model to be processed, and match the historical data of the substrate processing of the same model from the database;

[0024] S32, acquire the deviation value in the historical data, and establish the compensation relationship model between the nozzle and the die;

[0025] S33, acquire the initial processing parameters from the historical data, and use the initial processing parameters as the preset processing parameters, and finally adjust the preset processing parameters by using the compensation relationship model.

[0026] The selected historical processing parameters are the historical data of the last time, which is closest to the current processing situation in time. The production environment, equipment state and other factors usually do not change greatly in a short time, so these parameters can better reflect the current possible best processing state, and have high applicability.

[0027] As a preferred embodiment, the acquisition of the deviation value in the historical data and the establishment of the compensation relationship model between the nozzle and the die include the following steps:

[0028] S321, select a linear regression model to describe the compensation relationship between the nozzle and the die, and the formula is:

[0029] W x =a x ·p+b x ·m+c x ·t+d x ;

[0030] W y =a y ·p+b y ·m+c y ·t+d y ;

[0031] Wherein, W x , W y respectively represent the deviation values of the X-axis and the Y-axis, a x , b x , c x , d x and a y , b y , c y , dy are linear regression coefficients for X-axis and Y-axis respectively, p is the die bonding pressure, m is the die bonding temperature, and t is the die bonding time;

[0032] The compensation relationship formula is:

[0033] p' = p + Δp;

[0034] m' = m + Δm;

[0035] t' = t + Δt;

[0036] wherein p" is the compensation pressure, m' is the compensation temperature, and t" is the compensation time;

[0037] S322, output the calculated compensation parameters as the actual determined processing parameters.

[0038] As a preferred embodiment, the die bonding operation using the obtained initial parameters is performed, and after each time a die is grabbed, the grabbed die is photographed, the image is recorded, and the external surface of the die is detected for defects according to the photographed image, including the following steps:

[0039] S41, input the determined processing parameters and perform a grabbing action;

[0040] S42, the photographing device automatically performs a photographing function to photograph the grabbed die, and the photographing content includes an external image of the die and a position image of the die;

[0041] S43, pre-process the photographed external image of the die, including denoising, enhancing contrast, adjusting brightness, and enlarging;

[0042] S44, use an image processing algorithm to process the image, automatically identify defects on the surface of the die, including cracks, scratches, foreign matter, and uneven coating;

[0043] S45, when a defect is identified on the external surface of the die, the defective die is grabbed to a special processing channel for rework or scrap, and when no defect is identified on the external surface of the die, the die bonding operation is continued.

[0044] The photographing mode is a continuous photographing mode, which can obtain multiple instantaneous images, increase the observation angle and information amount of the die state, and also capture the performance of the die under different postures and lighting conditions.

[0045] As a preferred embodiment, the position of the currently grabbed die is compared with the position of the last die, a relative deviation value is calculated, and a compensation value is calculated according to the deviation value, including the following steps:

[0046] S51, preprocessing the position image of the currently photographed die and the position image of the last photographed die;

[0047] S52, locating the position of the die using an image processing algorithm to determine the centroid position of the currently captured die and the centroid position of the last die;

[0048] S53, calculating the difference between the current die position and the last die position to obtain a relative deviation value, and using the obtained deviation value to calculate a corresponding compensation value;

[0049] S54, adjusting the processing parameters using the calculated compensation value before the next die bonding operation.

[0050] Since there will be errors in each capture, image shooting will be performed each time the capture operation is performed, and the specific deviation in the image is calculated. This method can adapt to different working environments and working conditions, and improve the accuracy of capture.

[0051] As a preferred embodiment, after the processing is completed, the processing data is integrated, input into the database, and the historical data of the current category substrate processing in the database is replaced, including the following steps:

[0052] S71, collecting all related data during processing;

[0053] S72, screening the collected raw data to remove invalid or abnormal data, and arranging the data into a unified format;

[0054] S73, integrating all data into a complete data package, and connecting the software to the database;

[0055] S74, adding the data of the current processing to the related table of the database, querying the database, screening out the historical data associated with the current processing category substrate, and performing a deletion operation.

[0056] In the case of limited database storage capacity, replacing data can avoid database overgrowth, save storage space, and also make database management more efficient.

[0057] The beneficial effects of the present application are:

[0058] 1. The present application can quickly determine the first processing parameter of the subsequent different type substrates by collecting the historical data of the last processing of different type substrates and classifying and storing the data, in addition, it also effectively simplifies the operation process, so that subsequent die bonding does not need to search a large number of parameters for parameter determination.

[0059] 2. This invention obtains historical deviation values ​​and processing parameters from historical data, and uses the historical deviation values ​​to establish a compensation relationship model between the nozzle and the grain. This allows for precise adjustment of the differences between the nozzle and the grain. Thus, the preset parameters can be adjusted using this model before die bonding to obtain accurate initial processing parameters, thereby greatly improving the accuracy of the initial processing parameters.

[0060] 3. This invention first performs defect detection on the outside of the grain by taking an image each time the grain is picked up. When a defect is detected, the defective grain is picked up to a special channel to facilitate the grain to be reworked or scrapped. This improves the control level of product quality, and timely identification and separation of defective grains can prevent defective products from flowing into subsequent processes, reducing unnecessary processing and waste of resources.

[0061] For die-bonding without external defects, the die position image is processed to calculate the deviation and compensation values. Then, the compensation value is output and the die bonding operation is performed according to the compensated target position. This method can dynamically adjust the die bonding position in real time to ensure that each die can be bonded to the substrate as accurately as possible, which greatly improves the accuracy and consistency of die bonding and effectively reduces the product defect rate caused by inaccurate positioning. In addition, this method can also adapt to various changes and uncertainties in the production process, making the production process more stable and reliable.

[0062] 4. By integrating the processed data and inputting it into the database to replace the relevant historical data, this invention ensures that the database always stores the latest processing information, providing the most timely reference for the processing of subsequent substrates of the same type. In addition, replacing the data can also prevent the database from becoming excessively large and save storage space.

[0063] 5. This invention uses historical data to obtain historical deviation data and historical initial setting data, so that the staff does not need to perform a lot of data search and calculation to determine the initial processing parameters. In addition, the preset initial parameters are adjusted by the deviation value of the most recent historical data, so that the parameters during the first die bonding operation are more accurate. Furthermore, during the die bonding operation, it also has the function of dynamically adjusting the die bonding position in real time according to the captured grain image, thereby improving the accuracy of each die bonding operation. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 is a flowchart of a sorting die bonding precision compensation method according to an embodiment of the present application. DETAILED DESCRIPTION

[0066] To further illustrate the embodiments, the present application provides accompanying drawings, which are part of the disclosure of the present application, mainly used to illustrate the embodiments, and can be used to explain the operating principle of the embodiments in conjunction with the related description of the specification. Those skilled in the art should understand other possible implementations and advantages of the present application by referring to these contents. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0067] According to an embodiment of the present application, a sorting die bonding precision compensation method is provided.

[0068] The present application will be further described in conjunction with the accompanying drawings and specific embodiments. As shown in the drawings, according to an embodiment of the present application, a sorting die bonding precision compensation method includes the following steps: Figure 1

[0069] S1, using a sorting device to sort die grains;

[0070] S2, collecting historical data of the last time of processing of different types of substrates, and classifying the collected historical data according to the substrate type, and establishing a database to store the data;

[0071] Further, the collecting historical data of the last time of processing of different types of substrates, and classifying the collected historical data according to the substrate type, and establishing a database to store the data includes the following steps:

[0072] S21, collecting processing parameters, quality parameters, equipment status, and environmental parameters, the processing parameters including die bonding pressure, temperature, and time, the quality parameters including error parameters and compensation amount, the equipment status including equipment number and calibration record, and the environmental parameters including temperature and humidity and vibration level;

[0073] S22, classifying the collected data according to the substrate type;

[0074] S23, cleaning and formatting the data to remove error records;

[0075] S24, creating a database and importing the sorted data.

[0076] It should be noted that classifying the historical data according to the substrate type in advance can facilitate subsequent parameter selection and determination of the same type of substrate. When the same type of substrate is processed subsequently, the historical data can be quickly matched through the substrate type name.

[0077] ​S3, match the substrate to be processed with the data in the database, obtain the processing parameters of the historical data, establish a compensation relationship model between the suction nozzle and the die based on the deviation value in the historical data, use the initial processing parameters in the historical data as the preset processing parameters, adjust the preset processing parameters using the compensation model before die bonding, and obtain accurate initial processing parameters;

[0078] Further, the matching of the substrate to be processed with the data in the database, the obtaining of the processing parameters of the historical data, the establishment of the compensation relationship model between the suction nozzle and the die based on the deviation value in the historical data, the use of the initial processing parameters in the historical data as the preset processing parameters, the adjustment of the preset processing parameters using the compensation model before die bonding, and the obtaining of accurate initial processing parameters include the following steps:

[0079] S31, input the substrate model to be processed, and match the historical data of the same model substrate processing from the database;

[0080] S32, obtain the deviation value in the historical data, and establish a compensation relationship model between the suction nozzle and the die;

[0081] S33, obtain the initial processing parameters from the historical data, and use the initial processing parameters as the preset processing parameters, and finally adjust the preset processing parameters using the compensation relationship model.

[0082] It should be noted that the preset processing parameters are usually set before the die bonding operation, and when the substrate type is the same, the preset processing parameters are usually the same value. The historical processing parameters have been set, so the initial processing parameters in the historical data are directly used as the preset parameters, which can save a lot of parameter debugging time and avoid the process of repeatedly testing different parameter combinations, thereby improving the processing efficiency.

[0083] Further, the obtaining of the deviation value in the historical data and the establishment of the compensation relationship model between the suction nozzle and the die include the following steps:

[0084] S321, select a linear regression model to describe the compensation relationship between the suction nozzle and the die, and the formula is:

[0085] W x =a x ·p+b x ·m+c x ·t+d x ;

[0086] W y =a y ·p+b y ·m+c y ·t+d y ;

[0087] wherein, W x , W y respectively represent the deviation values of the X-axis and the Y-axis, a x , b x , c x , d x and a y , b y , c y , d y are linear regression coefficients for the X-axis and the Y-axis respectively, p is the die bonding pressure, m is the die bonding temperature, and t is the die bonding time;

[0088] The compensation relationship formula is:

[0089] p' = p + Δp;

[0090] m' = m + Δm;

[0091] t' = t + Δt;

[0092] wherein, p" is the compensation pressure, m' is the compensation temperature, and t" is the compensation time;

[0093] S322, output the calculated compensation parameters as the actual determined processing parameters.

[0094] It should be noted that the compensation relationship model is established according to the deviation values in the historical data, which can be accurately adjusted for different processing conditions. Adjusting the parameters before die bonding can effectively reduce the production of defective products caused by the mismatch between the suction nozzle and the die, and improve the yield of the product. In addition, the compensation model is used to adjust the preset processing parameters, without the need for a long trial-and-error process, and the accurate initial processing parameters can be quickly obtained.

[0095] S4, using the obtained initial parameters to perform die bonding operation, after each time the die is grabbed, the grabbed die is photographed, the image is recorded, and the external surface of the die is detected according to the photographed image;

[0096] Further, using the obtained initial parameters to perform die bonding operation, after each time the die is grabbed, the grabbed die is photographed, the image is recorded, and the external surface of the die is detected according to the photographed image, including the following steps:

[0097] S41, input the determined processing parameters and execute the grabbing action;

[0098] S42, the photographing device automatically executes the photographing function, photographs the grabbed die, and the photographing content includes the external image of the die and the position image of the die;

[0099] S43, pre-process the captured external image of the die, including denoising, enhancing contrast, adjusting brightness, and enlarging;

[0100] S44, process the image using image processing algorithms to automatically identify defects on the surface of the die, including cracks, scratches, foreign matter, and uneven coating;

[0101] S45, when a defect is identified on the exterior of the die, the defective die is picked up and sent to a special processing channel for rework or scrap, and if no defect is identified on the exterior of the die, the die continues to the die bonding operation.

[0102] It should be noted that when a defect is found on the exterior of the die, the next time the die is picked up for die bonding, the position information from the previous time will be used for die bonding to ensure the continuity and stability of the die bonding process.

[0103] S5, compare the position of the currently picked-up die with the position of the previous die to calculate a relative deviation value, and calculate a compensation value based on the deviation value;

[0104] Further, the comparison of the position of the currently picked-up die with the position of the previous die to calculate a relative deviation value, and the calculation of a compensation value based on the deviation value include the following steps:

[0105] S51, pre-process the position image of the currently captured die and the position image of the previously captured die;

[0106] S52, use image processing algorithms to locate the position of the die to determine the centroid position of the currently picked-up die and the centroid position of the previous die;

[0107] S53, calculate the difference between the position of the current die and the position of the previous die to obtain a relative deviation value, and use the obtained deviation value to calculate a corresponding compensation value;

[0108] S54, before the next die bonding operation, use the calculated compensation value to adjust the processing parameters.

[0109] It should be noted that this way can monitor the position change in the picking process in real time, find possible deviations in time, improve the accuracy and stability of picking, secondly, the calculation of the compensation value can accurately adjust the picking position, effectively reduce the position error caused by various factors, and ensure that each die can be accurately picked and placed, thereby improving production efficiency and product quality;

[0110] In the application of multi-nozzle structure, in order to improve the overall system relative accuracy, the following adjustment steps can be taken: first, for each nozzle in different grabbing situations, the system will obtain the specific deviation value of each nozzle when grabbing the die, which reflects the difference between the actual operation and the ideal position, and is the basis for subsequent compensation calculation;

[0111] Next, according to the obtained deviation value, the compensation amount is calculated, and through these calculations, the specific compensation amount required for each nozzle is obtained;

[0112] Finally, the compensation amount calculated by these calculations will be accurately applied to the corresponding nozzle.

[0113] This compensation not only includes fine-tuning of position, but also involves adjustment of force and speed to ensure that each nozzle can grab and place the die with the highest accuracy under various operating conditions, so that the overall relative accuracy of the multi-nozzle structure is significantly improved, thereby further ensuring the efficient and stable operation of the entire system.

[0114] S6, output the compensation value and perform die bonding operation according to the compensated target position to ensure that the die is accurately bonded to the substrate;

[0115] S7, after processing, integrate the processing data, input the database, and replace the historical data of the current category substrate processing in the database.

[0116] Further, the processing data is integrated after processing, input into the database, and the historical data of the current category substrate processing in the database is replaced, including the following steps:

[0117] S71, collect all related data during processing;

[0118] S72, filter the collected raw data to remove invalid or abnormal data, and arrange the data into a unified format;

[0119] S73, integrate all data into a complete data package, and connect the software to the database;

[0120] S74, add the data of the current processing to the related table of the database, query the database, filter out the historical data associated with the current processing category substrate, and perform the delete operation.

[0121] It should be noted that the processing data is integrated after processing and input into the database, and the historical data of the current category substrate processing is replaced, which can ensure that the data in the database is always up-to-date, providing the most realistic reference for subsequent processing of the same category substrate, thereby greatly improving the accuracy and efficiency of parameter setting.

[0122] In summary, the present application can quickly determine the first processing parameter of subsequent different models of substrates by collecting the historical data of the last processing of different models of substrates, classifying and storing the data, and effectively simplifying the operation process, so that subsequent die bonding does not need to search a large number of parameters for parameter determination; the present application can accurately adjust the difference between the suction nozzle and the die by obtaining the historical deviation value and the number of processing parameters from the historical data and establishing a compensation relationship model between the suction nozzle and the die using the historical deviation value, so as to adjust the preset parameter using the model before die bonding to obtain accurate initial processing parameters, thereby greatly improving the accuracy of the initial processing parameters; the present application can detect defects on the outside of the die in the image first when grabbing the die each time, and when defects are detected, the defective die is grabbed to a special channel, so as to facilitate the next rework or scrap processing of the die, thereby improving the control level of product quality, identifying and separating defective dies in time, avoiding the flow of defective products into subsequent processes, and reducing unnecessary processing and resource waste; for dies without defects on the outside, the die position image is processed to calculate the deviation value and the compensation value, and then the compensation value is output and die bonding operation is performed according to the compensated target position, which can dynamically adjust the die bonding position in real time to ensure that each die can be bonded to the substrate as accurately as possible, greatly improving the precision and consistency of die bonding, effectively reducing the product failure rate caused by inaccurate position, and secondly, this method can also adapt to various changes and uncertainties in the production process, making the production process more stable and reliable; the present application can ensure that the database always stores the latest processing information by integrating the processed data and inputting the database to replace the related historical data, providing the most time-effective reference for subsequent processing of similar substrates, and in addition, replacing the data can also avoid database overgrowth and save storage space; the present application uses historical data to obtain historical deviation data and historical initial setting data, so that the staff does not need to search and calculate a large amount of data to determine the initial processing parameter, and in addition, the deviation value of the last historical data is used to adjust the preset initial parameter, so that the parameter of the first die bonding operation is more accurate, and in addition, the die bonding operation also has the function of adjusting the die bonding position in real time and dynamically according to the photographed die image, improving the accuracy of each die bonding.

[0123] The above only describes the preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for compensating the accuracy of die sorting, characterized in that, The method comprises the following steps: S1, using a sorting device to sort the die; S2, collecting historical data of the last time the different models of substrates were processed, classifying the collected historical data according to the substrate model, and establishing a database to store the data; S21, collecting processing parameters, quality parameters, equipment states, and environmental parameters, the processing parameters including die bonding pressure, temperature, and time, the quality parameters including error parameters and compensation amounts, the equipment states including equipment numbers and calibration records, and the environmental parameters including temperature and humidity and vibration levels; S3, matching the substrate to be processed with the data in the database to obtain the processing parameters of the historical data, establishing a compensation relationship model between the suction nozzle and the die according to the deviation value in the historical data, using the initial processing parameters in the historical data as preset processing parameters, adjusting the preset processing parameters using the compensation model before die bonding to obtain accurate initial processing parameters; S4, using the obtained initial parameters to perform die bonding operations, taking a photo of the die after each time the die is grabbed, recording the image, and detecting defects on the external surface of the die according to the photographed image; S41, inputting the determined processing parameters to perform a grabbing action; S42, the photographing device automatically performs a photographing function to take a photo of the grabbed die, and the photographing content includes the external image of the die and the position image of the die; S43, preprocessing the external image of the photographed die, including denoising, enhancing contrast, adjusting brightness, and enlarging; S44, using an image processing algorithm to process the image to automatically identify defects on the surface of the die, including cracks, scratches, foreign matter, and uneven coating; S45, when a defect is identified on the external surface of the die, the defective die is grabbed to a special processing channel to make the die be reworked or scrapped, and when no defect is identified on the external surface of the die, the die bonding operation continues; S5, comparing the position of the currently grabbed die with the position of the last die to calculate a relative deviation value, and calculating a compensation value according to the deviation value; S6, outputting the compensation value and performing die bonding operations according to the target position after compensation to ensure that the die is accurately bonded to the substrate; S7, after the processing is completed, integrating the processing data, inputting the data into the database, and replacing the historical data of the current category of substrates in the database.

2. The method of claim 1, wherein the method further comprises: The collecting of the historical data of the last time the different models of substrates were processed, the classification of the collected historical data according to the substrate model, and the establishment of a database to store the data further comprise the following steps: S22, classifying the collected data according to the substrate model; S23, cleaning and formatting the data to remove error records; S24, creating a database and importing the sorted data.

3. The method of claim 1, wherein the method further comprises: The matching of the substrate to be processed with the data in the database to obtain the processing parameters of the historical data, the establishment of a compensation relationship model between the suction nozzle and the die according to the deviation value in the historical data, the use of the initial processing parameters in the historical data as preset processing parameters, the adjustment of the preset processing parameters using the compensation model before die bonding to obtain accurate initial processing parameters comprise the following steps: S31, input the substrate model to be processed, match the historical data of the same model substrate processing from the database; S32, obtain the deviation value in the historical data, and establish a compensation relationship model between the suction nozzle and the die; S33, obtain the initial processing parameters from the historical data, and use the compensation relationship model to adjust the preset processing parameters.

4. The sorting and die-bonding accuracy compensation method according to claim 3, characterized in that, The compensation relationship model between the suction nozzle and the die includes the following steps: S321, a linear regression model is selected to describe the compensation relationship between the suction nozzle and the die, and the formula is: wherein, respectively represent the deviation values of the X-axis and the Y-axis, and is the linear regression coefficient for the X-axis and the Y-axis, respectively, is the die bonding pressure, is the die bonding temperature, is the die bonding time; The compensation relationship formula is: wherein, is a compensation pressure, is a compensation temperature, and is a compensation time; S322, output the calculated compensation parameters as the actual determined processing parameters.

5. The method of claim 1, wherein the method further comprises: The relative deviation value is calculated by comparing the position of the current grabbed die with the position of the last die, and the compensation value is calculated according to the deviation value, which includes the following steps: S51, pre-process the position image of the current captured die and the position image of the last captured die; S52, use image processing algorithm to locate the position of the die and determine the centroid position of the current grabbed die and the centroid position of the last die; S53, calculate the difference between the current die position and the last die position to obtain the relative deviation value, and calculate the corresponding compensation value using the obtained deviation value; S54, adjust the processing parameters using the calculated compensation value before the next die bonding operation.

6. The method of claim 1, wherein the method further comprises: After the processing is completed, the processing data is integrated, input into the database, and the historical data of the current category substrate processing in the database is replaced, which includes the following steps: S71, collect all related data in the processing process; S72, filter the collected raw data to remove invalid or abnormal data, and arrange the data into a unified format; S73, integrate all data into a complete data package, and connect the software with the database; S74, add the current processing data to the related table of the database, query the database, filter out the historical data associated with the current processing category substrate, and perform the delete operation.

Images