Semiconductor die bonder and control method thereof

Abstract

The invention relates to the technical field of semiconductor manufacturing, in particular to a semiconductor die bonder and a control method thereof. The die bonder comprises a swing arm, the swing arm is rotatably connected with a rack through a rotating shaft, the swing arm is driven by a first driving source, and the first driving source is controlled by a controller; the first base plate and the second base plate are arranged below the swing arm and can move front and back or left and right relative to the rotating shaft in the horizontal plane, movement of the first base plate and movement of the second base plate are controlled by a second driving source and a third driving source respectively, and the second driving source and the third driving source are both controlled by the controller. According to the semiconductor die bonder and the control method thereof provided by the invention, real-time monitoring and feedback are carried out through a camera, and a control strategy is combined, so that the position and the posture of a chip can be sensed in real time in the whole die bonding process, and flexible adjustment is carried out according to actual conditions, thereby ensuring accurate alignment of the chip on a target substrate.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor manufacturing, and in particular to a semiconductor die bonding machine and a control method thereof. Background Art

[0002] In semiconductor manufacturing, die bonding is a critical step in precisely attaching chips to substrates to form electronic components or integrated circuits. The rapid advancement of semiconductor technology, coupled with shrinking chip sizes and increasing integration density, presents unprecedented challenges to the precision and efficiency of die bonding. Traditional die bonding processes often rely on manual labor or simple mechanical devices, which are insufficient to meet the high-precision and high-efficiency demands of modern semiconductor manufacturing.

[0003] Manual operation has numerous drawbacks in the die bonding process. First, due to human physiological limitations, prolonged operation can lead to fatigue, resulting in decreased positioning accuracy. Second, manual operation is significantly influenced by subjective factors, and differences in operator skill levels can lead to inconsistent product quality. Furthermore, manual operation is inefficient, making it difficult to meet the demands of large-scale production.

[0004] While simple mechanical devices have improved production efficiency to a certain extent, they still lack flexibility when dealing with tiny chips, high-density integration, and complex substrate structures. These mechanical devices often lack real-time monitoring and feedback mechanisms, making it impossible to accurately sense the position and posture changes of the chip during the die bonding process, making it difficult to ensure accurate alignment of the chip on the target substrate.

[0005] For example, patent publication number CN111370350B, titled "Crystal Bonding Machine," includes a stand; a glue dispensing device; a glue dispensing shift mechanism; a feeding mechanism; a crystal bonding swing arm device; a crystal feeding platform; a crystal bonding shift mechanism; and a material receiving mechanism. The crystal bonding swing arm device includes a rotating frame, multiple crystal bonding swing arms, a lifter, and a crystal bonding motor, each crystal bonding swing arm being equipped with a suction nozzle; and the glue dispensing device includes multiple glue dispensing modules. The disadvantage is that while this crystal bonding machine improves crystal bonding efficiency to a certain extent, it lacks real-time monitoring and feedback devices such as cameras, and cannot sense the position and posture of the chip in real time. Therefore, it is insufficient in ensuring the accurate alignment of the chip on the target substrate.

[0006] Existing die bonders also have limitations in their control strategies. They often employ a single, fixed control strategy, unable to flexibly adjust to the characteristics of the chip and substrate, as well as variations in the actual production process. This makes it difficult to ensure accurate and stable bonding when encountering chips and substrates of varying sizes, shapes, or materials during the die bond process. Summary of the Invention

[0007] In response to the problem that the existing technology has deficiencies in ensuring the accurate alignment of chips on the target substrate, the present invention provides a semiconductor die bonding machine and a control method thereof. Through real-time monitoring and feedback by a camera, combined with a control strategy, the entire die bonding process can sense the position and posture of the chip in real time, and flexibly adjust according to actual conditions, thereby ensuring the accurate alignment of the chip on the target substrate.

[0008] To achieve the above technical objectives, the present invention provides a technical solution, which is a semiconductor die bonder, comprising: frame; a swing arm, the swing arm being rotatably connected to the frame via a rotating shaft, the swing arm being driven by a first driving source, and the first driving source being controlled by a controller; a first substrate and a second substrate, wherein the first substrate is used to carry a chip to be soldered, and the second substrate is used to receive and solder the chip, the first substrate and the second substrate are respectively disposed below the swing arm and are capable of moving forward and backward or left and right relative to the rotation axis in a horizontal plane, the movement of the first substrate and the second substrate are respectively controlled by a second driving source and a third driving source, and the second driving source and the third driving source are both controlled by a controller; A vacuum adsorption device, provided on the swing arm, for moving the chip from the first substrate to the second substrate; A solder spraying device, provided on the swing arm, includes a soldering head and is used to spray solder on the chip; a heating device, disposed on the second substrate, for locally heating the welding area; Cameras are provided above and below the first substrate and the second substrate, and are used to capture image information of the chip, the first substrate, and the second substrate. The cameras are connected to the controller to transmit the captured image information to the controller; The controller implements the following linkage control strategy: During the welding process, the camera captures the image information of the current welding point and obtains the image information of the next welding target area at the same time as the previous welding; Analyze the captured image information based on the image processing algorithm to determine the optimal welding strategy for the current welding point and the optimal welding point for the next welding target area; Based on the determined optimal welding strategy and optimal welding point, control instructions are generated, and the controller simultaneously controls the movement of the swing arm, the first substrate, and the second substrate, so that while the swing arm is welding the current chip to the second substrate, the first substrate moves to the position of the next chip to be welded, and the second substrate moves to the target position for the next welding.

[0009] In this technical solution, the swing arm is flexibly connected to the frame via a rotating shaft and precisely driven by a first drive source, achieving fast and stable movement. The first and second substrates are positioned below the swing arm, respectively, and can move freely in the horizontal plane. Controlled by independent second and third drive sources, they ensure rapid switching and precise positioning of the chip between different positions. A vacuum suction device and solder sprayer are strategically positioned on the swing arm, enabling rapid chip transfer and precise soldering. A heating device, located on the second substrate, provides localized heating of the soldering area, improving soldering quality and reliability. Cameras are positioned in all directions to capture real-time image information of the chip, the first substrate, and the second substrate, providing rich data support for the controller. The controller implements a sophisticated linkage control strategy, using image processing algorithms to rapidly analyze image information and determine the optimal soldering strategy and position. During the soldering process, the controller simultaneously controls the movement of the swing arm, the first substrate, and the second substrate, achieving synchronized and efficient soldering. While the swing arm is soldering the current chip to the second substrate, the first substrate is already moving to the next chip to be soldered, and the second substrate is also moving to the target position for the next soldering step. This significantly shortens the soldering cycle and improves production efficiency.

[0010] The present invention is further configured as follows: a hollow disk is provided on each of the first substrate and the second substrate, and the chip is arranged on the hollow disk.

[0011] In this technical solution, first, both the first and second substrates are equipped with hollow plates, allowing the chips to be arranged more orderly on the plates, facilitating subsequent transfer and soldering operations. This hollow structure not only reduces the weight of the substrates but also improves ventilation and heat dissipation for the chips, helping to maintain chip stability and reliability during the die bonding process. Secondly, a heating device is installed on the second substrate, providing more precise temperature control during the chip soldering process.

[0012] Another technical solution provided by the present invention is a control method for a semiconductor die bonder, comprising the following steps: S1, using the vacuum adsorption device on the swing arm to adsorb the chip on the first substrate, using a camera to perform appearance inspection on the chip to determine the optimal welding point of the chip; S2, controlling the swing arm to position the optimal soldering point of the chip downward, capturing a preliminary position image of the soldering point on the second substrate through a camera, and performing preliminary alignment; S3, analyzing the image data captured by the camera according to an image processing algorithm, calculating the deviation between the current position of the chip soldering point and the target soldering position of the second substrate, and generating a motion instruction to drive the first substrate and / or the second substrate to move until the soldering points are aligned; S4, activating the heating device on the second substrate to preheat the second substrate, and after confirming that the soldering points are aligned, controlling the solder spraying device to spray solder on the soldering points; S5, after the solder spraying is completed, the welding result is detected in real time by a camera to obtain the detection result; based on the detection result, the welding quality is predicted using a pre-trained machine learning model to obtain the prediction result information; S6. According to the prediction result information, the corresponding action instruction set is retrieved from a pre-built database, where the action instruction set includes a swing arm instruction, a first substrate action instruction, and a second substrate action instruction, and is used to adjust the position or state of the swing arm, the first substrate, and the second substrate to optimize the subsequent welding process or correct the deviation in the current welding.

[0013] In this technical solution, the chip is first suctioned to the chip using a vacuum suction device on the swing arm, and a camera is used to perform an appearance inspection. This ensures the integrity and accuracy of the chip before transfer and determines the optimal solder joint, laying the foundation for subsequent alignment and soldering operations. Next, the swing arm is controlled to position the chip with the optimal solder joint facing downward, and the camera captures an image of the preliminary position of the solder joint on the second substrate for preliminary alignment. This utilizes the camera's high-precision positioning capabilities to ensure the initial alignment accuracy between the chip and substrate. Then, an image processing algorithm analyzes the image data captured by the camera, calculates the deviation between the current position of the chip solder joint and the target solder joint position on the second substrate, and generates motion instructions to drive the first and / or second substrates for fine-tuning, achieving precise control of the soldering process and ensuring accurate solder joint alignment. During the soldering preparation phase, the heating device on the second substrate is activated to preheat the substrate, providing the appropriate temperature conditions for subsequent solder spraying. After confirming the solder joint alignment, the solder spraying device is used to spray the solder joint, completing the chip soldering process. After welding is complete, a camera monitors the welding results in real time. Based on these results, necessary adjustments are made to the swing arm, first substrate, and / or second substrate to ensure weld quality. This demonstrates the control method's feedback mechanism, ensuring the stability and reliability of the welding results through real-time monitoring and adjustments. Finally, a machine learning algorithm analyzes real-time and historical data from the welding process to predict and optimize the operating parameters of the solder spray device. This demonstrates the intelligent nature of the control method, continuously improving welding efficiency and quality through data analysis and algorithm optimization.

[0014] The present invention is further configured such that: the use of a camera to perform appearance inspection on the chip to determine the optimal welding point of the chip includes: Identify the electrodes and pads on the chip using image recognition methods, evaluate the flatness and cleanliness of the chip surface, and select areas with no surface defects and clear electrodes as potential welding points; Collect chip image data, train a solder joint model based on a machine learning algorithm, automatically identify the optimal solder joint area on the chip, and input the real-time chip image into the trained solder joint model. The solder joint model predicts the location of the optimal solder joint by comparing the input image with the features in the training data. Laser ranging is used to detect the three-dimensional topography of the chip surface, and a three-dimensional model of the chip surface is constructed and analyzed. By analyzing the surface undulations and defect information in the three-dimensional model, the flattest and most defect-free area on the chip surface is determined as the welding point. The temperature distribution of the chip after preheating is detected using a thermal imaging method. The temperature distribution in the thermal imaging image is analyzed using a temperature analysis algorithm, and the corresponding temperature distribution uniformity value is calculated. The area where the temperature distribution uniformity value is less than the preset threshold is selected as the welding point.

[0015] This technical solution first uses image recognition to identify key chip structures such as electrodes and pads. It also assesses physical properties such as the chip's surface flatness and cleanliness. Leveraging advanced image processing technology, this method quickly and accurately captures essential chip surface information, providing crucial insights for subsequent solder joint selection. Next, chip image data is collected and a machine learning algorithm is used to train a model. Leveraging the advantages of big data and artificial intelligence, this model automatically identifies optimal solder joint areas on the chip by learning from a large number of samples. During real-time operation, simply inputting captured chip images into the trained model allows for rapid prediction of optimal solder joint locations, significantly improving detection efficiency and accuracy. Furthermore, laser ranging is used to perform three-dimensional topography detection of the chip surface. A three-dimensional model of the chip surface is constructed and analyzed, providing a more intuitive visualization of the chip's surface topography and helping to identify the flattest, most defect-free areas on the chip surface as solder joints. This three-dimensional topography detection allows for more precise control of solder joint location and shape, improving solder joint quality. Finally, thermal imaging methods were used to detect the temperature distribution of the chip after preheating. The thermal characteristics of the chip during the preheating process were taken into consideration. By selecting areas with uniform temperature distribution and good heat dissipation performance as welding points, the thermal stress during the welding process can be further reduced, thereby improving the reliability and life of the chip.

[0016] The present invention is further configured as follows: in step S2, capturing a preliminary position image of the welding point on the second substrate by a camera, and performing preliminary alignment includes: According to the edge detection method, the edge of the welding point is identified, and the characteristic points of the welding point are extracted and matched with the preset standard characteristic points; Use a light source with a specific wavelength to illuminate the welding point, and adjust the brightness and irradiation angle of the light source according to actual needs; The captured images are processed in real time to identify the location of the welding points and compare them with the preset positions; According to the recognition result, the position of the first substrate and / or the second substrate is automatically adjusted.

[0017] In this technical solution, the edges of the weld are first identified using edge detection methods, accurately outlining the weld contour and providing a basis for subsequent feature point extraction and matching. Next, the weld feature points, such as shape, size, or color, are extracted and matched against pre-set standard feature points. Leveraging the accuracy of feature point matching, the weld is precisely aligned, improving welding accuracy. A light source with a specific wavelength is used to illuminate the weld, and the brightness and angle of the light source are adjusted as needed. This fully considers the impact of lighting conditions on image capture and feature point recognition. By optimizing light source parameters, clear and accurate weld images are captured, improving alignment reliability. Regarding image capture and processing, the captured image is processed in real time to identify the weld position and compare it with the pre-set position. This enables dynamic monitoring and real-time adjustment of the weld, ensuring stability and accuracy during the welding process. Finally, based on the recognition results, the position of the first and / or second substrates is automatically adjusted to achieve precise alignment of the weld. This demonstrates the automated and intelligent nature of the control method. Through real-time feedback and adjustments, a smooth welding process and high welding quality standards are ensured.

[0018] The present invention is further configured as follows: in step S3, analyzing the image data captured by the camera according to the image processing algorithm, calculating the deviation between the current position of the chip welding point and the target welding position of the second substrate, and generating a motion instruction to drive the first substrate and / or the second substrate to move includes: Extracting feature points of the welding point and the target welding position from the image according to an image processing algorithm, and matching the extracted feature points of the welding point with the feature points of the target welding position; According to the matched feature points, the precise position of the welding point in the current image is calculated by geometric transformation method; Comparing the calculated welding point position with the target welding position and calculating the deviation between the two; generating a motion instruction for driving the first substrate and / or the second substrate to move according to the calculated deviation; Images captured from multiple angles are extracted and matched based on deep learning algorithms, and overall weld point information is obtained through image fusion technology. During the movement process, the position changes of the welding points are analyzed in real time, and the movement instructions are dynamically adjusted according to the analysis results.

[0019] This technical solution first uses an image processing algorithm to extract and precisely match the characteristic points of the weld point and the target weld position from the image. This ensures accurate feature point extraction and efficient matching, laying a solid foundation for subsequent position calculation and deviation analysis. Next, a geometric transformation method is used to calculate the precise position of the weld point in the current image. This position is then compared with the target weld position to determine the deviation between the two. Through precise mathematical calculations, this enables quantitative analysis of the weld point position, providing a scientific basis for the generation of motion instructions. When generating motion instructions, parameters such as the movement range, speed, and acceleration of the first and second substrates are fully considered, ensuring smooth and accurate motion. This demonstrates the refinement and intelligence of the control method, allowing for flexible adjustment of motion parameters to accommodate different welding requirements. Furthermore, the description mentions the use of deep learning algorithms to extract and match features from images captured from multiple angles, and image fusion technology to obtain comprehensive weld point information. This leverages the powerful capabilities of deep learning in image processing to achieve comprehensive, multi-angle analysis of the weld point, improving welding accuracy and reliability. During the movement process, the position changes of the welding points are analyzed in real time, and the movement instructions are dynamically adjusted according to the analysis results, which reflects the real-time and dynamic nature of the control method. The movement strategy can be adjusted in time according to the actual situation during the welding process to ensure the smooth progress of the welding process.

[0020] The present invention is further configured as follows: in step S5, the real-time detection of the welding result by the camera and the adjustment of the swing arm, the first substrate and / or the second substrate according to the detection result include: Automatically detect welding results based on weld spot size, shape deviation, voids, and cracks to identify potential welding defects; Based on the image analysis results, the deviation between the actual position of the welding point and the ideal position, as well as the deviation between the welding quality index and the standard value, are calculated; According to the deviation calculation result, the position and posture of the swing arm, the first substrate and / or the second substrate are dynamically adjusted.

[0021] In this technical solution, the importance of real-time detection of welding results by camera not only ensures immediate feedback on welding quality, but also provides an accurate basis for subsequent adjustments. By automatically detecting potential defects such as weld size, shape deviation, voids and cracks in welding points, the system can quickly identify possible problems in the welding process, thereby avoiding the production of defective products. Based on the image analysis results, the deviation between the actual position and the ideal position of the welding point, as well as the deviation between the welding quality index and the standard value are calculated. The deviation calculation provides data support for subsequent dynamic adjustments, making the adjustment process more scientific and accurate. Based on the deviation calculation results, the system can dynamically adjust the position and posture of the swing arm, the first substrate and / or the second substrate to correct the deviation in the welding process in real time to ensure that the welding quality is optimized, which not only improves the accuracy and stability of welding, but also greatly improves the production efficiency and product qualification rate of semiconductor crystal bonding machines.

[0022] The present invention is further configured as follows: the calculation formula of the welding quality index is: ; in, and is the weight coefficient, and + =1, is the actual area of ​​the solder joint, Aideal is the ideal area of ​​the solder joint, h is the height of the solder joint, and the ideal height range of the solder joint is , Q is the welding quality index.

[0023] The present invention is further configured as follows: in step S6, according to the prediction result information, a corresponding action instruction set is retrieved from a pre-built database, wherein the action instruction set includes a swing arm instruction, a first substrate action instruction, and a second substrate action instruction, and is used to adjust the position or state of the swing arm, the first substrate, and the second substrate to optimize the subsequent welding process or correct the deviation in the current welding, including: The image, temperature and pressure data of the welding point during the welding process are collected in real time through cameras, temperature sensors and pressure sensors; Integrate historical data from past welding processes, including welding results under different working parameters, chip types, and substrate material data; Clean, normalize, and preprocess the collected real-time and historical data for feature extraction, and use the preprocessed historical data to train the deep neural network model; During the welding process, the real-time collected image, temperature, and pressure data are input into the trained deep neural network model; Based on the prediction results output by the model, it is determined whether the current welding process meets the requirements. If the prediction results show that the welding quality is poor or there is room for optimization of the working parameters, the parameters are adjusted. The working parameters of the solder spray device are dynamically adjusted through the optimization algorithm. The optimization algorithm calculates new parameter values ​​based on the prediction results and the current working parameters to optimize the welding quality expectations.

[0024] This technical solution achieves comprehensive monitoring and intelligent analysis of the welding process through the integrated application of sensor technology and machine learning algorithms. It not only collects multi-dimensional data (images, temperature, and pressure) of the welding point in real time, but also integrates historical welding data, including key information such as welding results under different operating parameters, chip type, and substrate material. Preprocessing provides a high-quality data foundation for subsequent machine learning model training. During the welding process, the trained deep neural network model can receive real-time data and quickly predict the current welding results and the effects of operating parameters. This enables the system to promptly identify potential welding problems and dynamically adjust the operating parameters of the solder spray device through optimization algorithms, thereby ensuring the stability and consistency of welding quality.

[0025] The present invention is further configured as follows: after step S6, the process further includes: after confirming that the welding quality meets the requirements, controlling the swing arm to release the welded chip from the second substrate and preparing for the welding operation of the next chip.

[0026] This technical solution emphasizes not only the automation and intelligence of the system but also its efficiency and continuity. The control method seamlessly connects each welding step, ensuring a smooth process while guaranteeing both welding quality and production efficiency. This coherent workflow design is crucial for improving the overall performance and reliability of semiconductor die bonders.

[0027] The beneficial effects of the present invention are as follows: (1) by real-time monitoring and feedback through the camera, combined with the control strategy, the entire die bonding process can sense the position and posture of the chip in real time, and flexibly adjust according to the actual situation, thereby ensuring the accurate alignment of the chip on the target substrate; (2) the chip is adsorbed by the vacuum adsorption device on the swing arm, and the appearance inspection is performed by the camera, thereby ensuring the integrity and accuracy of the chip before transfer, and at the same time determining the optimal welding point of the chip, laying the foundation for subsequent alignment and welding operations, controlling the swing arm to position the optimal welding point of the chip downward, and capturing the preliminary position image of the welding point on the second substrate through the camera for preliminary alignment, utilizing the high-precision positioning capability of the camera to ensure the preliminary alignment accuracy between the chip and the substrate, analyzing the image data captured by the camera according to the image processing algorithm, calculating the deviation between the current position of the chip welding point and the target welding position of the second substrate, and generating motion instructions to drive the first substrate and / or the second substrate for fine-tuning, thereby achieving precise control of the welding process and ensuring the accurate alignment of the welding point. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a structural schematic diagram of a semiconductor die bonding machine according to the present invention; Figure 2 The present invention is a flow chart of a control method for a semiconductor die bonder.

[0029] In the figure: 1. frame; 2. swing arm; 3. rotating shaft; 4. first substrate; 5. second substrate; 6. camera; 7. hollow disk. DETAILED DESCRIPTION

[0030] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] like Figure 1 As shown in the first embodiment of the present invention, a semiconductor die bonder includes: Rack 1; a swing arm 2, the swing arm 2 being rotatably connected to the frame 1 via a rotating shaft 3, the swing arm 2 being driven by a first driving source, which is controlled by a controller; a first base plate 4 and a second base plate 5, wherein the first base plate 4 is used to carry the chip to be soldered, and the second base plate 5 is used to receive and solder the chip. The first base plate 4 and the second base plate 5 are respectively arranged below the swing arm 2 and can move forward and backward or left and right relative to the rotating shaft 3 in the horizontal plane. The movement of the first base plate 4 and the second base plate 5 is controlled by a second driving source and a third driving source, respectively. The second driving source and the third driving source are both controlled by a controller; A vacuum adsorption device is provided on the swing arm 2 and is used to move the chip from the first substrate 4 to the second substrate 5; A solder spraying device, provided on the swing arm 2, includes a soldering head and is used to spray solder on the chip; A heating device, provided on the second substrate 5, for locally heating the welding area; Cameras 6 are provided above and below the first substrate 4 and the second substrate 5, and are used to capture image information of the chip, the first substrate 4 and the second substrate 5. The cameras 6 are connected to the controller to transmit the captured image information to the controller; The controller implements the following linkage control strategy: During the welding process, the camera 6 captures the image information of the current welding point and simultaneously obtains the image information of the target area of ​​the next welding at the same time as the previous welding; Analyze the captured image information based on the image processing algorithm to determine the optimal welding strategy for the current welding point and the optimal welding point for the next welding target area; Based on the determined optimal welding strategy and optimal welding point, control instructions are generated, and the controller simultaneously controls the movement of the swing arm 2, the first substrate 4, and the second substrate 5, so that while the swing arm 2 is welding the current chip to the second substrate 5, the first substrate 4 moves to the next chip position to be welded, and the second substrate 5 moves to the target position for the next welding.

[0032] In this embodiment, the swing arm 2 is flexibly connected to the frame 1 via a rotating shaft 3 and precisely driven by a first drive source, achieving rapid and stable movement. The first and second substrates 4 and 5 are respectively positioned below the swing arm 2 and can move freely in the horizontal plane. Controlled by independent second and third drive sources, they ensure rapid switching and precise positioning of the chip between different positions. A vacuum adsorption device and solder spraying device are cleverly positioned on the swing arm 2, enabling rapid chip transfer and precise soldering. A heating device is positioned on the second substrate 5 to locally heat the soldering area, improving soldering quality and reliability. The camera 6 is positioned in all directions, capable of capturing real-time image information of the chip, the first substrate 4 and the second substrate 5, providing rich data support for the controller. The controller implements a sophisticated linkage control strategy, rapidly analyzing image information through image processing algorithms to determine the optimal soldering strategy and position. During the soldering process, the controller can simultaneously control the movement of the swing arm 2, the first substrate 4 and the second substrate 5, achieving a synchronized and efficient soldering process. While the swing arm 2 is welding the current chip to the second substrate 5, the first substrate 4 has moved to the position of the next chip to be welded, and the second substrate 5 has also moved to the target position for the next welding, which greatly shortens the welding cycle and improves production efficiency.

[0033] As can be understood, the controller implements the following linkage control strategy to achieve the optimal control strategy: During the welding process, the camera captures the image information of the current welding point and obtains the image information of the next welding target area at the same time as the previous welding; Analyze the captured image information based on the image processing algorithm, calculate the deviation between the current position of the chip welding point and the target welding position, and determine the optimal alignment strategy for the current welding point; At the same time, based on machine learning algorithms, real-time and historical data of the welding process are analyzed to predict and optimize the operating parameters of the solder spray device, such as solder amount, spray speed, and heating temperature, to form an optimal welding parameter strategy; Based on the determined optimal alignment strategy and optimal welding parameter strategy, control instructions are generated, and the controller simultaneously controls the movement of the swing arm, the first substrate, and the second substrate. When the swing arm welds the current chip to the second substrate, the first substrate moves to the position of the next chip to be welded, and the second substrate moves to the target position for the next welding, so as to achieve synchronization, efficiency, and high quality of the welding process.

[0034] In one embodiment of the present invention, a hollow plate 7 is provided on each of the first and second substrates 4 and 5, and the chips are arranged on the hollow plate 7. Firstly, the hollow plate 7 on each of the first and second substrates 4 and 5 allows for a more orderly arrangement of the chips on the plate, facilitating subsequent transfer and soldering operations. This hollow structure not only reduces the weight of the substrates but also improves ventilation and heat dissipation for the chips, helping to maintain chip stability and reliability during the die bonding process.

[0035] It can be understood that the motor drives the rotating shaft 3 to drive the swing arm 2 to rotate between the first base plate 4 and the second base plate 5.

[0036] like Figure 2 As shown in the second embodiment of the present invention, a control method for a semiconductor die bonder includes the following steps: S1, using the vacuum adsorption device on the swing arm 2 to adsorb the chip on the first substrate 4, using the camera 6 to perform appearance inspection on the chip to determine the optimal welding point of the chip; S2, controlling the swing arm 2 to position the optimal soldering point of the chip downward, capturing a preliminary position image of the soldering point on the second substrate 5 through the camera 6, and performing preliminary alignment; S3, analyzing the image data captured by the camera 6 according to the image processing algorithm, calculating the deviation between the current position of the chip soldering point and the target soldering position of the second substrate 5, and generating a motion instruction to drive the first substrate 4 and / or the second substrate 5 to move until the soldering points are aligned; S4, activating the heating device on the second substrate 5 to preheat the second substrate 5, and after confirming that the soldering points are aligned, controlling the solder spraying device to spray solder on the soldering points; S5, after the solder spraying is completed, the welding result is detected in real time by the camera 6 to obtain the detection result; based on the detection result, the welding quality is predicted using the pre-trained machine learning model to obtain the prediction result information; S6. According to the prediction result information, the corresponding action instruction set is retrieved from the pre-built database, and the action instruction set includes the swing arm 2 instruction, the first substrate 4 action instruction and the second substrate 5 action instruction, which are used to adjust the position or state of the swing arm 2, the first substrate 4 and the second substrate 5 to optimize the subsequent welding process or correct the deviation in the current welding.

[0037] In this embodiment, the chip is first held by the vacuum suction device on the swing arm 2, and visual inspection is performed using the camera 6. This ensures the integrity and accuracy of the chip before transfer and determines the optimal solder joint, laying the foundation for subsequent alignment and soldering operations. Next, the swing arm 2 is controlled to position the chip with the optimal solder joint facing downward, and the camera 6 captures an image of the preliminary position of the solder joint on the second substrate 5 for preliminary alignment. This utilizes the high-precision positioning capability of the camera 6 to ensure the initial alignment accuracy between the chip and the substrate. Next, the image data captured by the camera 6 is analyzed using an image processing algorithm to calculate the deviation between the current position of the chip solder joint and the target solder position on the second substrate 5. This generates motion commands to drive the first substrate 4 and / or the second substrate 5 for fine-tuning, achieving precise control of the soldering process and ensuring accurate solder joint alignment. During the soldering preparation phase, the heating device on the second substrate 5 is activated to preheat the substrate, providing suitable temperature conditions for subsequent solder spraying. After confirming the alignment of the solder joints, the solder spraying device is used to spray the solder joints, completing the chip soldering process. After welding is complete, camera 6 monitors the welding results in real time. Based on these results, necessary adjustments are made to swing arm 2, first substrate 4, and / or second substrate 5 to ensure welding quality. This demonstrates the control method's feedback mechanism, ensuring the stability and reliability of the welding results through real-time monitoring and adjustment. Finally, a machine learning algorithm analyzes real-time and historical data from the welding process to predict and optimize the operating parameters of the solder spray device. This demonstrates the intelligent nature of the control method, continuously improving welding efficiency and quality through data analysis and algorithm optimization.

[0038] It can be understood that the generation of the pre-trained machine learning model includes: collecting a large amount of welding process data, including images of welding points and welding quality indicators; performing pre-processing operations on the data such as cleaning, labeling and normalization, learning and extracting key features from the training data according to the neural network algorithm to predict welding quality; adjusting the model parameters through the optimization algorithm to minimize the loss function so that the model can better fit the training data; evaluating the performance of the model on independent validation sets and test sets to ensure that it has good generalization ability; deploying the trained machine learning model to the control system of the semiconductor die bonder; in the actual production process, the welding result data obtained by the camera 6 will be input into the model, and the model will output the prediction result information to guide subsequent adjustment and optimization operations.

[0039] It can be understood that the database contains action instruction sets corresponding to various welding conditions. These action instruction sets are designed based on a large amount of historical welding data and experimental results to ensure the accuracy and stability of welding.

[0040] In one embodiment of the present invention, in step S1, the use of the camera 6 to perform appearance inspection on the chip to determine the optimal welding point of the chip includes: Identify the electrodes and pads on the chip using image recognition methods, evaluate the flatness and cleanliness of the chip surface, and select areas with no surface defects and clear electrodes as potential welding points; Collect chip image data, train a solder joint model based on a machine learning algorithm, automatically identify the optimal solder joint area on the chip, and input the real-time chip image into the trained solder joint model. The solder joint model predicts the location of the optimal solder joint by comparing the input image with the features in the training data. Laser ranging is used to detect the three-dimensional topography of the chip surface, and a three-dimensional model of the chip surface is constructed and analyzed. By analyzing the surface undulations and defect information in the three-dimensional model, the flattest and most defect-free area on the chip surface is determined as the welding point. The temperature distribution of the chip after preheating is detected using a thermal imaging method. The temperature distribution in the thermal imaging image is analyzed using a temperature analysis algorithm, and the corresponding temperature distribution uniformity value is calculated. The area where the temperature distribution uniformity value is less than the preset threshold is selected as the welding point.

[0041] This technical solution first uses image recognition to identify key chip structures such as electrodes and pads. It also assesses physical properties such as the chip's surface flatness and cleanliness. Leveraging advanced image processing technology, this method quickly and accurately captures essential chip surface information, providing crucial insights for subsequent solder joint selection. Next, chip image data is collected and a machine learning algorithm is used to train a model. Leveraging the advantages of big data and artificial intelligence, this model automatically identifies optimal solder joint areas on the chip by learning from a large number of samples. During real-time operation, simply inputting captured chip images into the trained model allows for rapid prediction of optimal solder joint locations, significantly improving detection efficiency and accuracy. Furthermore, laser ranging is used to perform three-dimensional topography detection of the chip surface. A three-dimensional model of the chip surface is constructed and analyzed, providing a more intuitive visualization of the chip's surface topography and helping to identify the flattest, most defect-free areas on the chip surface as solder joints. This three-dimensional topography detection allows for more precise control of solder joint location and shape, improving solder joint quality. Finally, thermal imaging methods were used to detect the temperature distribution of the chip after preheating. The thermal characteristics of the chip during the preheating process were taken into consideration. By selecting areas with uniform temperature distribution and good heat dissipation performance as welding points, the thermal stress during the welding process can be further reduced, thereby improving the reliability and life of the chip.

[0042] It can be understood that the welding point model can automatically identify the best welding area on the chip by learning features (such as electrode shape, pad position, surface defects, etc.) in a large amount of chip image data.

[0043] It can be understood that the image recognition method is a feature-based image recognition method, which performs recognition by extracting key features (such as edges, corners, and textures) in the image and uses an edge detection method to identify the edges of the welding points.

[0044] Furthermore, in step S2, capturing a preliminary position image of the welding point on the second substrate 5 by the camera 6 and performing preliminary alignment includes: According to the edge detection method, the edge of the welding point is identified, and the characteristic points of the welding point are extracted and matched with the preset standard characteristic points; Use a light source with a specific wavelength to illuminate the welding point, and adjust the brightness and irradiation angle of the light source according to actual needs; The captured images are processed in real time to identify the location of the welding points and compare them with the preset positions; According to the recognition result, the position of the first substrate 4 and / or the second substrate 5 is automatically adjusted.

[0045] In this technical solution, the edges of the weld are first identified using edge detection methods, accurately outlining the weld contour and providing a basis for subsequent feature point extraction and matching. Next, the weld feature points, such as shape, size, or color, are extracted and matched with pre-set standard feature points. Leveraging the accuracy of feature point matching, the weld is precisely aligned, improving welding accuracy. A light source with a specific wavelength is used to illuminate the weld, and the brightness and angle of the light source are adjusted as needed. This fully considers the impact of lighting conditions on image capture and feature point recognition. By optimizing light source parameters, clear and accurate weld point images are captured, improving alignment reliability. Regarding image capture and processing, the captured image is processed in real time to identify the weld position and compare it with the pre-set position. This enables dynamic monitoring and real-time adjustment of the weld, ensuring stability and accuracy during the welding process. Finally, based on the recognition results, the position of the first substrate 4 and / or the second substrate 5 is automatically adjusted to achieve precise alignment of the weld. This demonstrates the automated and intelligent nature of the control method. Through real-time feedback and adjustments, a smooth welding process and high welding quality standards are ensured.

[0046] It can be understood that the characteristic points of the welding point are shape, size or color.

[0047] It can be understood that the light source of a specific wavelength refers to a laser light source in the near-infrared band (such as 800-1100nm), because lasers of these wavelengths can be better absorbed by metal materials, improving welding efficiency and quality, and tin-based solder has good absorption capacity in the near-infrared band. Lasers of these wavelengths can effectively heat and melt tin-based solder to achieve good welding effects.

[0048] In one embodiment of the present invention, in step S3, analyzing the image data captured by the camera 6 according to the image processing algorithm, calculating the deviation between the current position of the chip soldering point and the target soldering position of the second substrate 5, and generating a motion instruction to drive the first substrate 4 and / or the second substrate 5 to move includes: Extracting feature points of the welding point and the target welding position from the image according to an image processing algorithm, and matching the extracted feature points of the welding point with the feature points of the target welding position; According to the matched feature points, the precise position of the welding point in the current image is calculated by geometric transformation method; Comparing the calculated welding point position with the target welding position and calculating the deviation between the two; Generate a motion instruction to drive the first substrate 4 and / or the second substrate 5 to move according to the calculated deviation; Images captured from multiple angles are extracted and matched based on deep learning algorithms, and overall weld point information is obtained through image fusion technology. During the movement process, the position changes of the welding points are analyzed in real time, and the movement instructions are dynamically adjusted according to the analysis results.

[0049] In this technical solution, an image processing algorithm is first used to extract the characteristic points of the weld point and the target weld position from the image and perform precise matching. This ensures accurate feature point extraction and efficient matching, laying a solid foundation for subsequent position calculation and deviation analysis. Next, a geometric transformation method is used to calculate the precise position of the weld point in the current image. This position is compared with the target weld position to determine the deviation between the two. Through precise mathematical calculations, a quantitative analysis of the weld point position is achieved, providing a scientific basis for the generation of motion instructions. When generating motion instructions, parameters such as the movement range, speed, and acceleration of the first and second substrates 4 and 5 are fully considered, ensuring smooth and accurate motion. This demonstrates the refinement and intelligence of the control method, allowing for flexible adjustment of motion parameters to accommodate different welding requirements. Furthermore, the description mentions the use of deep learning algorithms to extract and match features from images captured from multiple angles, and image fusion technology to obtain comprehensive weld point information. This leverages the powerful capabilities of deep learning in image processing to achieve comprehensive, multi-angle analysis of weld points, improving welding accuracy and reliability. During the movement process, the position changes of the welding points are analyzed in real time, and the movement instructions are dynamically adjusted according to the analysis results, which reflects the real-time and dynamic nature of the control method. The movement strategy can be adjusted in time according to the actual situation during the welding process to ensure the smooth progress of the welding process.

[0050] It can be understood that the image processing algorithms are SIFT, SURF, and ORB.

[0051] It can be understood that the deviation can be expressed as a coordinate difference in a two-dimensional plane.

[0052] It can be understood that the motion instructions take into account the movement range, speed and acceleration parameters of the first substrate 4 and the second substrate 5 .

[0053] It can be understood that in image processing, feature points refer to pixels in the image with significant features (such as corner points, edge points, texture change points, etc.). These feature points are unique and stable in the image and can be used for image matching and positioning. These significant feature points are extracted from the image, usually using specific algorithms (such as SIFT, SURF, ORB, etc.).

[0054] It is understood that in image processing, geometric transformation refers to the process of converting an image from one coordinate system to another. Common geometric transformation methods include translation, rotation, scaling, affine transformation, and perspective transformation. In the control method of a semiconductor die bonder, the geometric transformation method is used to calculate the precise position of the weld point in the current image based on the matched feature points. For example, if the coordinates of the target weld position in the image and the relative positional relationship between the weld point feature point and the target feature point are known (such as described by an affine transformation), the precise coordinates of the weld point in the current image can be calculated through geometric transformation.

[0055] As you can understand, deep learning algorithms can be used in semiconductor die bonder control methods to extract solder joint features from images captured from multiple angles. Information from these multiple images can be fused to obtain more comprehensive and accurate information. Image fusion technology can be used in semiconductor die bonder control methods to obtain comprehensive solder joint information, improving soldering accuracy and reliability.

[0056] In step S5, the real-time detection of the welding result by the camera 6 and the adjustment of the swing arm 2, the first substrate 4 and / or the second substrate 5 according to the detection result include: Automatically detect welding results based on weld spot size, shape deviation, voids, and cracks to identify potential welding defects; Based on the image analysis results, the deviation between the actual position of the welding point and the ideal position, as well as the deviation between the welding quality index and the standard value, are calculated; According to the deviation calculation result, the position and posture of the swing arm 2, the first substrate 4 and / or the second substrate 5 are dynamically adjusted.

[0057] In this technical solution, the importance of real-time detection of welding results by camera 6 not only ensures immediate feedback on welding quality, but also provides an accurate basis for subsequent adjustments. By automatically detecting potential defects such as weld size, shape deviation, voids and cracks in welding points, the system can quickly identify possible problems in the welding process, thereby avoiding the production of defective products. The deviation between the actual position and the ideal position of the welding point, as well as the deviation between the welding quality index and the standard value are calculated based on the image analysis results. The deviation calculation provides data support for subsequent dynamic adjustments, making the adjustment process more scientific and accurate. Based on the deviation calculation results, the system can dynamically adjust the position and posture of the swing arm 2, the first substrate 4 and / or the second substrate 5 to correct the deviation in the welding process in real time, ensuring that the welding quality is optimized, which not only improves the accuracy and stability of welding, but also greatly improves the production efficiency and product qualification rate of semiconductor crystal bonding machines.

[0058] It can be understood that the welding quality indicators include welding spot area, height, and uniformity indicators.

[0059] It can be understood that the calculation formula of the welding quality index is: ; in, and is the weight coefficient, and + =1, is the actual area of ​​the solder joint, Aideal is the ideal area of ​​the solder joint, h is the height of the solder joint, and the ideal height range of the solder joint is , Q is the welding quality index.

[0060] In this formula, the welding quality index Q takes into account the area and height of the weld and is used to evaluate the welding quality.

[0061] In step S6, based on the prediction result information, a corresponding action instruction set is retrieved from a pre-built database. The action instruction set includes a swing arm instruction, a first substrate action instruction, and a second substrate action instruction. The action instruction set is used to adjust the position or state of the swing arm, the first substrate, and the second substrate to optimize the subsequent welding process or correct the deviation in the current welding process. The image, temperature and pressure data of the welding point during the welding process are collected in real time by the camera 6, the temperature sensor and the pressure sensor; Integrate historical data from past welding processes, including welding results under different working parameters, chip types, and substrate material data; Clean, normalize, and preprocess the collected real-time and historical data for feature extraction, and use the preprocessed historical data to train the deep neural network model; During the welding process, the real-time collected image, temperature, and pressure data are input into the trained deep neural network model; Based on the prediction results output by the model, it is determined whether the current welding process meets the requirements. If the prediction results show that the welding quality is poor or there is room for optimization of the working parameters, the parameters are adjusted. The working parameters of the solder spray device are dynamically adjusted through the optimization algorithm. The optimization algorithm calculates new parameter values ​​based on the prediction results and the current working parameters to optimize the welding quality expectations.

[0062] In this embodiment, comprehensive monitoring and intelligent analysis of the welding process are achieved through the integrated application of sensor technology and machine learning algorithms. This not only collects multi-dimensional data (images, temperature, and pressure) of the weld point in real time, but also integrates historical welding data, including key information such as welding results under different operating parameters, chip type, and substrate material. Preprocessing provides a high-quality data foundation for subsequent machine learning model training. During the welding process, the trained deep neural network model can receive real-time data and quickly predict the current welding results and the effects of operating parameters. This enables the system to promptly identify potential welding issues and dynamically adjust the operating parameters of the solder spray device through optimization algorithms, thereby ensuring stable and consistent welding quality.

[0063] It can be understood that the relationship between the prediction results and the parameters is established through the mapping relationship learned by the model during the training process, and the optimization process will continue to iterate until the best parameter combination is found.

[0064] It can be understood that the optimization algorithm is a gradient descent algorithm and a genetic algorithm.

[0065] After step S6, the process also includes controlling the swing arm 2 to release the bonded chip from the second substrate 5 after confirming that the welding quality meets the requirements, preparing for the welding of the next chip. This system not only emphasizes its automation and intelligence, but also highlights its efficiency and continuity. The control method seamlessly connects each welding step, ensuring a smooth welding process while guaranteeing welding quality and production efficiency. A coherent workflow design is crucial for improving the overall performance and reliability of semiconductor die bonders.

[0066] The specific description of the present invention in the above embodiments is only used to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Technical engineers in this field may make some non-essential improvements and adjustments to the present invention based on the contents of the above invention, which fall within the scope of protection of the present invention.

Claims

1. A semiconductor die bonder, characterized in that: include: frame; a swing arm, the swing arm being rotatably connected to the frame via a rotating shaft, the swing arm being driven by a first driving source, and the first driving source being controlled by a controller; a first substrate and a second substrate, wherein the first substrate is used to carry a chip to be soldered, and the second substrate is used to receive and solder the chip, the first substrate and the second substrate are respectively disposed below the swing arm and are capable of moving forward and backward or left and right relative to the rotation axis in a horizontal plane, the movement of the first substrate and the second substrate are respectively controlled by a second driving source and a third driving source, and the second driving source and the third driving source are both controlled by a controller; A vacuum adsorption device, provided on the swing arm, for moving the chip from the first substrate to the second substrate; A solder spraying device, provided on the swing arm, includes a soldering head and is used to spray solder on the chip; a heating device, disposed on the second substrate, for locally heating the welding area; Cameras are provided above and below the first substrate and the second substrate, and are used to capture image information of the chip, the first substrate and the second substrate. The cameras are connected to the controller to transmit the captured image information to the controller; The controller implements the following linkage control strategy: During the welding process, the camera captures the image information of the current welding point and obtains the image information of the next welding target area at the same time as the previous welding; Analyze the captured image information based on the image processing algorithm to determine the optimal welding strategy for the current welding point and the optimal welding position for the next welding target area; Based on the determined optimal welding strategy and optimal welding position, control instructions are generated, and the controller simultaneously controls the movement of the swing arm, the first substrate, and the second substrate, so that while the swing arm is welding the current chip to the second substrate, the first substrate moves to the position of the next chip to be welded, and the second substrate moves to the target position for the next welding.

2. A semiconductor die bonder according to claim 1, characterized in that: A hollow disk is provided on each of the first substrate and the second substrate, and the chip is arranged on the hollow disk.

3. A control method for a semiconductor die bonder, characterized in that: A semiconductor die bonding machine according to claim 1 or 2, comprising the following steps: S1, using the vacuum adsorption device on the swing arm to adsorb the chip on the first substrate, using a camera to perform appearance inspection on the chip to determine the optimal welding point of the chip; S2, controlling the swing arm to position the optimal soldering point of the chip downward, capturing a preliminary position image of the soldering point on the second substrate through a camera, and performing preliminary alignment; S3, analyzing the image data captured by the camera according to an image processing algorithm, calculating the deviation between the current position of the chip soldering point and the target soldering position of the second substrate, and generating a motion instruction to drive the first substrate and / or the second substrate to move until the soldering points are aligned; S4, activating the heating device on the second substrate to preheat the second substrate, and after confirming that the soldering points are aligned, controlling the solder spraying device to spray solder on the soldering points; S5, after the solder spraying is completed, the welding result is detected in real time by a camera to obtain the detection result; based on the detection result, the welding quality is predicted using a pre-trained machine learning model to obtain the prediction result information; S6. According to the prediction result information, the corresponding action instruction set is retrieved from a pre-built database, where the action instruction set includes a swing arm instruction, a first substrate action instruction, and a second substrate action instruction, and is used to adjust the position or state of the swing arm, the first substrate, and the second substrate to optimize the subsequent welding process or correct the deviation in the current welding.

4. The control method of a semiconductor die bonder according to claim 3, characterized in that: In step S1, the use of a camera to perform appearance inspection on the chip to determine the optimal welding point of the chip includes: Identify the electrodes and pads on the chip using image recognition methods, evaluate the flatness and cleanliness of the chip surface, and select areas with no surface defects and clear electrodes as potential welding points; Collect chip image data, train a solder joint model based on a machine learning algorithm, automatically identify the optimal solder joint area on the chip, and input the real-time chip image into the trained solder joint model. The solder joint model predicts the location of the optimal solder joint by comparing the input image with the features in the training data. Laser ranging is used to detect the three-dimensional topography of the chip surface, and a three-dimensional model of the chip surface is constructed and analyzed. By analyzing the surface undulations and defect information in the three-dimensional model, the flattest and most defect-free area on the chip surface is determined as the welding point. The temperature distribution of the chip after preheating is detected using a thermal imaging method. The temperature distribution in the thermal imaging image is analyzed using a temperature analysis algorithm, and the corresponding temperature distribution uniformity value is calculated. The area where the temperature distribution uniformity value is less than the preset threshold is selected as the welding point.

5. The control method of a semiconductor die bonder according to claim 3, wherein: In step S2, capturing a preliminary position image of the welding point on the second substrate by a camera and performing preliminary alignment includes: According to the edge detection method, the edge of the welding point is identified, and the characteristic points of the welding point are extracted and matched with the preset standard characteristic points; Use a light source with a specific wavelength to illuminate the welding point, and adjust the brightness and irradiation angle of the light source according to actual needs; The captured images are processed in real time to identify the location of the welding points and compare them with the preset positions; According to the recognition result, the position of the first substrate and / or the second substrate is automatically adjusted.

6. The control method of a semiconductor die bonder according to claim 3, characterized in that: In step S3, analyzing the image data captured by the camera according to the image processing algorithm, calculating the deviation between the current position of the chip soldering point and the target soldering position of the second substrate, and generating a motion instruction to drive the first substrate and / or the second substrate to move includes: Extracting feature points of the welding point and the target welding position from the image according to an image processing algorithm, and matching the extracted feature points of the welding point with the feature points of the target welding position; According to the matched feature points, the precise position of the welding point in the current image is calculated by geometric transformation method; Comparing the calculated welding point position with the target welding position and calculating the deviation between the two; generating a motion instruction for driving the first substrate and / or the second substrate to move according to the calculated deviation; Images captured from multiple angles are extracted and matched based on deep learning algorithms, and overall weld point information is obtained through image fusion technology. During the movement process, the position changes of the welding points are analyzed in real time, and the movement instructions are dynamically adjusted according to the analysis results.

7. The control method of a semiconductor die bonder according to claim 3, characterized in that: In step S5, the real-time detection of the welding result by the camera and the adjustment of the swing arm, the first substrate and / or the second substrate according to the detection result include: Automatically detect welding results based on weld spot size, shape deviation, voids, and cracks to identify potential welding defects; Based on the image analysis results, the deviation between the actual position of the welding point and the ideal position, as well as the deviation between the welding quality index and the standard value, are calculated; According to the deviation calculation result, the position and posture of the swing arm, the first substrate and / or the second substrate are dynamically adjusted.

8. The control method of a semiconductor die bonder according to claim 7, characterized in that: The calculation formula of the welding quality index is: ； in, and is the weight coefficient, and + =1, is the actual area of ​​the solder joint, Aideal is the ideal area of ​​the solder joint, h is the height of the solder joint, and the ideal height range of the solder joint is , Q is the welding quality index.

9. The control method of a semiconductor die bonder according to claim 3, wherein: In step S6, based on the prediction result information, a corresponding action instruction set is retrieved from a pre-built database. The action instruction set includes a swing arm instruction, a first substrate action instruction, and a second substrate action instruction. The action instruction set is used to adjust the position or state of the swing arm, the first substrate, and the second substrate to optimize the subsequent welding process or correct the deviation in the current welding process. The image, temperature and pressure data of the welding point during the welding process are collected in real time through cameras, temperature sensors and pressure sensors; Integrate historical data from past welding processes, including welding results under different working parameters, chip types, and substrate material data; Clean, normalize, and preprocess the collected real-time and historical data for feature extraction, and use the preprocessed historical data to train the deep neural network model; During the welding process, the real-time collected image, temperature, and pressure data are input into the trained deep neural network model; Based on the prediction results output by the model, it is determined whether the current welding process meets the requirements. If the prediction results show that the welding quality is poor or there is room for optimization of the working parameters, the parameters are adjusted. The working parameters of the solder spray device are dynamically adjusted through the optimization algorithm. The optimization algorithm calculates new parameter values ​​based on the prediction results and the current working parameters to optimize the welding quality expectations.

10. The control method of a semiconductor die bonder according to any one of claims 3 to 9, characterized in that: After step S6, the method further includes: after confirming that the welding quality meets the requirements, controlling the swing arm to release the welded chip from the second substrate and preparing for the welding operation of the next chip.

Images