Running control system of uniform glue baking machine

Abstract

The invention relates to the technical field of equipment operation control, and discloses an operation control system of a spin coating roaster. The system comprises an initialization module used for carrying out parameter initialization on the uniform glue baking machine and generating equipment initial state data; the time sequence control module is used for performing time sequence control on an upper shovel hand servo mechanism and a lower shovel hand servo mechanism of the uniform glue baking machine to obtain wafer position data and adsorption state data; the rotating speed grading control module is used for carrying out five-section type rotating speed grading control on rotating servo equipment of the spin-coating baking machine to obtain rotating speed control data; the gluing control module is used for carrying out gluing control on gluing equipment and generating gluing parameter data; and the compensation control module is used for performing inter-region heat coupling compensation control on the four-region heating units of the spin-coating baking machine and outputting a temperature uniformity control result. According to the invention, the gluing positioning precision control is realized, the consistency of the gluing amount is improved, and the variation between batches of products is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of equipment operation control, and in particular to an operation control system of a glue-splitting and baking machine. Background Art

[0002] In the semiconductor manufacturing process, the coating baking process is a key link in wafer processing, and its quality directly affects the yield and performance of the chip. Traditional coating baking machines have many technical bottlenecks in operation control: the servo system parameters are fixed and it is difficult to adapt to environmental changes and process differences; the speed control accuracy is low, and the speed fluctuates greatly during the wafer rotation, resulting in uneven coating; the position accuracy of the scraper mechanism and the response lag of the vacuum adsorption control system increase the risk of wafer damage; the heat cross-effect between the heating areas is serious, and the problem of uneven temperature is difficult to solve. These problems are particularly prominent in high-precision semiconductor manufacturing processes.

[0003] As chip manufacturing processes continue to shrink, the control accuracy requirements for the coating baking process are getting higher and higher. In the existing technology, the single PID control method cannot effectively cope with system parameter changes and external interference. There is a lack of coordination mechanism between the control units, making it difficult to achieve overall optimal control. Especially in terms of coating flow control and heat management, due to the lack of dynamic compensation mechanism, the process consistency is poor and the batch-to-batch variation is large, which seriously restricts the stability of semiconductor manufacturing and product quality. Summary of the invention

[0004] The invention provides an operation control system for a glue spreading and baking machine, which realizes the control of the glue coating positioning accuracy, improves the consistency of the glue coating amount, and reduces the variation between product batches.

[0005] In a first aspect, the present invention provides an operation control system for a glue-spreading and baking machine, the operation control system for the glue-spreading and baking machine comprising: The initialization module is used to detect and initialize the parameters of the servo motor and stepper motor of the glue baking machine and generate the initial state data of the equipment; A timing control module is used to perform position accuracy control and vacuum adsorption timing control on the upper shovel hand servo mechanism and the lower shovel hand servo mechanism of the coating baking machine according to the initial state data of the equipment, and obtain wafer position data and adsorption state data; A rotation speed classification control module, used for performing five-stage rotation speed classification control on the rotation servo device of the coating baking machine based on the wafer position data and the adsorption state data to obtain rotation speed control data; A glue coating control module, used to use the speed control data to control the glue coating equipment to perform fixed-point glue coating or reciprocating glue coating, adjust the position of the glue coating stepping motor and the opening and closing time of the glue valve, and generate glue coating parameter data; The compensation control module is used to perform inter-region heat coupling compensation control on the four-region heating unit of the glue coating baking machine according to the glue coating parameter data, and output a temperature uniformity control result.

[0006] Optionally, in a first implementation of the first aspect of the present invention, the initialization module is specifically used to: Perform status detection on the upper shovel hand servo motor, lower shovel hand servo motor, lifting Z-axis servo motor and rotating axis servo motor of the glue spreading and baking machine to obtain the initial status information of the motor; According to the motor initial state information, each servo motor is respectively subjected to an origin return operation to obtain motor zero position reference data; Based on the motor zero position reference data, the operating parameters of each servo motor are collected to obtain a motor operating parameter set including the speed before the near point DOG photoelectric, the speed after the near point DOG photoelectric, the automatic high speed, the automatic low speed, the acceleration time and the deceleration time during origin return; Performing pressure value detection on the air source system and the glue barrel pressure system of the glue spreading and baking machine to obtain air source pressure data and glue barrel pressure data; The initial temperature of the glue baking machine is collected, and the difference is calculated with the set temperature to generate initial temperature deviation data; The device initial state data is generated based on the motor operation parameter set, the air source pressure data, the glue barrel pressure data and the temperature initial deviation data.

[0007] Optionally, in a second implementation of the first aspect of the present invention, the timing control module is specifically used to: Detecting the wafer receiving and lifting mechanism of the coating baking machine to obtain the wafer boat layer status information and the wafer initial position information; According to the wafer boat layer status information and the wafer initial position information, upper and lower shovel hand coordinated control analysis is performed on the initial state data of the equipment to obtain a dual-axis linkage motion control instruction; Based on the dual-axis linkage motion control instruction, the upper shovel hand servo mechanism is moved to the standby position, the lower shovel hand servo mechanism is moved to the wafer boat material taking position, and the real-time trajectory deviation data is monitored; According to the real-time trajectory deviation data, the motion trajectories of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism are compensated in real time to obtain wafer position data; For the vacuum adsorption system of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism, the adsorption and release processes are controlled according to the preset vacuum adsorption time and vacuum breaking time to obtain adsorption state data.

[0008] Optionally, in a third implementation of the first aspect of the present invention, the speed classification control module is specifically used to: Read the current process recipe parameters from the data area of ​​the programmable logic controller according to the wafer position data and the adsorption state data, and obtain the speed control parameters including the starting speed of the coating, the first speed value, the second speed value, the third speed value, the fourth speed value, the fifth speed value and the duration of each section; Inputting the speed control parameters into an extended state observer model, observing the unmodeled dynamics of the system and external disturbances as extended states, and calculating a speed state estimation value and a system disturbance estimation value; The rotation axis servo motor control signal is calculated using the speed state estimation value and the system interference estimation value, and a motor drive instruction is output through a linear state feedback control strategy, so that the rotation axis servo motor runs according to the set five-stage speed requirement, and real-time speed measurement data is obtained; Calculating the difference between the real-time speed measurement data and the target speed value in the speed control parameter to obtain real-time speed deviation data; The dynamic compensation control amount is calculated according to the real-time speed deviation data and the system interference estimation value, and the acceleration curve is dynamically calculated according to the difference between the current speed and the target speed to obtain the speed control data including the actual operating values ​​of five speed sections, the duration of each section and the speed stability data.

[0009] Optionally, in a fourth implementation of the first aspect of the present invention, the gluing control module is specifically used to: Determine a fixed-point gluing mode or a reciprocating gluing mode according to the speed control data and generate a corresponding gluing mode control signal; Based on the glue coating mode control signal, the glue coating stepping motor is controlled to move to a preset glue coating starting position, and the actual position deviation between the wafer center and the glue coating valve nozzle is calculated through the position sensor feedback signal to obtain the position fine-tuning parameter; The position fine-tuning parameters are used to perform position compensation control on the glue coating stepping motor to obtain the target glue coating position control result, and at the same time, the glue barrel pressure system is monitored in real time to generate pressure control data; Based on the glue coating mode control signal and the target glue coating position control result, the glue valve opening time is accurately controlled in the fixed-point glue coating mode, or the glue coating stepping motor is controlled to reciprocate between the glue coating start position and the glue coating end position in the reciprocating glue coating mode, and the glue valve is controlled to open and close according to a preset number of times, so as to obtain the glue valve action data; The pressure control data and the glue valve action data are correlated and analyzed to generate glue coating parameter data including glue coating position, glue valve opening time, glue coating pressure change and glue coating uniformity evaluation index.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the compensation control module further includes: An extraction unit, used to extract process type information from the glue coating parameter data, and read the temperature setting value and heating time requirement according to the process type information to generate four-zone temperature control target parameters; A construction unit is used to construct a temperature closed-loop control model based on the temperature control target parameters of the four zones, and calculate the temperature deviation value, deviation integral value and deviation change rate of each temperature zone respectively to obtain a temperature closed-loop basic control amount; A division unit, used for dividing the working range into a low temperature section, a medium temperature section and a high temperature section according to the temperature closed-loop basic control quantity, and generating segmented temperature control instructions corresponding to different temperature sections; The heat interference compensation unit is used to model the heat transfer characteristics of the temperature intervals of the four-zone heating units and to compensate for the heat interference, so as to obtain the inter-zone heat coupling compensation data; An advance compensation unit, used to combine the inter-region heat coupling compensation data, to compensate in advance for the system deviation caused by the change in external ambient temperature, and to generate a comprehensive temperature compensation control instruction; The real-time monitoring unit is used to synthesize the segmented temperature control instruction and the comprehensive temperature compensation control instruction and output them to the solid-state relay control unit, and at the same time monitor the heating power of each temperature interval in real time to obtain the temperature uniformity control result.

[0011] Optionally, in a sixth implementation of the first aspect of the present invention, the heat interference compensation unit is specifically used to: Sequentially apply standard heating pulses to the four-zone heating units of the first zone on the first side, the second zone on the first side, the first zone on the second side, and the second zone on the second side in the coating baking machine, and record the temperature response curves of the other three zones to generate initial test data of heat transfer; Calculating the heat transfer coefficient between the regions based on the initial test data of heat transfer to obtain a transfer coefficient matrix, and performing structural processing on the transfer coefficient matrix to obtain a heat coupling mathematical model; According to the thermal coupling mathematical model, the deviation data of the current temperature of the four zones and the target temperature collected in real time are analyzed, and the heat interference amount from the other three temperature zones is calculated for each temperature zone to obtain the quantitative data of the cross-effect of the temperature zones; Based on the quantitative data of the cross-effect of the temperature intervals, the reverse compensation principle is applied to adjust the heating control output of each temperature zone in a partitioned manner, and a pre-compensation control instruction is calculated; The pre-compensation control instruction is compared and analyzed with the actual temperature change, and the parameter values ​​in the 4×4 coupling matrix are optimized by the least square method to generate inter-regional thermal coupling compensation data.

[0012] In the technical solution provided by the present invention, an extended state observer and dynamic parameter adjustment technology are used to achieve precise control of the rotation axis speed, and the speed stability is improved from the traditional ±10rpm to ±1rpm, and the uniformity of glue spreading is significantly improved. By constructing a collaborative control mechanism between the functional units, the upper and lower shovels, the rotation system, the glue coating system and the heating system are fed back in a closed loop, and the overall system operation is more coordinated, which effectively solves the problem of incoordination caused by the independent control of each unit. The 4×4 thermal coupling matrix model is introduced to accurately describe the heat transfer relationship between the heating areas. Through the inter-area thermal coupling compensation control, the temperature control accuracy is improved from ±2℃ to ±0.5℃, which significantly improves the overall heating uniformity of the wafer. The adaptive flow control algorithm combined with pressure feedback realizes the control of the glue coating positioning accuracy within the range of ±0.1mm, greatly improves the consistency of the glue coating amount, and reduces the variation between product batches. Based on the abnormal detection and processing mechanism of multi-sensor fusion, the system can monitor the operating status in real time, automatically identify and handle various abnormal situations, prevent equipment damage and product scrapping, and ensure production safety. By combining human-machine interface with automatic control, operators can achieve complex process control through simple parameter settings and process selection, reducing the dependence on operator skills. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0014] Figure 1 It is a schematic diagram of an embodiment of the control system of the glue-spinning and baking machine in the embodiment of the present invention. DETAILED DESCRIPTION

[0015] An embodiment of the present invention provides an operation control system for a glue-spinning baking machine. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, systems, products or devices.

[0016] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the control system of the glue-spinning and baking machine in the embodiment of the present invention includes: Initialization module 101, used to detect and initialize parameters of the servo motor and stepper motor of the glue baking machine, and generate initial state data of the equipment; In this embodiment, the upper shovel hand servo motor, the lower shovel hand servo motor, the lifting Z axis servo motor and the rotating axis servo motor of the glue baking machine are detected by the PLC controller during the startup phase. By analyzing the feedback signal of each motor, its power-on state, response delay, encoder signal, etc. are judged to be normal, and the initial state information of the motor is obtained. The control system performs the origin return operation on the above motors in turn according to the initial state information of the motor, that is, guides each motor to return to its mechanical zero position through the motion control instruction and confirms the final position through the near-point DOG photoelectric switch to form a set of stable zero-position reference data. Based on the motor zero-position reference data. The dynamic operating parameters of each servo motor are collected, including the movement speed (pre-photoelectric speed) before the motor approaches the near-point DOG photoelectric switch when performing origin return and the deceleration speed (post-photoelectric speed) after crossing the switch, as well as the high-speed movement speed and low-speed precision positioning speed of the motor in the normal automatic operation state, as well as the acceleration time required to accelerate from a stationary state to a stable speed and the deceleration time required to decelerate from a stable state to a stop. All data constitute the motor operation parameter set. At the same time, real-time data acquisition is performed on the two types of pressure subsystems of the glue baking machine. Multiple sampling and mean filtering are performed on the gas source pressure in the gas supply system and the internal pressure of the glue barrel to eliminate the influence of occasional fluctuations and obtain gas source pressure data and glue barrel pressure data. The internal temperature of the current device is read, and the temperature sensor value on the heating module is compared with the target temperature preset in the controller in real time. The initial temperature deviation data is calculated by difference calculation to evaluate whether the system is started in a stable thermal environment and implement the early compensation strategy for the temperature control system. The motor operation parameter set, gas source pressure data, glue barrel pressure data, and initial temperature deviation data are structured and combined, and written into the state buffer area of ​​the PLC controller to form the initial state data of the device.

[0017] The timing control module 102 is used to perform position accuracy control and vacuum adsorption timing control on the upper shovel hand servo mechanism and the lower shovel hand servo mechanism of the coating baking machine according to the initial state data of the equipment, and obtain the wafer position data and adsorption state data; In this embodiment, the wafer receiving and lifting mechanism of the glue baking machine is tested, and the height encoder and photoelectric position sensor installed on the wafer boat lifting platform are used to obtain the current state information of the number of layers of the wafer boat stack in real time, and the initial position information of each layer of wafers is calculated in combination with the current position of the platform and the slot number of the wafer boat, so as to construct the three-dimensional spatial coordinate reference of the current material collection target. According to the state information of the number of layers of the wafer boat and the initial position information of the wafer, the control system calls the servo control model and acceleration and deceleration parameters generated in the initial state data of the equipment, and performs collaborative control analysis on the motion planning of the upper and lower shovels. The analysis process adopts the interpolation strategy of dual-axis linkage, establishes synchronization constraints in the speed, position and time dimensions, and generates a set of dual-axis linkage motion control instructions including acceleration section, uniform speed section and deceleration section. The controller drives the upper shovel servo mechanism precisely to the preset standby position, and moves the lower shovel servo mechanism to the designated material picking position of the wafer boat. During the whole movement process, the high-frequency position feedback system is used to continuously monitor the trajectory execution of the upper and lower shovels. The position data returned by the servo motor encoder is collected every 10 milliseconds, and compared with the theoretical trajectory in real time to calculate the trajectory deviation data between the actual running path and the set path. Based on the trajectory deviation information, dynamic adjustment is performed through the embedded PID compensation module or the adaptive controller with an extended state observer to correct the motor output command in real time to ensure that the actual motion path is close to the theoretical trajectory and obtain the wafer position data. The precise timing control subroutine of the vacuum adsorption system controls the opening and closing sequence of the vacuum generator and vacuum breaking solenoid valve of the upper and lower shovels respectively. The adsorption process follows the preset vacuum adsorption time and maintains a stable negative pressure state within the set time to ensure that the wafer is reliably adsorbed during the pick-up and placement process without the risk of sliding or falling; the release process performs the exhaust operation according to the vacuum breaking time, so that the airflow impact and the release action are completed synchronously to ensure that the wafer is stably released to the designated position. During this process, the vacuum pressure sensor feeds back the pressure curve in real time. The system determines whether the adsorption is successful based on the curve characteristics and outputs the adsorption status data based on the time node of the adsorption state conversion.

[0018] The rotation speed classification control module 103 is used to perform five-stage rotation speed classification control on the rotation servo device of the coating baking machine based on the wafer position data and the adsorption state data to obtain the rotation speed control data; In this embodiment, according to the wafer position data and adsorption state data, the process recipe parameters corresponding to the current product are called from the data area inside the programmable logic controller (PLC), and the speed control parameter set is extracted from it. The parameter set includes the initial speed value required for the initial stage of glue spreading, and also includes five speed level values ​​executed in sequence, that is, the specific speed settings of the first to fifth sections. At the same time, the duration that each speed state should be maintained is defined, forming a segmented speed control target including multiple physical processes such as acceleration, uniform speed, and deceleration. The control system passes the speed control parameters as input to the extended state observer model, which is used to perform robust estimation of the speed dynamic response process in the presence of system modeling errors and external disturbances, wherein the unmodeled dynamic behavior of the system and various external disturbance signals are abstracted as extended state variables, which are evolved and numerically tracked in real time in the form of differential equations inside the observer, and the output includes the state estimation value of the current speed of the rotating shaft and the estimation value of the interference signal. Using the speed state estimate and the system disturbance estimate, the control system calculates the control command of the rotary axis servo motor based on the linear state feedback strategy, in which the control law combines the offset of the state estimate and the disturbance estimate result to realize the dual control mechanism of feedforward regulation and error feedback correction of the speed change trend. While the motor executes the control command, the system's built-in encoder feedback channel collects the current speed data of the motor in real time, records it as speed measurement data in real time, and compares it with the target speed value in the speed control parameter one by one, performs difference calculation operation, and obtains the real-time speed deviation data under the current operating state, which is used for dynamic error correction and subsequent compensation calculation. The control system integrates and analyzes the real-time speed deviation data and the system disturbance estimate, and dynamically calculates the compensation control amount according to the weight ratio of the two and the current speed transition stage, especially in the speed transition area between segments. The system automatically constructs a flexible acceleration curve, and constructs an acceleration function with continuous differentiable properties according to the difference between the target speed and the current actual speed and the switching time window to prevent drastic speed mutations, thereby avoiding abnormal risks such as wafer eccentricity and glue liquid throwing out. Through iterative closed-loop control, the final output includes speed control data consisting of the actual operating values ​​of the five speed sections, the precise duration of each section, and the degree of fluctuation during the speed maintenance process of each section, and is stored in the PLC log area in real time as a basis for quality tracking.

[0019] The glue coating control module 104 is used to control the glue coating equipment to perform fixed-point glue coating or reciprocating glue coating by using the speed control data, adjust the glue coating stepper motor position and the glue valve opening and closing time, and generate glue coating parameter data; In this embodiment, according to the speed control data, combined with the preset logic of the glue coating strategy in the current process formula, it is determined whether to adopt the fixed-point glue coating mode or the reciprocating glue coating mode, and the corresponding glue coating mode control signal is generated accordingly. The control signal includes parameters such as the glue coating starting position, the stepper motor movement mode, the glue valve opening mode, the reciprocating frequency and the glue coating time window, which serve as the basic input for the subsequent glue coating execution action. According to the glue coating mode control signal instruction, the glue coating stepper motor is driven to move along the preset trajectory to the glue coating starting position, and combined with the feedback signal collected by the high-precision position sensor, the actual spatial deviation between the current glue coating nozzle position and the wafer geometric center is calculated. The deviation is measured in sub-millimeter units and constitutes the position fine-tuning parameter. The position fine-tuning parameter is input into the position compensation control algorithm, and the position output trajectory of the glue coating stepper motor is corrected in real time to ensure that the nozzle is aligned with the target glue coating area, and the target glue coating position control result is formed. At the same time, the glue barrel pressure system is sampled and monitored in real time, and a pressure sensor installed near the glue barrel outlet is used to obtain a continuous pressure signal stream, and the noise is eliminated through filtering to form stable pressure control data, which reflects the flow stability of the current glue supply system. Based on the glue coating mode control signal and the target glue coating position control result, the glue coating action execution stage is entered. In the fixed-point glue coating mode, the control system triggers the glue valve opening and closing control channel, and uses a high-precision timer to ensure that the glue valve opening time is strictly consistent with the process setting, thereby ensuring that the release volume of the glue per unit time meets the uniform distribution requirements; in the reciprocating glue coating mode, the system controls the stepper motor to complete a stable reciprocating motion between the starting position and the end position according to the set acceleration curve, and at the same time controls the glue valve to open and close multiple times on the reciprocating path according to the preset trigger frequency, to ensure that the glue evenly covers the specific path area of ​​the wafer, and outputs the glue valve action data of the entire glue coating stage, including the time point, duration and relative position of each opening within the motion cycle. After the gluing process is completed, the previously collected pressure control data is correlated and analyzed with the glue valve action data generated during the gluing process. The gluing position, glue valve opening time, pressure change trend and glue amount fluctuation amplitude are integrated and calculated to construct a multi-dimensional glue quality evaluation model, extract the glue uniformity index, evaluate the consistency of the spatial distribution and the local coating accuracy of the gluing process, and generate structured glue parameter data, recording multiple key variables including the target gluing position coordinates, the specific time node of each glue valve action, the real-time pressure change curve during gluing and the coating distribution uniformity evaluation index.

[0020] The compensation control module 105 is used to perform inter-region heat coupling compensation control on the four-region heating units of the glue coating baking machine according to the glue coating parameter data, and output the temperature uniformity control result.

[0021] In this embodiment, the extraction unit performs structured analysis on the glue coating parameter data generated by the upstream glue coating control module, and extracts the core type information representing the product process attributes, including key variables such as glue coating process category, colloid viscosity grade, rotation speed bandwidth and coating thickness grade. Based on the process type information, the preset process temperature control parameter library is queried, and the corresponding four-zone temperature setting values ​​and heating time requirements of each zone are extracted to generate structured four-zone temperature control target parameters. This parameter set is used as the basis for downstream temperature control logic modeling, and the temperature control target and thermal field response sequence are distinguished by two heating units on each side of A / B. The construction unit uses the above control target parameters to establish a temperature closed-loop control model, and calculates the temperature closed-loop variables for the four temperature control zones, including the real-time temperature deviation value, the deviation integral value and the deviation change rate, which correspond to the system's dynamic response capabilities to the instantaneous temperature difference, the cumulative temperature difference trend and the temperature change speed, respectively. The three indicators are combined to construct the basic control quantity, and used as the input variable of PID regulation for closed-loop control scheduling. The division unit divides the current temperature control state of the system into intervals according to the absolute value of the temperature, and clearly divides it into three types of thermal states: low temperature section, medium temperature section and high temperature section, and generates corresponding segmented temperature control instructions according to the response inertia and heat demand characteristics of different temperature sections. At the same time, the heat interference compensation unit models and processes the thermal energy coupling phenomenon existing in the four-zone heating system. The system builds a coupled thermal resistance network model based on the heat conduction path and the layout structure of the heating unit, and calculates the intensity of the heat interaction between the regions to obtain the heat coupling compensation data between the regions. The compensation data corrects the original heating output in real time during the thermal response adjustment process to prevent the temperature rise in a certain area from causing passive overheating or temperature control instability in other areas. On this basis, the advance compensation unit introduces a prediction mechanism for external environmental temperature disturbances. By collecting the historical change curve of the environmental temperature and combining the response sensitivity of each region to external temperature fluctuations during the operation cycle, an environmental disturbance trend model is constructed, and the model prediction results are integrated with the regional coupling compensation data to generate a comprehensive temperature compensation control instruction with feedforward characteristics in advance. The real-time monitoring unit performs fusion processing on the aforementioned segmented temperature control instructions and comprehensive temperature compensation control instructions, and uniformly outputs them to the solid-state relay control unit to achieve real-time adjustment and dynamic loading of the power of each heating unit; at the same time, while executing the output, the monitoring unit performs high-frequency sampling and data fitting analysis on the heating power change curve of each temperature control area. When it is detected that the power of any temperature control unit exceeds the preset safety upper limit threshold, the system automatically triggers the protection mechanism, suspends the heating output and feeds back an abnormal signal; under normal control conditions, the module continuously tracks the temperature control accuracy, and forms a temperature uniformity control result based on the temperature stability, fluctuation amplitude and temperature difference between each area.

[0022] According to the thermal coupling modeling procedure, the four independent temperature control areas arranged in the glue-spreading baking cavity, namely the first area of ​​the first side, the second area of ​​the first side, the first area of ​​the second side, and the second area of ​​the second side, were subjected to standardized thermal response tests. During the test, a standard heating pulse signal with a constant amplitude and a set duration was applied to each heating unit in turn, and the remaining three areas were controlled to maintain an isolated state with zero thermal power. At the same time, the temperature change process of the three areas without heating was recorded in real time by high-precision thermocouples to form a response curve sequence, and the initial test data of heat transfer containing unidirectional heat transfer information between all areas was generated. Based on the initial test data, by fitting the lag time, maximum temperature rise amplitude and heat decay rate of each pulse response curve, the equivalent heat reception value of each target area when thermally stimulated by other areas was calculated, and a 4×4 heat transfer coefficient matrix with heating sources as rows and heated areas as columns was constructed. Since the matrix diagonal represents the self-heating response and the non-diagonal elements express the cross-interference intensity, the matrix is ​​structured, its numerical range is standardized, and weak coupling channels below the set threshold are eliminated to generate a thermal coupling mathematical model, which reflects the dynamic thermal coupling relationship between the temperature control areas within the system and is used for subsequent real-time interference estimation. After the model is built, the current temperature values ​​of the four temperature control areas are collected in real time, and the difference is calculated with the control target temperature to obtain the temperature deviation vector; the control system combines the deviation vector with the coupling model to analyze the deviation value of each area, trace the reason for its deviation from the target to what proportion of the disturbance of heat transfer in the neighboring area, and estimate the unit heat contribution of the three neighboring areas respectively, and obtain the quantitative data of the cross-effect of the temperature range. The temperature rise share caused by the non-self-controlled heating currently in each area is characterized, providing a numerical basis for the reverse regulation mechanism. Based on the reverse compensation control principle, these cross-thermal interference quantities are decomposed, and the heating control signal of each temperature control channel is adjusted in different zones to generate pre-compensation control instructions, which include the correction increments of the original PID output. These increments are reflected in reserving some thermal power margin to eliminate external heat input, or actively reducing power during the strong thermal disturbance stage in the neighboring area to offset the superposition effect. The pre-compensation control instructions are put into execution, and the actual temperature change curve after execution is recorded synchronously, and the error analysis is performed with the theoretical prediction curve. By constructing the least squares optimization function, the parameter values ​​in the original heat transfer coefficient matrix are adaptively updated, and the matrix structure and coupling strength weight distribution are iteratively optimized, so as to refine the thermal coupling compensation data between regions, and finally form a thermal interference suppression mechanism with dynamic closed-loop update capability.

[0023] The initialization module 101 is specifically used for: Perform status detection on the upper shovel hand servo motor, lower shovel hand servo motor, lifting Z-axis servo motor and rotating axis servo motor of the glue spreading and baking machine to obtain the initial status information of the motor; According to the initial state information of the motor, each servo motor is respectively subjected to the origin return operation to obtain the motor zero position reference data; Based on the motor zero position reference data, the operating parameters of each servo motor are collected to obtain the motor operating parameter set including the speed before the near point DOG photoelectric, the speed after the near point DOG photoelectric, the automatic high speed, the automatic low speed, the acceleration time and the deceleration time during origin return; Perform pressure value detection on the air source system and glue barrel pressure system of the glue baking machine to obtain air source pressure data and glue barrel pressure data; Collect the initial temperature of the glue baking machine, calculate the difference with the set temperature, and generate the initial temperature deviation data; Generate equipment initial state data based on the motor operation parameter set, air source pressure data, glue barrel pressure data and temperature initial deviation data.

[0024] In this embodiment, the servo motor status detection program for the key actuators is started to establish a control benchmark. The main control unit performs startup signal verification and response feedback analysis on the upper shovel hand servo motor, the lower shovel hand servo motor, the lifting Z-axis servo motor, and the rotating axis servo motor in turn according to the preset timing. The system determines whether the motor is in a controllable and drivable stable state by reading the position signal, current load curve, motor start response time, and driver status bit returned by the encoder in real time, and obtains the initial state information of the motor based on this, and identifies whether the servo system has error conditions such as overcurrent protection, position loss, excitation abnormality, and driver failure that do not allow it to enter the operation stage. When the system confirms that the above four core motor modules are in a healthy state and have a complete feedback path, the initialization logic enters the next stage and performs the origin return operation. According to the encoder resolution of each motor, the position relationship between the mechanical limit structure and the proximity sensor, the origin movement instructions are issued to the four servo mechanisms respectively. The upper shovel hand and the lower shovel hand servo motors move the clamping arm to the limit zero position through horizontal drive; the lifting Z-axis servo motor drives the vertical platform back to the lower limit position of the Z axis; the rotating axis servo motor adjusts the pallet rotation mechanism to the set mechanical reference angle, and reduces the movement speed when approaching the proximity DOG photoelectric sensor to ensure the reliability of zero point capture. The zero point arrival confirmation of each axis is triggered by the proximity DOG signal and is stabilized for not less than the specified time threshold. The system then records the corresponding encoder value to obtain the motor zero position reference data. These reference data calibrate the physical zero point of each servo mechanism in space, constitute the coordinate system benchmark for subsequent position control, and are cached in the control memory as the boundary conditions for real-time control of the system. After obtaining the zero-point reference data, the operating parameters of each servo motor are collected through the high-frequency communication bus interface between the servo drive and the controller, including: the feed speed before the DOG photoelectric trigger (i.e., the speed before the photoelectric trigger) during the origin return process, the deceleration precision positioning speed after the DOG trigger (i.e., the speed after the photoelectric trigger), the automatic high-speed speed and low-speed precision control speed corresponding to the normal operation stage, and the acceleration time required for the servo motor to accelerate from static to high-speed operation and the deceleration time required from high-speed stop. These parameters reflect the dynamic response capability of the motor and are also the basic variables of the speed profile in motion planning. The system automatically classifies and stores the collected parameters to form a motor operating parameter set, providing a basic input data source for subsequent trajectory planning, speed segmentation control and PID adjustment. At the same time, the system synchronously starts the evaluation of the pressure state of the pneumatic control circuit and the glue supply system to obtain stable physical output conditions.For the air source system, the current air pressure value is collected by the pressure sensor installed in the main air intake pipe, and the average filtering algorithm is used to offset the short-term fluctuation caused by the start / stop of the compressor to generate the air source pressure data; for the glue barrel pressure system, the current working pressure of the glue liquid transmission system is read in real time by the embedded pressure sensor in the glue barrel, and it is verified whether it is within the pressure bandwidth where the glue liquid can be stably pushed out, and the glue barrel pressure data is obtained. The initial temperature state in the glue baking cavity is collected. The temperature acquisition unit reads the current actual temperature value through the high-precision thermistor or thermocouple sensor located at the key node of the heating cavity, and performs a difference operation with the pre-start temperature reference value set in the batch process recipe to generate the initial temperature deviation data. The deviation value is used to determine whether the current thermal field meets the environmental requirements for immediate execution of glue coating or rotation. If the deviation exceeds the set threshold, the system automatically starts the preheating subroutine to increase the hot zone temperature to the process start-up requirements; if the deviation is acceptable, it automatically enters the next control process. The initial temperature deviation data is also recorded in the data cache table in the initialization stage, which is used for the temperature control adaptive control module to dynamically determine the adjustment direction and compensation amplitude when performing preheating or coupling compensation control. The above-mentioned various types of collected initialization operation data, namely the motor operation parameter set, air source pressure data, glue barrel pressure data and temperature initial deviation data are integrated, and a structured data structure unit is constructed inside the controller and saved in the system operation memory as the initial state data of the equipment.

[0025] The timing control module 102 is specifically used for: Detect the wafer receiving and lifting mechanism of the coating baking machine to obtain the wafer boat layer status information and the wafer initial position information; According to the wafer boat layer status information and wafer initial position information, the upper and lower shovel hands coordinated control analysis is performed on the equipment initial status data to obtain the dual-axis linkage motion control instructions; Based on the dual-axis linkage motion control command, the upper shovel hand servo mechanism moves to the standby position, the lower shovel hand servo mechanism moves to the crystal boat material taking position, and the real-time trajectory deviation data is monitored; According to the real-time trajectory deviation data, the motion trajectories of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism are compensated in real time to obtain the wafer position data; For the vacuum adsorption system of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism, the adsorption and release processes are controlled according to the preset vacuum adsorption time and vacuum breaking time to obtain adsorption state data.

[0026] In this embodiment, the wafer receiving and lifting mechanism in the core mechanical structure is detected to identify the number of stacked layers of the current wafer boat and the initial position information of each layer of wafers in three-dimensional space. The control system monitors the vertical motion range and stop position of the wafer boat in real time through the height encoder or laser displacement sensor installed on the lifting platform. Combined with the standard thickness and spacing of each layer of wafers in the structural parameters, the number of stacked wafer boat layers is quickly identified through the calculation formula, and combined with the preset calibration values ​​of the wafer spacing and the tray position in the wafer boat, the initial spatial coordinates of each layer of wafers are calculated to obtain the wafer boat layer state information and the wafer initial position information. On this basis, the initial state data of the device generated by the completed initialization process is retrieved, which contains key control variables such as the motion capacity parameters of the servo motor, the position control boundary, and the acceleration and deceleration time constant. The controller uses this initial state data, combined with the acquired wafer target position, to call the dual-axis collaborative control algorithm to coordinate and analyze the movement of the upper and lower shovel hands as a whole. The analysis includes the synchronous start and stop of the two mechanisms on the time axis, and also requires the avoidance strategy on the position axis, that is, to ensure that the motion trajectories of the two do not interfere with each other, and at the same time, the speed profile design must meet the stability requirements of wafer safe handling. Through the above analysis, a set of dual-axis linkage motion control instructions including the starting position, target position, speed segment planning, and synchronous control signals are generated, and sent to the corresponding servo driver control channel in real time. The control system drives the upper shovel servo mechanism to move slowly to the set standby position according to the control instructions. The standby position is located in the middle buffer area between the wafer loading area and the lifting platform to ensure that the subsequent pick-up and placement actions have path redundancy; at the same time, the lower shovel servo mechanism starts the precise positioning mode, driving the clamping structure to move to the material position of the current target layer of the wafer boat. Its vertical movement is jointly constrained by the height sensor and the wafer boat layer number recognition results to ensure that the position of the aligned target wafer does not deviate. During the entire movement process, the system starts the trajectory tracking subroutine, collects the position data fed back by the servo motor encoder every 10 milliseconds, and compares it with the theoretical trajectory in real time, calculates the error between the actual trajectory and the planned trajectory, and generates trajectory deviation data. Based on the real-time trajectory deviation data, the motion trajectory of the upper and lower shovels is dynamically corrected through the embedded compensation control module. The compensation method combines PID regulation and extended state observer (ESO) modeling principles. On the one hand, the position change is smoothed through proportional, integral and differential actions. On the other hand, ESO is used to estimate and feedback the non-modeled disturbance, thereby maintaining high-precision trajectory following capabilities throughout the entire movement process, so that the error between the actual wafer receiving position and the theoretical design position is controlled within the specified range. After this stage is completed, the final position after correction is recorded as the actual wafer position data and marked as the spatial input for subsequent rotation control and glue alignment. When the position accuracy is guaranteed, the system enters the adsorption control stage, and executes two sequential subtasks of vacuum adsorption and vacuum breaking for the vacuum adsorption system equipped with the upper and lower shovels.In the material picking action, according to the vacuum adsorption time parameters set by the process, the vacuum pump is turned on and the corresponding solenoid valve is opened, so that the clamping structure quickly establishes a negative pressure area to stably adsorb the wafer. During the adsorption process, the pressure value of the negative pressure chamber is continuously monitored by the vacuum pressure sensor, and a safe holding time is maintained after reaching the set threshold. If the pressure drops to a dangerous range, the system will issue an alarm or terminate the handling operation. Subsequently, during the wafer placement stage, the control system calls the vacuum breaking time parameter, quickly releases the vacuum state, opens the exhaust channel to restore the pressure to normal pressure, and realizes the undisturbed release of the wafer. During the above adsorption and release process, the system will record the adsorption state change curve in real time and generate adsorption state data, which reflects whether the wafer is completely adsorbed, whether there are hidden dangers such as slippage or adsorption failure, and is assigned "successful", "abnormal", "unadsorbed" and other marks through the standardized logic judgment module. The actual position data of the wafer and the adsorption state data are packaged together and uploaded to the data cache module.

[0027] The speed classification control module 103 is specifically used for: Read the current process recipe parameters from the data area of ​​the programmable logic controller according to the wafer position data and the adsorption state data, and obtain the speed control parameters including the starting speed of the coating, the speed value of the first section, the speed value of the second section, the speed value of the third section, the speed value of the fourth section, the speed value of the fifth section, and the duration of each section; Input the speed control parameters into the extended state observer model, observe the unmodeled dynamics and external disturbances of the system as extended states, and calculate the speed state estimation value and the system disturbance estimation value; The rotation axis servo motor control signal is calculated using the speed state estimation value and the system disturbance estimation value, and the motor drive instruction is output through the linear state feedback control strategy, so that the rotation axis servo motor runs according to the set five-stage speed requirements and obtains real-time speed measurement data; The difference between the real-time speed measurement data and the target speed value in the speed control parameter is calculated to obtain the real-time speed deviation data; The dynamic compensation control amount is calculated based on the real-time speed deviation data and the system interference estimation value, and the acceleration curve is dynamically calculated based on the difference between the current speed and the target speed to obtain the speed control data including the actual operating values ​​of the five speed segments, the duration of each segment and the speed stability data.

[0028] In this embodiment, according to the wafer position data and adsorption state data, by reading the data structure associated with the current wafer process recipe in the data area of ​​the programmable logic controller (PLC), the glue control parameter set corresponding to the current category is called and parsed, and the rotation speed planning information is obtained in the parameter set, including the glue starting speed, the first speed value, the second speed value, the third speed value, the fourth speed value and the fifth speed value, and the set duration of each speed stage is read to form a five-stage speed control target model driven by process logic. The five-stage speed control parameters are input into the extended state observer (ESO) model to enhance the fault tolerance of the rotation control system to modeling errors and environmental disturbances. The ESO model is a state estimator built based on the observer structure. It not only observes the main state variable of the system, that is, the estimated value of the current angular velocity of the rotating shaft, but also introduces the unmodeled dynamics (such as motor friction changes, load disturbances, wind resistance changes, etc.) and external disturbances (such as air disturbances generated near the glue nozzle, power supply fluctuations, etc.) into the system state space, and uniformly estimates them as part of the extended state. The model uses the actual speed measurement value as feedback input, and drives the differential equation to iteratively calculate through error feedback, and obtains two core results: the estimated value of the rotation state and the estimated value of the system disturbance, which represent the actual operating trend of the system under non-ideal conditions. The control system calls the linear state feedback control strategy and constructs the control signal of the rotating axis servo motor according to the above ESO output results. The control system performs a difference operation between the estimated value of the rotation state and the target speed of the current stage to calculate the speed error term; at the same time, the estimated value of the system disturbance is introduced as a feedforward compensation to construct a closed-loop control law with disturbance suppression capability. The control signal output by the control law will be transmitted to the rotating motor driver, and the motor will be driven by voltage, current or PWM duty cycle adjustment to achieve real-time control of angular acceleration, so that the rotating platform is executed according to the five-stage speed model: from the starting speed of the glue-splitting in the initial stage, it smoothly switches to the set value of each speed stage, and keeps the speed stable within each specified time, forming a smooth and high-precision multi-stage speed operation trajectory. To ensure the feedback loop of the execution process, the system's supporting angle encoder continuously outputs the real-time speed measurement data of the current rotating shaft at a high sampling frequency. These data are recorded and compared with the set value of the current speed stage in real time. After the difference operation, the real-time speed deviation data is generated. The positive and negative distribution, duration and convergence speed of the deviation value are analyzed in real time, and combined with the system interference estimation value output in the ESO, it is judged whether the current deviation is caused by a temporary disturbance or a long-period structural error. On this basis, the dynamic compensation control amount is calculated to adjust the gain distribution and acceleration change trend of the rotation control.Especially in the critical interval where switching occurs between speed segments, the system calculates the dynamic acceleration trajectory based on the difference between the current speed and the target speed, converts the speed transition process into a smooth and derivable acceleration curve, and avoids impact speed mutations at the switching point, thereby effectively suppressing the occurrence of wafer offset, glue throwing, or adsorption failure due to excessive inertia. At the same time, the dynamic compensation control quantity is used to adjust the proportional band and integral time constant of the servo control, so that the system control strategy maintains a good balance between fast response and stable accuracy. After the above steps, the system summarizes the actual operating values ​​of the five speed segments in real time, including the measured average speed of each segment, the instantaneous speed at start and end, the actual running time of each segment, and the fluctuation amplitude within the speed stability range, to construct a structured speed control data set.

[0029] The gluing control module 104 is specifically used for: Determine the fixed-point gluing mode or the reciprocating gluing mode according to the speed control data and generate a corresponding gluing mode control signal; Based on the glue coating mode control signal, the glue coating stepper motor is controlled to move to the preset glue coating starting position, and the actual position deviation between the wafer center and the glue coating valve nozzle is calculated through the position sensor feedback signal to obtain the position fine-tuning parameter; Use the position fine-tuning parameters to perform position compensation control on the glue-spreading stepper motor to obtain the target glue-spreading position control result. At the same time, the glue barrel pressure system is monitored in real time to generate pressure control data. Based on the glue coating mode control signal and the target glue coating position control result, the glue valve opening time is accurately controlled in the fixed-point glue coating mode, or the glue coating stepper motor is controlled to reciprocate between the glue coating start position and the glue coating end position in the reciprocating glue coating mode, and the glue valve is controlled to open and close according to the preset number of times to obtain the glue valve action data; The pressure control data and the glue valve action data are correlated and analyzed to generate glue coating parameter data including glue coating position, glue valve opening time, glue coating pressure change and glue coating uniformity evaluation index.

[0030] In this embodiment, the speed control data output in the current rotation stage is used as the input basis to make a decision logic judgment on the glue coating strategy. The system reads the speed segment characteristics, speed stability, time continuity and rotational inertia change trend contained in the speed control data, and compares and analyzes it with the glue viscosity grade, coating thickness requirements and wafer size indicators in the process formula parameters. When the judgment result meets the low-speed start, high stable speed and small wafer size, the system selects the fixed-point glue coating mode; if high-speed glue spreading and long speed duration or process requirements for large-area distributed glue layer are detected, the system selects the reciprocating glue coating mode. Based on this logic, the controller generates a glue coating mode control signal, which includes glue coating type information (fixed point or reciprocating), and also carries detailed parameters such as the coordinates of the glue coating starting point, the direction of the glue coating path, the preset opening time or the number of reciprocating times, as the main control instruction sent to the glue coating actuator. The control system drives the glue coating stepper motor to move to the preset glue coating starting position according to the glue coating mode control signal. The starting position is slightly outside the geometric center of the wafer. The preset path is aligned with the edge of the wafer and is configured as the starting point of the spiral, radial or linear reciprocating mode according to the specific process requirements. After the stepper motor completes the rough positioning, the system starts the position sensor sampling process, uses a high-precision photoelectric sensor, laser rangefinder or inductive displacement sensor to scan the current position of the nozzle in real time, and compares its measured spatial coordinates with the actual center coordinates of the wafer, so as to calculate the spatial offset between the nozzle outlet and the center of the wafer, that is, the position fine-tuning parameter, which is measured at the sub-millimeter level and reflects the impact of installation tolerance, actuator repeatability error or slight wafer offset on the glue coating positioning accuracy. The position fine-tuning parameter is input into the compensation control submodule to generate a fine-tuning compensation control instruction to correct the end position of the glue coating stepper motor so that its output meets the set target glue coating position, forming a target glue coating position control result, and ensuring that the glue coating action is initiated from the correct reference point. At the same time, the control system activates the glue barrel pressure monitoring channel in parallel, calls the pressure sensor installed at the output end of the glue barrel to collect real-time pressure values, and performs filtering and linear fitting processing during the sampling process to form highly stable pressure control data. This data is used to determine whether the pressure is in the normal range when the glue valve is opened, and adjust the pressure supply intensity of the compressed gas inside the glue barrel or the output of the driving hydraulic system through feedback control. The control system starts the glue execution logic synchronously according to the glue coating mode control signal and the target glue coating position control result. In the fixed-point glue coating mode, the system controls the glue valve to remain stationary at the target nozzle position, and controls the opening and closing time of the glue valve solenoid switch through a high-precision timer. The opening time is calculated based on the viscosity of the glue liquid, the flow coefficient and the target coating thickness, and the control accuracy per unit time reaches the millisecond level; during the opening of the glue valve, the system continuously reads the pressure data to ensure that the glue liquid output process is stable and continuous, without bubbles, blockages or pulses; after the glue valve is closed, the system records the actual opening time and pressure fluctuations to form the first set of glue coating process data.In the reciprocating glue coating mode, the control system starts the stepper motor to reciprocate along the predetermined path according to the glue coating path planning information. The path has a set length, speed curve and repetition frequency, and triggers the opening and closing of the glue valve according to the control logic when the path moves forward to the key node. Each opening and closing is completed synchronously at the key point of the path to ensure that the glue is evenly distributed on the wafer surface. The number of times the glue valve is opened during the entire reciprocating process, the start time and closing time of each opening, the corresponding position of the path, and the corresponding pressure fluctuation data are all recorded in real time by the system to form glue valve action data containing path information and dynamic pressure behavior. During the entire glue coating process, the control system jointly analyzes the real-time collected pressure control data and the synchronously generated glue valve action data to construct a multi-dimensional behavior model of the glue coating process. The pressure change curve at each time the glue valve is opened is differentiated to extract the pressure rise time, steady-state plateau period and decline trend, and compared with the theoretical pressure-flow response model to evaluate whether the glue output is stable; at the same time, the glue coating position and the glue valve opening time are temporally and spatially correlated to determine whether the glue landing point is consistent with the predetermined trajectory; the path repetition accuracy is analyzed in combination with the stepper motor trajectory feedback data to evaluate the uniformity of the coating distribution on the surface. The above analysis results are structured and packaged to generate glue coating parameter data including the glue coating space position, the actual glue valve opening time, the pressure change curve during glue coating and its stability fluctuation index, and the coating coverage uniformity evaluation index.

[0031] The compensation control module 105 further includes: An extraction unit is used to extract process type information from the glue coating parameter data, and read the temperature setting value and heating time requirement according to the process type information to generate four-zone temperature control target parameters; A construction unit is used to construct a temperature closed-loop control model based on the temperature control target parameters of the four zones, and calculate the temperature deviation value, deviation integral value and deviation change rate of each temperature zone respectively to obtain the temperature closed-loop basic control quantity; A division unit is used to divide the working range into a low temperature section, a medium temperature section and a high temperature section according to the temperature closed-loop basic control quantity, and generate segmented temperature control instructions corresponding to different temperature sections; The heat interference compensation unit is used to model the heat transfer characteristics of the temperature intervals of the four-zone heating units and to compensate for the heat interference, so as to obtain the inter-zone heat coupling compensation data; The advance compensation unit is used to combine the inter-regional thermal coupling compensation data to make advance compensation for the system deviation caused by the change of external ambient temperature and generate a comprehensive temperature compensation control instruction; The real-time monitoring unit is used to synthesize the segmented temperature control instructions and the comprehensive temperature compensation control instructions and output them to the solid-state relay control unit. At the same time, the heating power of each temperature interval is monitored in real time to obtain the temperature uniformity control result.

[0032] In this embodiment, the gluing parameter data generated by the gluing control module is parsed by the extraction unit, and the parameter data structure contains a series of process type information associated with heat treatment, such as the coating type, glue type, gluing path morphology, coating thickness requirements and wafer size information corresponding to the current process. The key fields representing the change law of the thermal field demand are extracted from it by the feature index algorithm, and matched with the thermal treatment parameter template defined in the preset process library as input variables, so as to identify the thermal treatment strategy required for the current product. After the identification is completed, the corresponding temperature setting value and the target heating time of each area are called out from the process template. These values ​​are divided into the first area of ​​the A side, the second area of ​​the A side, the first area of ​​the B side, and the second area of ​​the B side, respectively corresponding to the set thermal field state of the four heating areas, thereby forming a four-area temperature control target parameter set. The construction unit takes the temperature control target parameter as a reference input, and combines the real-time temperature acquisition value, the historical temperature control error data and the response inertia of the servo heating system in the current system operation to establish a multi-channel temperature control model based on closed-loop PID control. The control model establishes independent temperature adjustment equations for the four heating zones, and calculates the temperature deviation value (i.e., the difference between the set value and the current value), the deviation integral value (used to reflect the long-term steady-state error trend), and the deviation change rate (reflecting the current temperature rise or drop trend) in real time in each round of sampling cycle. The three indicators together constitute the temperature closed-loop basic control quantity of each temperature control loop. The construction unit inputs the basic control quantity as a driving variable into the temperature control law to generate the power output control signal of each zone to the solid-state relay or PID power adjustment module, ensuring that each temperature zone can track the set temperature target in the form of an autonomous closed loop. The division unit performs logical segmentation processing on the temperature control task, and classifies the current temperature control state into the low temperature section, medium temperature section, or high temperature section according to the temperature closed-loop basic control quantity calculated for each zone, especially the deviation value and the deviation change rate. There are significant differences in the thermal response characteristics under different temperature sections, such as fast response but low error tolerance in the low temperature section, high stability requirements in the medium temperature section, and significant power output restrictions in the high temperature section. Therefore, the division unit generates PID parameter groups and output control curves matching different temperature sections according to the segmentation logic, forming segmented temperature control instructions, and automatically switching control strategies for each temperature zone to ensure that the entire heating process achieves the optimal balance between stability and response speed. At the same time, in order to effectively solve the mutual interference between temperature zones caused by heat conduction, radiation and equipment structure coupling, the thermal interference compensation unit in the system starts the heat transfer model construction program. This program is based on the internal thermal structure parameters of the equipment, combined with the standard thermal response data obtained in the previous experimental stage, and introduces the heat transfer matrix form in the control logic to quantify the thermal coupling relationship between the four areas. The thermal interference compensation unit represents the influence of the power change of each temperature control area on the temperature response of the other three areas with a matrix coupling coefficient, and updates the parameters in the matrix in real time to obtain the thermal coupling compensation data between the areas.In the stage of drastic temperature change or control command transition, the system superimposes this compensation data on the control output of each temperature control loop, and performs feedforward correction on the bias temperature error caused by the interference path, thereby preventing the cumulative effect of local overheating or regional thermal imbalance. Considering the impact of non-structural factors such as external environmental temperature changes, ventilation disturbances or aging of heat dissipation structures on the system thermal field control, the advance compensation unit establishes an environmental interference prediction model by long-term recording of environmental temperature fluctuations, temperature control error change trends and historical control response behaviors. When the external environmental temperature change trend is detected, the system combines the thermal coupling compensation data to calculate the error direction and magnitude generated in each area in advance, and injects the corresponding compensation amount in advance to form a comprehensive temperature compensation control instruction. The real-time monitoring unit merges the segmented temperature control instruction with the comprehensive temperature compensation control instruction, and uniformly outputs it to the solid-state relay control unit or digital power module to achieve precise driving of the power output of the four heating areas; at the same time, the unit collects the power output curve, heating rate and steady-state maintenance capability of each area at a high sampling rate, and compares and analyzes the time consistency and spatial synchronization between the target temperature and the actual temperature to obtain the temperature uniformity control result.

[0033] Wherein, the heat interference compensation unit is specifically used for: Sequentially apply standard heating pulses to the four-zone heating units of the first zone on the first side, the second zone on the first side, the first zone on the second side, and the second zone on the second side in the coating baking machine, and record the temperature response curves of the other three zones to generate initial test data of heat transfer; The heat transfer coefficients between the regions are calculated based on the initial test data of heat transfer to obtain a transfer coefficient matrix, and the transfer coefficient matrix is ​​structured to obtain a heat coupling mathematical model; According to the thermal coupling mathematical model, the deviation data of the current temperature and the target temperature of the four zones collected in real time are analyzed, and the heat interference from the other three zones is calculated for each temperature zone to obtain the quantitative data of the cross-effect of the temperature zones; Based on the quantitative data of cross-effects of temperature intervals, the reverse compensation principle is applied to adjust the heating control output of each temperature zone and calculate the pre-compensation control instructions; The pre-compensation control instructions are compared and analyzed with the actual temperature changes, and the parameter values ​​in the 4×4 coupling matrix are optimized by the least squares method to generate the inter-regional thermal coupling compensation data.

[0034] In this embodiment, the initial mathematical expression of the thermal coupling relationship is constructed by an active excitation experiment. The control system applies standardized heating pulse signals to the first zone of the first side, the second zone of the first side, the first zone of the second side, and the second zone of the second side according to a predetermined test procedure. The heating pulse has a constant power amplitude and a fixed time length, and keeps the three areas outside the test area in a closed state or a stable state during the heating process, thereby minimizing the influence of external interference, so that the thermal response process reflects the influence of the test area on the thermal radiation, heat conduction and heat diffusion of the remaining areas. During the application of the standard heating pulse, the temperature response curve of the non-heating area is recorded in a high-frequency sampling manner through thermocouples or thermistor sensors arranged at multiple points in each temperature control area, and the actual temperature rise curve of the heating area is recorded synchronously. For each set of tests, the curve data of the three cross-heat response channels are collected respectively to form the initial test data of heat transfer. The above test data is fitted and analyzed, and the heat transfer coefficient between the areas is calculated by analyzing the ratio of the response delay and the temperature rise amplitude between different areas. If the four temperature zones are numbered as A1, A2, B1, and B2, the system takes the heating zone as the source and the heated zone as the target, and sequentially constructs a 4×4 heat transfer coefficient matrix G, in which the diagonal elements represent the unit effect coefficient of the heating of the zone on the temperature increase of the zone, and the non-diagonal elements represent the unit transfer effect of the source zone on the other three target zones, that is, the degree of thermal coupling. After the matrix G is calculated, it is structured and the elements in the matrix are numerically normalized, outliers are removed, the dynamic response range is compressed, and the symmetry check is performed to improve its computability and robustness in the subsequent real-time compensation model, and the thermal coupling mathematical model is obtained. The model is used to describe the heat flow propagation path, influence intensity, and disturbance coupling structure between the four zones. The thermal coupling mathematical model is applied to the real-time control compensation logic in the actual operation process. In each temperature control cycle of the system, the controller collects the current actual temperature data of the four temperature control zones in real time, and calculates the difference between it and the target temperature set in the recipe parameters to form the current temperature deviation vector ΔT. The controller then inputs the deviation vector into the reverse analysis module based on the established thermal coupling mathematical model, and for each temperature zone, calculates the component of its temperature deviation caused by heat transfer from the other three temperature zones, that is, quantifies the coupled thermal interference effects from three non-local heat sources, and outputs quantitative data representing the intensity of mutual interference between zones, forming a quantitative data set of cross-effects of temperature intervals. After obtaining the cross-effect data, the control system corrects the heating power output of each current temperature control zone based on the reverse compensation principle. If the current temperature deviation of a certain area is positive, but most of it is caused by heat transfer from other areas, the actual heating power of this area should be actively reduced; if there is a negative deviation in a certain area but no significant thermal interference occurs in other areas, the power of this area should be increased to quickly restore the target temperature.By constructing a reverse compensation matrix ΔQ, in which each element represents the thermal power correction value required to compensate a temperature control area under coupling interference conditions, the system vector superimposes the original PID controller output with ΔQ to generate a set of dynamic pre-compensation control instructions. The temperature response after the execution of the pre-compensation control instruction is compared with the output of the prediction model, and the deviations between the two in the time axis and amplitude are compared, so as to analyze the fitting deviation between each parameter item in the thermal coupling model and the actual thermal response. On this basis, the least squares method is used to optimize the parameters of the original 4×4 coupling matrix. The optimization process takes the actual temperature rise data as the observation value and the original thermal coupling model as the prediction equation. By constructing the error minimization objective function and executing the matrix inversion or iterative optimization algorithm, the coupling coefficient in the matrix is ​​automatically adjusted to improve the accuracy and stability of the model under the current equipment state and environmental conditions. Through the closed-loop thermal coupling modeling and compensation mechanism including "excitation, observation, modeling, compensation, and optimization", a structured understanding of the internal heat conduction law of the equipment is achieved, and a thermal control logic with self-learning and self-correction capabilities is constructed.

[0035] In the embodiment of the present invention, the extended state observer and dynamic parameter adjustment technology are used to achieve precise control of the rotation axis speed, the speed stability is improved from the traditional ±10rpm to ±1rpm, and the uniformity of glue spreading is significantly improved. By constructing a collaborative control mechanism between the functional units, the upper and lower shovels, the rotation system, the glue coating system and the heating system are fed back in a closed loop, the overall system operation is more coordinated, and the incoordination problem caused by the independent control of each unit is effectively solved. The 4×4 thermal coupling matrix model is introduced to accurately describe the heat transfer relationship between the heating areas. Through the inter-area thermal coupling compensation control, the temperature control accuracy is improved from ±2℃ to ±0.5℃, which significantly improves the overall heating uniformity of the wafer. The adaptive flow control algorithm combined with pressure feedback realizes the control of the glue coating positioning accuracy within the range of ±0.1mm, the consistency of the glue coating amount is greatly improved, and the variation between product batches is reduced. Based on the abnormal detection and processing mechanism of multi-sensor fusion, the system can monitor the operating status in real time, automatically identify and handle various abnormal situations, prevent equipment damage and product scrapping, and ensure production safety. By combining human-machine interface with automatic control, operators can achieve complex process control through simple parameter settings and process selection, reducing the dependence on operator skills.

[0036] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned system embodiments and will not be repeated here.

[0037] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a glue baking machine operation control device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the system described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0038] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control system for a glue-spinning and baking machine, characterized in that: include: The initialization module is used to detect and initialize the parameters of the servo motor and stepper motor of the glue baking machine and generate the initial state data of the equipment; A timing control module is used to perform position accuracy control and vacuum adsorption timing control on the upper shovel hand servo mechanism and the lower shovel hand servo mechanism of the coating baking machine according to the initial state data of the equipment, and obtain wafer position data and adsorption state data; A rotation speed classification control module, used for performing five-stage rotation speed classification control on the rotation servo device of the coating baking machine based on the wafer position data and the adsorption state data to obtain rotation speed control data; A glue coating control module, used to use the speed control data to control the glue coating equipment to perform fixed-point glue coating or reciprocating glue coating, adjust the position of the glue coating stepping motor and the opening and closing time of the glue valve, and generate glue coating parameter data; The compensation control module is used to perform inter-region heat coupling compensation control on the four-region heating unit of the glue coating baking machine according to the glue coating parameter data, and output a temperature uniformity control result.

2. The operation control system of the glue baking machine according to claim 1 is characterized in that: The initialization module is specifically used for: Perform status detection on the upper shovel hand servo motor, lower shovel hand servo motor, lifting Z-axis servo motor and rotating axis servo motor of the glue spreading and baking machine to obtain the initial status information of the motor; According to the motor initial state information, each servo motor is respectively subjected to an origin return operation to obtain motor zero position reference data; Based on the motor zero position reference data, the operating parameters of each servo motor are collected to obtain a motor operating parameter set including the speed before the near point DOG photoelectric, the speed after the near point DOG photoelectric, the automatic high speed, the automatic low speed, the acceleration time and the deceleration time during origin return; Performing pressure value detection on the air source system and the glue barrel pressure system of the glue spreading and baking machine to obtain air source pressure data and glue barrel pressure data; The initial temperature of the glue baking machine is collected, and the difference is calculated with the set temperature to generate initial temperature deviation data; The device initial state data is generated based on the motor operation parameter set, the air source pressure data, the glue barrel pressure data and the temperature initial deviation data.

3. The operation control system of the glue baking machine according to claim 1 is characterized in that: The timing control module is specifically used for: Detecting the wafer receiving and lifting mechanism of the coating baking machine to obtain the wafer boat layer status information and the wafer initial position information; According to the wafer boat layer status information and the wafer initial position information, upper and lower shovel hand coordinated control analysis is performed on the initial state data of the equipment to obtain a dual-axis linkage motion control instruction; Based on the dual-axis linkage motion control instruction, the upper shovel hand servo mechanism is moved to the standby position, the lower shovel hand servo mechanism is moved to the wafer boat material taking position, and the real-time trajectory deviation data is monitored; According to the real-time trajectory deviation data, the motion trajectories of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism are compensated in real time to obtain wafer position data; For the vacuum adsorption system of the upper shovel hand servo mechanism and the lower shovel hand servo mechanism, the adsorption and release processes are controlled according to the preset vacuum adsorption time and vacuum breaking time to obtain adsorption state data.

4. The operation control system of the glue baking machine according to claim 1, characterized in that: The speed classification control module is specifically used for: Read the current process recipe parameters from the data area of ​​the programmable logic controller according to the wafer position data and the adsorption state data, and obtain the speed control parameters including the starting speed of the coating, the first speed value, the second speed value, the third speed value, the fourth speed value, the fifth speed value and the duration of each section; Inputting the speed control parameters into an extended state observer model, observing the unmodeled dynamics of the system and external disturbances as extended states, and calculating a speed state estimation value and a system disturbance estimation value; The rotation axis servo motor control signal is calculated using the speed state estimation value and the system interference estimation value, and a motor drive instruction is output through a linear state feedback control strategy, so that the rotation axis servo motor runs according to the set five-stage speed requirement, and real-time speed measurement data is obtained; Calculating the difference between the real-time speed measurement data and the target speed value in the speed control parameter to obtain real-time speed deviation data; The dynamic compensation control amount is calculated according to the real-time speed deviation data and the system interference estimation value, and the acceleration curve is dynamically calculated according to the difference between the current speed and the target speed to obtain the speed control data including the actual operating values ​​of five speed sections, the duration of each section and the speed stability data.

5. The operation control system of the glue baking machine according to claim 1, characterized in that: The gluing control module is specifically used for: Determine a fixed-point gluing mode or a reciprocating gluing mode according to the speed control data and generate a corresponding gluing mode control signal; Based on the glue coating mode control signal, the glue coating stepping motor is controlled to move to a preset glue coating starting position, and the actual position deviation between the wafer center and the glue coating valve nozzle is calculated through the position sensor feedback signal to obtain the position fine-tuning parameter; The position fine-tuning parameters are used to perform position compensation control on the glue coating stepping motor to obtain the target glue coating position control result, and at the same time, the glue barrel pressure system is monitored in real time to generate pressure control data; Based on the glue coating mode control signal and the target glue coating position control result, the glue valve opening time is accurately controlled in the fixed-point glue coating mode, or the glue coating stepping motor is controlled to reciprocate between the glue coating start position and the glue coating end position in the reciprocating glue coating mode, and the glue valve is controlled to open and close according to a preset number of times, so as to obtain the glue valve action data; The pressure control data and the glue valve action data are correlated and analyzed to generate glue coating parameter data including glue coating position, glue valve opening time, glue coating pressure change and glue coating uniformity evaluation index.

6. The operation control system of the glue baking machine according to claim 1, characterized in that: The compensation control module further includes: An extraction unit, used to extract process type information from the glue coating parameter data, and read the temperature setting value and heating time requirement according to the process type information to generate four-zone temperature control target parameters; A construction unit is used to construct a temperature closed-loop control model based on the temperature control target parameters of the four zones, and calculate the temperature deviation value, deviation integral value and deviation change rate of each temperature zone respectively to obtain a temperature closed-loop basic control amount; A division unit, used for dividing the working range into a low temperature section, a medium temperature section and a high temperature section according to the temperature closed-loop basic control quantity, and generating segmented temperature control instructions corresponding to different temperature sections; The heat interference compensation unit is used to model the heat transfer characteristics of the temperature intervals of the four-zone heating units and to compensate for the heat interference, so as to obtain the inter-zone heat coupling compensation data; An advance compensation unit, used to combine the inter-region heat coupling compensation data, to compensate in advance for the system deviation caused by the change in external ambient temperature, and to generate a comprehensive temperature compensation control instruction; The real-time monitoring unit is used to synthesize the segmented temperature control instruction and the comprehensive temperature compensation control instruction and output them to the solid-state relay control unit, and at the same time monitor the heating power of each temperature interval in real time to obtain the temperature uniformity control result.

7. The operation control system of the glue-spinning and baking machine according to claim 6, characterized in that: The heat interference compensation unit is specifically used for: Sequentially apply standard heating pulses to the four-zone heating units of the first zone on the first side, the second zone on the first side, the first zone on the second side, and the second zone on the second side in the coating baking machine, and record the temperature response curves of the other three zones to generate initial test data of heat transfer; Calculating the heat transfer coefficient between the regions based on the initial test data of heat transfer to obtain a transfer coefficient matrix, and performing structural processing on the transfer coefficient matrix to obtain a heat coupling mathematical model; According to the thermal coupling mathematical model, the deviation data of the current temperature of the four zones and the target temperature collected in real time are analyzed, and the heat interference amount from the other three temperature zones is calculated for each temperature zone to obtain the quantitative data of the cross-effect of the temperature zones; Based on the quantitative data of the cross-effect of the temperature intervals, the reverse compensation principle is applied to adjust the heating control output of each temperature zone in a partitioned manner, and a pre-compensation control instruction is calculated; The pre-compensation control instruction is compared and analyzed with the actual temperature change, and the parameter values ​​in the 4×4 coupling matrix are optimized by the least square method to generate inter-regional thermal coupling compensation data.

Images