Wafer correction method, conveying device, control apparatus, and storage medium

Abstract

The application discloses a wafer deviation correction method, a conveying device, a control device and a storage medium. The method obtains the calibration positions of a plurality of wafer sensors in a robot coordinate system, collects the code disc value, speed value and sampling time of a driving motor during the wafer deviation correction detection movement of the robot carrying the wafer, obtains the level change time stamp with higher time accuracy than the sampling time and the same time reference when the wafer sensor output signal level changes, extracts the local sampling data covering the level change time stamp and reconstructs the corresponding code disc value to determine the robot position and the trigger point, and then calculates the wafer offset according to the sensor calibration position and the trigger point and controls the robot deviation correction. The application can obtain the trigger instant position without relying on the high-precision position latching of the driver, and reduce the influence of false triggering through the geometric calculation between the sensor calibration position and the trigger point, thereby improving the wafer deviation correction accuracy, robustness and wafer conveying stability in the existing wafer conveying device.

Description

Technical Field

[0001] This application relates to the field of wafer transfer technology, and in particular to a wafer bias correction method, transfer device, control equipment, and storage medium. Background Technology

[0002] In semiconductor manufacturing, wafers typically need to be transferred between different workstations, chambers, or carrier positions. Wafer transfer robots operating in vacuum environments usually carry the wafer at their end effector. Since the relative position between the wafer and the robot's end effector is mainly maintained by contact support or frictional constraints, the wafer may shift relative to the robot's end effector during extension, retraction, acceleration, or deceleration. If this shift is not accurately identified and corrected before wafer placement, it can easily lead to inconsistencies between the wafer placement position and the taught or target position, affecting the stability and safety of subsequent process steps.

[0003] Existing wafer eccentricity correction solutions typically place wafer sensors near the wafer transport path. When the wafer edge passes through the sensor's detection area, the sensor's output signal changes. The system calculates the wafer's offset relative to the robot based on the geometric relationships between the robot's position, sensor position, and wafer radius, and controls the robot to perform position correction before wafer unloading. To obtain the robot's position at the instant the wafer sensor is triggered, some solutions rely on the driver's position latching function. That is, the driver directly latches the motor position when the sensor is triggered, and the controller then reads this latched position for eccentricity correction calculation.

[0004] However, in existing wafer transfer equipment, some ordinary drivers are not equipped with high-precision hardware probe latch interfaces, or their position latching response speed and synchronization accuracy are insufficient to meet the requirements of high-precision wafer spin correction. If a dedicated driver or high-performance controller is replaced to achieve high-precision position latching, it is usually necessary to redesign the driver installation method and cable connection relationship in the control cabinet, and may require extensive modification of the original lower-level control program, resulting in high system upgrade costs and affecting the stability of mature equipment architecture.

[0005] Furthermore, traditional two-sensor offset correction schemes are susceptible to spurious trigger points when handling wafers with chamfered edges or notches. For example, when handling wafers with chamfered edges, if both sensors simultaneously scan the chamfered edge instead of the curved edge, the two chamfered edge trigger points may fit a circle with the same radius as the wafer. This makes it difficult for the system to distinguish between genuine curved edge trigger points and spurious trigger points, thus affecting the reliability of wafer offset calculation. For instance, when handling wafers with chamfered edges, if both sensors simultaneously scan the chamfered edge instead of the curved edge, there may be a problem of failing to eliminate spurious trigger points.

[0006] Therefore, how to accurately obtain the robot position corresponding to the moment the wafer sensor is triggered without relying on the high-precision position latching function of the driver and without significantly modifying the existing wafer transfer equipment, and how to stably calculate the wafer offset when false trigger points exist due to edge cutting, notches, or obstructions, are the problems that need to be solved in the current active wafer offset correction. Summary of the Invention

[0007] This application provides a wafer spin correction method, a transfer device, a control device, and a storage medium to solve the problem that in existing wafer spin correction schemes, when the driver does not support high-precision position latching and there is spurious trigger interference, it is difficult to balance low modification cost, accuracy of position acquisition at the moment of triggering, and spin correction robustness.

[0008] This application provides a wafer correction method in a first aspect, comprising: acquiring the calibration positions of multiple wafer sensors in a robot coordinate system; during the correction detection motion of the robot carrying the wafer, acquiring the encoder value, speed value, and sampling time of the robot drive motor according to a sampling period; monitoring the output signals of the multiple wafer sensors, and acquiring a level change timestamp when the output signal of any wafer sensor changes level, wherein the time accuracy of the level change timestamp is higher than the time accuracy of the sampling time, and both have the same time reference; extracting local sampling data from the acquired encoder value, speed value, and sampling time, wherein the time range of the local sampling data covers the level change timestamp; reconstructing the encoder value corresponding to the level change timestamp based on the local sampling data; determining the robot position corresponding to the level change timestamp based on the reconstructed encoder value, thereby determining the trigger point of the wafer sensor corresponding to the level change timestamp; calculating the wafer offset based on the calibration positions of the multiple wafer sensors and the corresponding trigger points, and controlling the robot to perform correction.

[0009] Furthermore, the level change includes the output signal level change caused when the wafer edge enters the detection area of ​​the wafer sensor, and / or the output signal level change caused when the wafer edge leaves the detection area of ​​the wafer sensor.

[0010] Furthermore, local sampling data is extracted from the collected encoder values, speed values, and sampling times, including: determining the sampling period to which the level change timestamp belongs based on its position in the sampling time sequence; determining a local time window covering the level change timestamp based on the sampling period to which the level change timestamp belongs; and extracting the encoder values, speed values, and sampling times within the local time window to obtain local sampling data.

[0011] Furthermore, the local time window covers the sampling period to which the level change timestamp belongs, and the two sampling periods immediately before and after that sampling period.

[0012] Furthermore, the code disk value corresponding to the level change timestamp is reconstructed based on the local sampling data, including: constructing a continuous motion model of the robot drive motor based on the local sampling data; and substituting the level change timestamp into the continuous motion model to obtain the code disk value corresponding to the level change timestamp.

[0013] Furthermore, the continuous motion model is a cubic spline interpolation model, which is constructed based on the encoder value constraints and velocity value constraints in the local sampled data.

[0014] Furthermore, determining the robot position corresponding to the level change timestamp based on the reconstructed encoder value includes: determining the movement position of the robot drive motor at the time corresponding to the level change timestamp based on the reconstructed encoder value; and determining the robot position corresponding to the level change timestamp based on the movement position of the robot drive motor at the time corresponding to the level change timestamp and the robot kinematic model.

[0015] Furthermore, obtaining the calibration positions of multiple wafer sensors in the robot's coordinate system includes: controlling the robot to carry the calibration wafer along a first trajectory past the wafer sensor, obtaining a first candidate position based on the robot's position when the calibration wafer triggers the wafer sensor; controlling the robot to carry the calibration wafer along a second trajectory with an offset angle relative to the first trajectory past the wafer sensor, obtaining a second candidate position based on the robot's position when the calibration wafer triggers the wafer sensor; and determining the calibration position of the wafer sensor based on the distance between the first candidate position and the second candidate position.

[0016] Furthermore, the first candidate position includes multiple first candidate solutions, and the second candidate position includes multiple second candidate solutions; the calibration position of the wafer sensor is determined based on the distance between the first candidate position and the second candidate position, including: calculating the Euclidean distance between each first candidate solution and each second candidate solution respectively; determining the first candidate solution and the second candidate solution with the smallest Euclidean distance as the matching candidate solution; and determining the calibration position of the wafer sensor based on the matching candidate solution.

[0017] Furthermore, the multiple wafer sensors include a first wafer sensor, a second wafer sensor, and a third wafer sensor. The detection positions of the first wafer sensor and the third wafer sensor are located on both sides of the movement trajectory of the wafer center, respectively, and the detection position of the second wafer sensor is located between the first wafer sensor and the third wafer sensor.

[0018] Furthermore, the wafer offset is calculated based on the calibration positions of multiple wafer sensors and their corresponding trigger points, including: obtaining the trigger point sets corresponding to the first wafer sensor, the second wafer sensor, and the third wafer sensor respectively; selecting trigger points from the trigger point sets corresponding to the first wafer sensor, the second wafer sensor, and the third wafer sensor respectively to form a trigger point combination; and calculating the wafer offset based on the trigger point combination and the calibration position of the corresponding wafer sensor.

[0019] Furthermore, based on the combination of trigger points and the calibration position of the corresponding wafer sensor, the wafer offset is calculated, including: with the goal of satisfying the wafer radius constraint between the trigger point and the calibration position of the corresponding wafer sensor after compensating for the wafer offset, the wafer offset is solved using the least squares method.

[0020] Furthermore, the calculation of wafer offset based on the trigger point combination and the corresponding wafer sensor calibration position also includes: calculating the radius error for different trigger point combinations; and determining the wafer offset corresponding to the trigger point combination with the smallest radius error as the final wafer offset.

[0021] Furthermore, controlling the robotic arm to perform correction includes: if the wafer offset is not greater than the preset maximum correction distance, controlling the robotic arm to perform correction action; if the wafer offset is greater than the preset maximum correction distance, controlling the robotic arm to stop moving and outputting an error message.

[0022] This application provides a wafer transfer device in a second aspect, including a robot arm, a driver, multiple wafer sensors, an input acquisition unit, and a main controller. The robot arm is used to carry a wafer for spin correction and detection motion. The driver is connected to the drive motor of the robot arm to drive the robot arm and to acquire the encoder value and speed value of the drive motor. Multiple wafer sensors are disposed in the wafer spin correction and detection area and are used to output signals with level changes when the wafer edge passes through the corresponding detection area. The input acquisition unit is communicatively connected to the multiple wafer sensors and the main controller, respectively, to monitor the output signals of the multiple wafer sensors and to record the level change timestamp when the output signal of any wafer sensor changes level. The main controller is communicatively connected to the driver and the input acquisition unit, respectively, to acquire the encoder value, speed value, and sampling time collected according to the sampling period from the driver, to acquire the level change timestamp from the input acquisition unit, and to execute the wafer spin correction method provided in the first aspect based on the encoder value, speed value, sampling time, level change timestamp, and the calibration position of the multiple wafer sensors in the robot arm coordinate system.

[0023] This application provides a control device in a third aspect, which serves as the main controller of a wafer transfer device. The wafer transfer device includes a robotic arm, a driver, multiple wafer sensors, and an input acquisition unit. The input acquisition unit monitors the output signals of the multiple wafer sensors and records a time stamp of the level change when the output signal of any wafer sensor changes level. The control device includes a processor, a memory, and a communication interface. The communication interface is used to communicate with the driver and the input acquisition unit respectively. The memory stores a computer program. When the computer program is executed by the processor, the processor obtains the code disk value and speed value fed back by the driver, the sampling time, and the time stamp of the level change recorded by the input acquisition unit through the communication interface, and implements the wafer spin correction method provided in the first aspect.

[0024] In a fourth aspect, this application provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor of a control device, the processor acquires, through a communication interface, the code disk value and speed value fed back by the driver in the wafer transfer device, the sampling time, and the level change timestamp recorded by the input acquisition unit, and implements the wafer spin correction method provided in the first aspect.

[0025] The above technical solution has at least the following beneficial effects: By acquiring the encoder value, speed value, and sampling time of the robot's drive motor during the wafer spin correction detection process, and obtaining the level change timestamp with a time accuracy higher than the sampling time and under the same time base as the sampling time when the output signal of the wafer sensor changes level, the precise timing information of the wafer edge trigger sensor can be obtained first. Then, the encoder value and robot position corresponding to the trigger moment can be reconstructed using the local sampling data covering the level change timestamp. Thus, the robot position corresponding to the wafer sensor trigger point can be obtained without relying on the driver to perform high-precision position latching at the moment of sensor triggering, reducing the dependence on dedicated latching drivers or high-performance controllers, and improving the accuracy of wafer spin correction calculation under the premise of low modification cost of existing wafer transfer devices.

[0026] The above technical solution also has the following advantages: By calibrating the offset trajectory of the wafer sensor, the calibration position corresponding to the actual physical installation position can be selected from the candidate solutions, reducing the risk of calibration failure caused by sensor orientation uncertainty or wafer placement deviation; by setting three wafer sensors and calculating the wafer offset based on trigger point combination, wafer radius constraint, least squares method and radius error, pseudo trigger points introduced by factors such as edge cutting, notch or occlusion can be screened out from multiple trigger points, thereby improving the correction robustness and wafer placement reliability in irregular wafer scenarios such as wafers with edge cutting and wafers with notches. Attached Figure Description

[0027] The present application will be further described below with reference to the accompanying drawings and embodiments: Figure 1 A schematic diagram of the wafer transfer device provided in this application; Figure 2 A schematic flowchart illustrating the wafer polarization correction method provided in this application; Figure 3 A schematic diagram illustrating the principle of reconstructing the code disk value corresponding to the level change timestamp provided in this application; Figure 4 This is a schematic diagram of wafer sensor calibration provided in this application; Figure 5 This is a schematic diagram of wafer offset calculation provided in this application; Figure 6 A schematic diagram of the control device structure provided in this application.

[0028] Wherein: 110, robotic arm; 111, drive motor; 120, driver; 130, first wafer sensor; 131, second wafer sensor; 132, third wafer sensor; 140, input acquisition unit; 150, main controller; 160, wafer; 161, wafer with diced edge; 170, wafer spin correction detection area; 200, control device; 210, processor; 220, memory; 230, communication interface. Detailed Implementation

[0029] The embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the following embodiments are intended to help those skilled in the art understand and implement the technical solutions of this application, and do not constitute a limitation on the scope of protection of this application. Where there is no conflict, the technical features in the following embodiments can be combined with each other.

[0030] In the description of the embodiments of this application, the terms "first," "second," "third," etc., are used only to distinguish the same or similar objects, steps, signals, positions, or components, and do not indicate that these objects, steps, signals, positions, or components have a sequential order, a difference in importance, a limitation in quantity, or a limitation in structure. For example, "first wafer sensor," "second wafer sensor," and "third wafer sensor" are used only to distinguish different wafer sensors, and do not indicate that any one of the wafer sensors necessarily takes precedence over other wafer sensors in terms of function, installation order, or triggering order. The terms "comprising," "including," "having," and any variations thereof are intended to indicate non-exclusive inclusion. For example, an apparatus, method, or device that includes a component, step, unit, or data does not exclude the inclusion of other components, steps, units, or data not expressly listed. Unless the context expressly defines it otherwise, the expression "comprising one..." does not exclude the inclusion of one or more other objects of the same type. The terms "connection," "coupling," "communication connection," etc., should be interpreted broadly. For example, two components can be directly connected or indirectly connected through intermediate components; they can be mechanical, electrical, signal, communication, or data connections; they can be wired or wireless; they can be fixed or detachable. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific technical scenario. The terms "acquire," "collect," "receive," and "read" do not limit the data source or the method of data acquisition. For example, data can be directly obtained from a local sensor, driver, input acquisition unit, or memory, or indirectly obtained through a communication interface, bus, network, or other data path, or obtained by calculating, transforming, filtering, or reconstructing existing data. The terms "determine," "calculate," "obtain," "generate," and "reconstruct" can indicate direct calculation or acquisition through table lookup, fitting, interpolation, filtering, model processing, coordinate transformation, or multi-step operations. Unless otherwise explicitly defined, the above terms do not require a unique mathematical model, a fixed algorithm form, or a specific hardware implementation. The term "based on" can be understood as "at least partially based on." In other words, when a processing procedure is executed based on certain data, conditions, or parameters, it does not preclude the procedure from being executed in conjunction with other data, conditions, or parameters. For example, reconstructing encoder values ​​based on local sampled data can be done using only the local sampled data, or it can be further combined with speed constraints, robot kinematics models, filtering parameters, or equipment calibration parameters. The term "preset" indicates that the relevant parameters, conditions, thresholds, rules, or models can be determined before the corresponding steps are executed. Preset content can be obtained through user configuration, equipment factory configuration, automatic system calibration, historical data learning, issuance by the host computer, or invocation by the control program, and is not limited to fixed constants.

[0031] In the description of the embodiments of this application, unless otherwise specified, parameters such as numerical values, angles, distances, times, periods, and resolutions can be adjusted according to wafer size, robot structure, sensor type, control cycle, equipment safety clearance, and process precision requirements. The specific numerical values ​​given in the embodiments are only used to illustrate one possible implementation and do not mean that this application can only use those specific numerical values.

[0032] In the description of the embodiments in this application, the modules, units, components, lines, arrows, and geometric relationships in the accompanying drawings are used to help illustrate the structural, signal, or computational relationships of the technical solutions, and do not necessarily represent the size proportions, installation proportions, relative distances, or unique layout methods in actual products. Those skilled in the art can adjust the corresponding structures according to the actual equipment space, wiring methods, and control architecture without changing the technical concept of this application.

[0033] In this embodiment, the wafer can be a sheet-like workpiece that needs to be transferred between different workstations, chambers, carriers, or other carrier positions during the semiconductor manufacturing process. The robot can be a robotic arm, wafer transfer arm, or other wafer transfer actuator used to carry and transfer the wafer. The drive motor can be a motor used to drive the robot to extend, retract, rotate, or perform compound movements. The driver can be a servo driver, stepper driver, or other motion control drive unit connected to the drive motor and used to control the operation of the drive motor. The main controller can be a motion controller, industrial controller, host control device, embedded controller, or other control unit capable of performing data acquisition, calculation, and motion control. The robot coordinate system can be a reference coordinate system used to describe the robot's position, wafer sensor position, and wafer offset, which can be established based on the robot's taught position, HOME position, target workstation position, or device coordinate system.

[0034] In this embodiment, a wafer sensor refers to a sensor used to detect whether the edge of a wafer passes through a corresponding detection area. The wafer sensor can be a photoelectric sensor, a through-beam sensor, a reflective sensor, or other detection element capable of generating a change in output signal level when the wafer enters or leaves the detection area. The wafer bias correction detection area refers to the local sensing area corresponding to the wafer sensor. In one implementation, the wafer bias correction detection area can be the optical path area of ​​a photogate. When the wafer edge enters this optical path area and blocks the light signal, or when the wafer edge leaves the optical path area and the light signal is restored, the output signal level of the wafer sensor changes. It should be understood that... Figure 1 The multiple wafer biasing detection areas 170 in the model correspond to different wafer sensors and are not a large area that collectively surrounds multiple wafer sensors.

[0035] In this embodiment, the code disk value refers to the encoder feedback position value of the robot's drive motor, or a position feedback value that has an equivalent function to the encoder feedback position value. The sampling time refers to the moment when the main controller acquires the code disk value and speed value from the driver according to the sampling period. The level change timestamp refers to the timestamp recorded by the input acquisition unit when the output signal of the wafer sensor undergoes a level change. The time accuracy of the level change timestamp is higher than the time accuracy of the sampling time, and the level change timestamp and the sampling time are based on the same time reference. For ease of description, the moment when the wafer sensor output signal undergoes a level change, represented by the level change timestamp, can be denoted as... .

[0036] In this embodiment of the application, local sampling data refers to the time range covering the timestamps of level changes. A set of sampled data, which may include data located in Multiple sampling times in the vicinity, their corresponding encoder values, and speed values. A trigger point refers to a timestamp based on a certain level change. The reconstructed robot position. In other words, the trigger point is not a fixed physical point on the wafer edge, nor is it the mounting position of the wafer sensor, but rather the position of the robot when a certain wafer sensor experiences a voltage level change. The calibrated position of the wafer sensor refers to its position in the robot's coordinate system. The wafer offset refers to the offset vector generated by the actual wafer transport state relative to the reference state or taught state. The radius error is the error value calculated based on the trigger point combination, sensor calibration position, and wafer radius constraint, used to evaluate whether the trigger point combination conforms to the geometric relationship of true triggering at the wafer's arc edge.

[0037] Existing wafer spin correction solutions typically rely on the robot arm position corresponding to the sensor trigger moment for offset calculation. If the driver has a high-precision position latching function, the motor position can be directly latched at the moment of sensor triggering. However, in some existing wafer transport equipment, ordinary drivers are not equipped with high-precision hardware probe latching interfaces, or the position latching accuracy, response speed, and synchronization capability of the drivers cannot meet the requirements of high-precision wafer spin correction. Replacing dedicated drivers or reconstructing the control system would result in high equipment modification costs. In addition, factors such as wafers with chamfered edges, wafers with notches, or robot arm obstruction may cause false trigger points for wafer sensors, leading to unstable offset calculations. Based on this, the embodiments of this application reconstruct the code disk value at the moment of triggering by using high-precision level change timestamps and periodic sampling data, and combine the calibration positions and trigger point combinations of multiple wafer sensors to perform radius error screening, thereby achieving wafer spin correction without relying on high-precision position latching by the driver.

[0038] Figure 1 This is a schematic diagram of the wafer transfer device provided in this application. Figure 1 As shown, the wafer transfer device may include a robotic arm 110, a drive motor 111, a driver 120, a first wafer sensor 130, a second wafer sensor 131, a third wafer sensor 132, an input acquisition unit 140, and a main controller 150. The robotic arm 110 is used to carry the wafer 160 for transfer and correction / detection movements. The drive motor 111 is connected to the robotic arm 110, and the driver 120 is connected to the drive motor 111 to drive the drive motor 111 and acquire the encoder value and speed value of the drive motor 111.

[0039] The first wafer sensor 130, the second wafer sensor 131, and the third wafer sensor 132 are disposed near the wafer transport path and are used to generate a change in the output signal level when the wafer 160 passes through the corresponding wafer spin correction detection area 170. The detection positions of the first wafer sensor 130 and the third wafer sensor 132 can be located on opposite sides of the wafer's center movement trajectory, respectively, and the detection position of the second wafer sensor 131 can be located between the detection positions of the first wafer sensor 130 and the third wafer sensor 132. Each wafer sensor has a corresponding wafer spin correction detection area 170. When the wafer 160 blocks or leaves a certain wafer spin correction detection area 170, the output signal level of the corresponding wafer sensor changes.

[0040] The input acquisition unit 140 is connected to the first wafer sensor 130, the second wafer sensor 131, and the third wafer sensor 132 respectively, and is used to monitor the output signals of each wafer sensor and record the level change timestamp when the output signal of any wafer sensor changes level. The main controller 150 is communicatively connected to both the driver 120 and the input acquisition unit 140. The main controller 150 can acquire the encoder value, speed value, and sampling time from the driver 120, and can acquire the level change timestamps from the input acquisition unit 140. The input acquisition unit 140 can be an IO module with timestamp recording function, or an acquisition board, input interface circuit, FPGA acquisition unit, or other input acquisition device with edge capture and timestamp recording functions.

[0041] In this embodiment, the main controller 150, driver 120, and input acquisition unit 140 can be connected via an industrial control bus. For example, a communication link can be established via an EtherCAT bus, enabling the main controller 150 to read the encoder value and speed value fed back by the driver 120 according to the sampling period, and to obtain the time stamp of level change recorded by the input acquisition unit 140. In one specific embodiment, the sampling period can be 1ms, and the input acquisition unit 140 records the timestamps of level changes. The time resolution can reach the nanosecond level. It should be understood that, in the embodiments of this application, the sampling period and timestamp resolution can be configured according to the control system performance, sensor type, and correction accuracy requirements, as long as the level changes and the timestamp... The time precision must be higher than the time precision of the sampling time, and both must be under the same time reference.

[0042] Figure 2 This is a flowchart illustrating the wafer polarization correction method provided in this application. Figure 2 As shown, the wafer correction method may include steps S1 to S8.

[0043] In step S1, the calibration positions of multiple wafer sensors in the robot's coordinate system are obtained. The multiple wafer sensors may include a first wafer sensor 130, a second wafer sensor 131, and a third wafer sensor 132. The calibration positions of the wafer sensors can be obtained in advance through a calibration process, or they can be re-obtained after equipment maintenance, model change, or sensor position adjustment.

[0044] In step S2, during the spin correction and detection process of the robot arm 110 carrying the wafer 160, the code disk value, speed value, and sampling time of the robot arm drive motor 111 are collected according to the sampling period. The code disk value can be the encoder position feedback value of the drive motor 111, and the speed value can be the angular velocity, linear velocity, or speed information obtained by differentiating the code disk value of the drive motor 111.

[0045] In step S3, the output signals of multiple wafer sensors are monitored, and when the output signal of any wafer sensor changes level, a time stamp of the level change is obtained. The level change can include the output signal level change caused by the wafer edge entering the wafer sensor detection area, or the output signal level change caused by the wafer edge leaving the wafer sensor detection area. The input acquisition unit 140 can record the level change timestamp when a rising edge, falling edge, or preset edge change is detected. .

[0046] In step S4, local sampling data is extracted from the acquired encoder values, speed values, and sampling times. The time range of the local sampling data covers the level change timestamps. For example, it can be based on the timestamp of level changes. Determine the position of the level change timestamp in the sampling time sequence. The sampling period is determined; then, based on this sampling period, a local time window is determined, and the encoder value, speed value, and sampling time within the local time window are extracted.

[0047] In step S5, the time stamp of the level change is reconstructed based on the local sampling data. The corresponding code disk value. Due to the level change timestamp. The time precision is higher than the time precision of the sampling moment, and the level change timestamp They are usually located between adjacent sampling times. Therefore, the code disk value corresponding to a certain sampling time cannot be directly used as the code disk value at the time of level change. Instead, it is necessary to reconstruct the code disk value corresponding to the time of level change through local sampling data.

[0048] In step S6, the level change timestamp is determined based on the reconstructed code disk value. The corresponding robotic arm position is used to determine the time stamp of the level change. The corresponding trigger point of the wafer sensor. Specifically, the time stamp of the drive motor 111 at the level change can be determined first based on the reconstructed code disk value. Based on the motion position at the corresponding moment, and the transmission relationship between the drive motor 111 and the robot arm 110, or the kinematic model of the robot arm, the position of the robot arm 110 in the robot arm coordinate system at that moment is determined. This robot arm position can serve as the trigger point for the wafer sensor that experiences a level change.

[0049] In step S7, the wafer offset is calculated based on the calibration positions of multiple wafer sensors and their corresponding trigger points. The calibration positions of the multiple wafer sensors provide a fixed detection reference, and the trigger points characterize the position of the robot 110 when the wafer edge triggers the corresponding wafer sensor. The geometric constraint between the trigger point and the corresponding calibration position can be understood as follows: at the moment when a certain wafer sensor experiences a level change, the wafer edge corresponds to the detection position of that wafer sensor; therefore, the distance between the wafer center and the calibration position of that wafer sensor should be equal to the wafer radius. If the trigger point represents the position of the robot 110 at that moment, and the wafer offset represents the offset of the wafer relative to the robot 110, then the wafer center position can be determined by the trigger point, calibration position, and wafer offset, and a wafer radius constraint can be established. This geometric constraint will... Figure 5 The corresponding embodiments are further illustrated.

[0050] In step S8, the robot arm 110 is controlled to perform correction based on the wafer offset. If the wafer offset is not greater than the preset maximum correction distance, the main controller 150 can generate a correction control quantity based on the wafer offset and control the robot arm 110 to perform the correction action. If the wafer offset is greater than the preset maximum correction distance, it can be considered that the wafer 160 offset is too large or the detection result is abnormal. The main controller 150 can control the robot arm 110 to stop moving and output an error reminder. In a specific embodiment, the preset maximum correction distance can be set to 2mm. When the wafer offset modulus is not greater than 2mm, the main controller 150 controls the robot arm 110 to perform compensation movement based on the wafer offset; when the wafer offset modulus is greater than 2mm, the main controller 150 stops the robot arm 110's wafer placement action and outputs an error reminder. It should be understood that in the embodiments of this application, the preset maximum correction distance can also be set to other values ​​based on the wafer size, robot arm structure, workstation safety clearance, and process tolerance, such as 1mm, 3mm, or other distance values ​​adapted to the safety requirements of the equipment.

[0051] Figure 3 This is a schematic diagram illustrating the principle of reconstructing the code disk value corresponding to the level change timestamp provided in this application. (For example...) Figure 3 As shown, the main controller 150 acquires the sampling time according to the sampling period. Corresponding encoder value and speed value. Level change timestamp. It can be located between two adjacent sampling times. The main controller 150 can be... Based on the sampling period, determine the coverage. A local time window is defined, and the encoder value, speed value, and sampling time are extracted from the local time window as local sampling data.

[0052] In one embodiment, the local time window can cover the level change timestamp. The sampling period includes the current sampling period, the two sampling periods preceding it, and the two sampling periods following it. In other words, a local time window can include five sampling points. Using this local time window achieves a balance between computational complexity and trajectory fitting accuracy. For wafer alignment detection motion, the robot 110 is typically in an acceleration phase, deceleration phase, or start / end transition phase during extension or retraction. If only linear interpolation is used, the motion between adjacent sampling periods is easily approximated as uniform motion, leading to a large error in the code disk value reconstruction. Constructing a continuous motion model using multiple sampling points can more accurately reflect the speed changes during this phase.

[0053] In one embodiment, a cubic spline interpolation model can be constructed based on local sampled data, and the time stamps of level changes can be used. Substituting into the cubic spline interpolation model, we obtain The corresponding reconstructed code disk value. For the first... Each sampling period The continuous motion model of the encoder value changing over time can be represented as:

[0054] in, Indicates the first Code disk value function within a sampling period; , , and The coefficients to be solved; For the first Each sampling time. Let... The value of the encoder disk collected at any time , The value of the encoder disk collected at any time It can be obtained based on the positional continuity condition. and .set up express exist The second derivative at each sampling point can be obtained by solving for the second derivative at each sampling point based on the continuity condition of the first derivative between adjacent sampling periods and the endpoint velocity constraint, and then further determining the coefficients of the cubic spline interpolation model.

[0055] In one implementation, if the sampling period The length is Then we have:

[0056]

[0057]

[0058]

[0059] For internal sampling points, a linear equation can be constructed based on the continuity condition of the first derivative of adjacent sampling periods:

[0060] For the start and end points of a local time window, boundary conditions can be set based on the velocity values ​​in the local sampled data. For example, at the start point, the first derivative can be taken as the velocity value. At the endpoint, the first derivative can be taken as the velocity value. The solution yields each Then, the cubic spline interpolation model for the corresponding sampling period can be determined. Subsequently, the time stamps of the level changes are... Substitute the corresponding sampling period The reconstructed encoder value is obtained. .

[0061] It should be understood that, in the embodiments of this application, the cubic spline interpolation model is only one preferred implementation. Other models capable of reconstructing level change timestamps based on local sampled data... The continuous motion model corresponding to the encoder value can also be used in this application, such as a piecewise polynomial model, a Hermite interpolation model based on velocity constraints, or other trajectory fitting models.

[0062] Figure 4 This is a schematic diagram of the wafer sensor calibration provided in this application. Figure 4 As shown, the wafer sensor calibration process may include a first calibration motion along the calibration station direction and a second calibration motion along the bias direction. Both the first and second calibration motions can be performed by the robot arm 110 carrying the wafer 160 through the detection areas of the first wafer sensor 130, the second wafer sensor 131, and the third wafer sensor 132, at which point the wafer 160 is the calibration wafer.

[0063] In the first calibration motion, the robot arm 110 can extend from the HOME position towards the calibration station and record the robot arm position when the calibration wafer 160 contacts the wafer sensor and the robot arm position when the calibration wafer 160 leaves the wafer sensor. Taking a certain wafer sensor as an example, if the angle corresponding to the calibration station direction is... When the robot arm contacts the sensor, its position is... The position of the robotic arm when it leaves the sensor is The wafer radius is The sensor position to be calibrated is Then, a system of circular equations can be constructed based on the radii relationship between the sensor position and the wafer center position:

[0064] Solving the above system of circular equations yields several first candidate locations for the wafer sensor. However, because the circular equations have symmetric solutions, these first candidate locations cannot usually be directly determined as the true calibration locations.

[0065] During the second calibration motion, the robot arm 110 can rotate the offset angle at the HOME position. Then, along a path with an offset angle relative to the first trajectory. The second trajectory extends to the calibration station. Offset angle The offset angle can be determined based on the robot arm configuration, wafer size, and wafer sensor mounting location. The angle is not limited to a fixed angle, as long as it allows the calibration wafer 160 to pass through the detection area of ​​the sensor on the wafer to be calibrated during the second calibration movement. During the second calibration movement, if the robot's position is [missing information - likely a specific position or location] when contacting the sensor... The position of the robotic arm when it leaves the sensor is Similarly, a system of circle equations can be constructed and solved to obtain multiple second candidate positions.

[0066] Since the wafer sensor is mounted in a fixed position, the actual sensor position should remain consistent in the robot's coordinate system regardless of whether the robot arm 110 moves along the calibration station direction or the offset direction. Incorrect candidate positions will experience significant geometric displacement due to changes in the direction of movement. Therefore, the first candidate position obtained from the first calibration movement can be represented as... and The second candidate position obtained from the second calibration motion is represented as and And calculate the Euclidean distance between the two sets of candidate positions:

[0067] Pick .like The corresponding candidate positions are and Then, according to and Determine the calibration location of the wafer sensor. For example, it can be... This serves as the calibration position of the wafer sensor in the robot's coordinate system. By utilizing the characteristic that the spatial position of the true solution is consistent across two trajectories, while the spatial position of the erroneous solution deviates, the true calibration position can be selected.

[0068] In some embodiments, if the offset angle If improper settings cause a wafer sensor to fail to generate the preset number of trigger points during calibration, or if candidate results matching the wafer radius cannot be obtained based on the trigger points, the main controller 150 can output a calibration error alert and prompt the user to reset the bias angle. Or check the installation status of the wafer sensor.

[0069] Figure 5 This is a schematic diagram illustrating the wafer offset calculation provided in this application. Figure 5 As shown, the wafer can be a diced wafer 161. The diced wafer 161 passes through the detection positions corresponding to the first wafer sensor 130, the second wafer sensor 131, and the third wafer sensor 132 along the wafer movement direction. If a specific geometric relationship is formed between the diced edge of the diced wafer 161 and the detection position, the diced edge may pass through multiple detection positions simultaneously or approximately simultaneously, causing multiple wafer sensors to generate pseudo-trigger points f1, f2, and f3 with the same or similar timestamps. The pseudo-trigger points do not correspond to the actual triggering process of the wafer's arc edge. If they are directly involved in the offset calculation, they may form a pseudo-candidate result C2 with a large radius error.

[0070] Unlike spurious trigger points, the real trigger points p1, p2, and p3 correspond to the triggering results when the edge of the wafer arc passes through the detection positions of each wafer sensor. For each wafer sensor, the main controller 150 can obtain one or more sets of trigger points during the movement of the robot arm 110. For example, the set of trigger points corresponding to the first wafer sensor 130 can be represented as follows: The set of trigger points corresponding to the second wafer sensor 131 can be represented as follows: The set of trigger points corresponding to the third wafer sensor 132 can be represented as: .in, , and They can be the same or different. Under conditions such as obstruction, notch, edge cutting, or wafer offset, the number of trigger points generated by each wafer sensor may not be fixed.

[0071] The main controller 150 can be from , and Each trigger point is selected from the options to form a trigger point combination. Any trigger point combination can be denoted as... The calibration positions of the first wafer sensor 130, the second wafer sensor 131, and the third wafer sensor 132 are respectively denoted as... , and The wafer offset is denoted as... The wafer radius is denoted as .

[0072] In the embodiments of this application, wafer offset The goal of the solution is to make the trigger point compensated Then, the calibration position of the corresponding wafer sensor satisfies the wafer radius constraint. Specifically, for the first... A wafer sensor, if the trigger point is The calibrated position is ,but This can be understood as the wafer center position determined jointly by the robot arm position and the wafer offset at the trigger moment. Since the wafer edge corresponds to the sensor's detection position when the wafer sensor changes voltage, the distance from the wafer center to the sensor's calibrated position should be equal to the wafer radius. This geometric relationship can be expressed as:

[0073] For three wafer sensors, we can obtain:

[0074]

[0075]

[0076] in, It can be understood as the first Each trigger point is used to compensate for wafer offset. The observation radius obtained afterwards. If If the trigger originates from the edge of a real circular arc, the observation radius should be close to the standard wafer radius. ;like False triggers from edges, notches, or robotic arm obstruction may cause the corresponding observation radius to deviate significantly. .

[0077] Since the combination of trigger points may contain spurious trigger points, the above radius constraints may not be satisfied simultaneously. Therefore, the wafer offset can be... The solution is transformed into a least squares problem. In one embodiment, it can be solved with the following objective function. :

[0078] in, , This indicates the radius error of the corresponding trigger point combination. The smaller the value, the more the combination of trigger points conforms to the geometric relationship of triggering at the edge of the wafer arc.

[0079] In another embodiment, a linear equation can be constructed first by establishing the equality of radii between any two sensors. Let Then there is For any two trigger points and , can be made as:

[0080] After expanding and organizing, we can obtain information about Linear relationship:

[0081] Three point pairs can be constructed from three trigger points, for example , and Then construct a system of linear equations. .because The offset vector is two-dimensional, and the three point pairs provide three equations, which can be an overdetermined system of linear equations. The main controller 150 can be solved using the least squares method. ,For example:

[0082] Alternatively, QR decomposition, singular value decomposition, or other numerical methods can be used to solve the problem. .

[0083] The main controller 150 calculates the wafer offset based on a certain combination of trigger points. Then, the radius error corresponding to the trigger point combination can be calculated. The radius error can be the sum of the squares of the differences between each observed radius and the standard wafer radius, or it can be the sum of the absolute differences between each observed radius and the standard wafer radius. The main controller 150 can iterate through... , and The trigger point combinations are used to solve for the wafer offset for each combination. It also calculates the radius error, selects the trigger point combination with the smallest radius error, and sets the corresponding wafer offset. This is determined to be the final wafer offset.

[0084] like Figure 5 As shown, if a combination of trigger points includes pseudo-trigger points f1, f2, and f3 on the tangent edge, the pseudo-candidate result C2 corresponding to this combination may produce a large radius error and should be discarded. If a combination of trigger points is mainly formed by real trigger points p1, p2, and p3 on the arc edge, the candidate result C1 corresponding to this combination may produce a small radius error and can be retained. It should be understood that, in the embodiments of this application, C1 and C2 are used to illustrate the candidate results or equivalent circumferential relationships corresponding to different combinations of trigger points, and do not mean that the wafer offset must be solved by directly fitting the circumcircle of the trigger points. Thus, the influence of pseudo-trigger points caused by tangent edges, notches, or occlusions on the wafer offset calculation can be reduced.

[0085] After obtaining the final wafer offset Then, the main controller 150 can Compare with the preset maximum correction distance. If The module length is not greater than the preset maximum correction distance, and the main controller 150 can be based on Generate compensation motion commands for the robotic arm 110 and control the robotic arm 110 to perform correction actions to bring the wafer placement position close to the target position. If the wafer offset exceeds the preset maximum correction distance, the main controller 150 can determine that the current wafer offset exceeds the safe correction range, control the robot arm 110 to stop moving, and output an error message. The error message may include a display reminder, an audible and visual reminder, a control system error code, or abnormal information uploaded to the host computer.

[0086] Figure 6 A schematic diagram of the control device structure provided in this application. Figure 6As shown, the control device 200 may include a processor 210, a memory 220, and a communication interface 230. The communication interface 230 can communicate with both the driver 120 and the input acquisition unit 140. The processor 210 can obtain the encoder value, speed value, and sampling time fed back by the driver 120 through the communication interface 230, and can also obtain the time stamp of the level change recorded by the input acquisition unit 140 through the communication interface 230. The memory 220 can store computer programs, and when the processor 210 executes the computer program, it can perform the wafer correction method described above.

[0087] In one embodiment, the control device 200 can serve as... Figure 1 The main controller 150 is located in the control device 200, which can be a motion controller, an industrial computer, or an embedded controller, and communicates with the driver 120 and the input acquisition unit 140 via the communication interface 230.

[0088] In some embodiments, memory 220 may include volatile memory and / or non-volatile memory. The computer program stored in memory 220 may include a data acquisition program, a timestamp parsing program, a local sampling data extraction program, a code disk value reconstruction program, a sensor calibration program, a wafer offset calculation program, and a spin correction control program. When processor 210 executes the above programs, it can achieve coordinated control of the robot arm 110, driver 120, input acquisition unit 140, and multiple wafer sensors in the wafer transfer device.

[0089] It should be understood that, in the embodiments of this application, the number of multiple wafer sensors can also be increased according to actual detection needs. For cases with more than three wafer sensors, a combination of trigger points that meets preset combination conditions can be selected from the set of trigger points of multiple sensors, and the wafer offset and radius error can be calculated based on the corresponding sensor calibration position. For different wafer specifications, different wafer radii can be pre-stored. Different maximum correction distances and different sensor calibration positions are used, and the corresponding parameters are called according to the wafer specifications before performing correction.

[0090] It should be understood that, in the embodiments of this application, the input acquisition unit 140 records the time stamp of the level change. Recording can be performed based on rising edge, falling edge, or double edge triggering. This includes timestamps of multiple level changes generated by the same wafer sensor during a single spin correction detection motion. The main controller 150 can reconstruct the code disk values ​​separately and obtain multiple trigger points. The main controller 150 can determine which trigger points participate in the final wafer offset calculation based on the trigger point combination filtering results.

[0091] It should be understood that, in the embodiments of this application, the local time window can be adjusted according to the speed, acceleration, sampling period, and computing power of the robot arm 110. For example, under relatively stable motion conditions, fewer sampling points can be used to construct a continuous motion model; under conditions with significant acceleration and deceleration, more sampling points can be used to construct a continuous motion model. Preferably, the local time window covers the timestamps of level changes. The sampling period is set to the current sampling period plus two sampling periods before and after it, in order to balance computational efficiency and trajectory reconstruction accuracy.

[0092] It should be understood that, in the embodiments of this application, wafer offset It can be represented as a two-dimensional translation vector, or, if the robot arm 110 has rotational degrees of freedom, it can be further extended to an offset parameter including translational and angular components. For applications requiring compensation for rotational errors, the wafer center position and attitude angle can be jointly estimated by combining more trigger points or more sensor information. In the current implementation, a two-dimensional translational offset is used. This example is used for illustration, and it does not exclude the possibility of extending the form of the offset parameter according to the equipment structure.

[0093] In summary, this embodiment of the application obtains the time stamp of the level change when the output signal of the wafer sensor changes level through the input acquisition unit 140. and utilize coverage The local sampling data is used to reconstruct the corresponding encoder value and robot position, allowing the trigger point to be determined without relying on the driver 120 to directly latch the position at the moment of triggering. Furthermore, the reliable calibration position of the wafer sensor is obtained through offset trajectory calibration, and the wafer offset is calculated through trigger point combinations, wafer radius constraints, and radius error screening. This can reduce the impact of spurious triggering factors such as edge trimming, notches, or obstructions on the correction results, thereby improving the accuracy, robustness, and adaptability of wafer correction to equipment modifications.

[0094] The above embodiments are merely illustrative of the technical concept and features of this application, intended to enable those skilled in the art to understand the content of this application and implement it accordingly, and should not be construed as limiting the scope of protection of this application. It is obvious to those skilled in the art that this application is not limited to the details of the above exemplary embodiments, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of this application is defined by the appended claims rather than the foregoing description, and thus all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within this application.

Claims

1. A wafer bias correction method, characterized in that, include: Obtain the calibration positions of multiple wafer sensors in the robot's coordinate system; During the process of the robotic arm carrying the wafer for spin correction and detection, the encoder value, speed value and sampling time of the robotic arm drive motor are collected according to the sampling period; The system monitors the output signals of multiple wafer sensors and acquires a level change timestamp when the output signal of any wafer sensor changes level. The time accuracy of the level change timestamp is higher than the time accuracy of the sampling time, and the two timestamps have the same time reference. Local sampling data is extracted from the collected encoder values, speed values, and sampling times, and the time range of the local sampling data covers the timestamp of the level change; Reconstruct the code disk value corresponding to the level change timestamp based on the local sampled data; The robot position corresponding to the level change timestamp is determined based on the reconstructed code disk value, so as to determine the trigger point of the wafer sensor corresponding to the level change timestamp; The wafer offset is calculated based on the calibration positions of multiple wafer sensors and their corresponding trigger points, and the robot arm is controlled to perform correction.

2. The wafer bias correction method according to claim 1, characterized in that, The level changes include the output signal level changes caused when the wafer edge enters the detection area of ​​the wafer sensor, and / or the output signal level changes caused when the wafer edge leaves the detection area of ​​the wafer sensor.

3. The wafer bias correction method according to claim 1, characterized in that, Local sampling data is extracted from the collected encoder values, speed values, and sampling times, including: The sampling period to which the level change timestamp belongs is determined based on its position in the sampling time sequence. Based on the sampling period to which the level change timestamp belongs, determine the local time window covering the level change timestamp; Extract the encoder value, speed value, and sampling time within the local time window to obtain local sampling data.

4. The wafer bias correction method according to claim 3, characterized in that, The local time window covers the sampling period to which the timestamp of the level change belongs, and the two sampling periods immediately before and after that sampling period.

5. The wafer bias correction method according to claim 1, characterized in that, The code disk value corresponding to the time stamp of the level change is reconstructed based on local sampling data, including: A continuous motion model of the robot's drive motor is constructed based on local sampling data; Substituting the level change timestamps into the continuous motion model yields the code disk values ​​corresponding to the level change timestamps.

6. The wafer bias correction method according to claim 5, characterized in that, The continuous motion model is a cubic spline interpolation model, which is constructed based on the encoder value constraints and velocity value constraints in the local sampled data.

7. The wafer bias correction method according to claim 1, characterized in that, The robot's position corresponding to the time stamp of the level change is determined based on the reconstructed encoder value, including: The movement position of the robot's drive motor at the time corresponding to the level change timestamp is determined based on the reconstructed encoder value; Based on the motion position of the robot's drive motor at the time corresponding to the level change timestamp and the robot's kinematic model, the robot's position corresponding to the level change timestamp is determined.

8. The wafer bias correction method according to claim 1, characterized in that, Obtain the calibration positions of multiple wafer sensors in the robot's coordinate system, including: The robot arm is controlled to carry the calibration wafer along the first trajectory past the wafer sensor. Based on the robot arm position corresponding to the triggering of the wafer sensor by the calibration wafer, the first candidate position is obtained. The robot arm is controlled to carry the calibration wafer along a second trajectory with an offset angle relative to the first trajectory, passing through the wafer sensor. Based on the robot arm position corresponding to the triggering of the wafer sensor by the calibration wafer, the second candidate position is obtained. The calibration position of the wafer sensor is determined based on the distance between the first candidate position and the second candidate position.

9. The wafer bias correction method according to claim 8, characterized in that, The first candidate position includes multiple first candidate solutions, and the second candidate position includes multiple second candidate solutions; The calibration position of the wafer sensor is determined based on the distance between the first candidate position and the second candidate position, including: Calculate the Euclidean distance between each first candidate solution and each second candidate solution; The first and second candidate solutions with the minimum Euclidean distance are determined as matching candidate solutions; The calibration position of the wafer sensor is determined based on the matching candidate solution.

10. The wafer bias correction method according to claim 1, characterized in that, The multiple wafer sensors include a first wafer sensor, a second wafer sensor, and a third wafer sensor. The detection positions of the first wafer sensor and the third wafer sensor are located on both sides of the movement trajectory of the wafer center, respectively, and the detection position of the second wafer sensor is located between the first wafer sensor and the third wafer sensor.

11. The wafer bias correction method according to claim 10, characterized in that, The wafer offset is calculated based on the calibration positions and corresponding trigger points of multiple wafer sensors, including: Obtain the trigger point sets corresponding to the first wafer sensor, the second wafer sensor, and the third wafer sensor respectively; Trigger points are selected from the trigger point sets corresponding to the first wafer sensor, the second wafer sensor, and the third wafer sensor, respectively, to form a trigger point combination; The wafer offset is calculated based on the combination of trigger points and the calibration position of the corresponding wafer sensor.

12. The wafer bias correction method according to claim 11, characterized in that, Based on the trigger point combination and the corresponding wafer sensor calibration position, the wafer offset is calculated, including: With the goal of satisfying the wafer radius constraint between the trigger point and the corresponding wafer sensor calibration position after compensating for wafer offset, the wafer offset is solved using the least squares method.

13. The wafer bias correction method according to claim 12, characterized in that, Based on the trigger point combination and the corresponding wafer sensor calibration position, the wafer offset is calculated, which also includes: Calculate the radius error for different combinations of trigger points; The wafer offset corresponding to the trigger point combination with the smallest radius error is determined as the final wafer offset.

14. The wafer bias correction method according to claim 1, characterized in that, Controlling the robotic arm to perform correction operations includes: If the wafer offset is not greater than the preset maximum correction distance, then control the robot arm to perform the correction action; If the wafer offset exceeds the preset maximum correction distance, the robot arm will stop moving and an error message will be displayed.

15. A wafer transfer device, characterized in that, Includes a robotic arm, actuators, multiple wafer sensors, an input acquisition unit, and a main controller; The robotic arm is used to carry the wafer for spin correction and detection movements. The driver is connected to the drive motor of the robot arm and is used to drive the robot arm to move and to obtain the encoder value and speed value of the drive motor. The plurality of wafer sensors are respectively disposed in the wafer bias detection area, and are used to output a signal with a level change when the edge of the wafer passes through the corresponding detection area; The input acquisition unit is communicatively connected to the plurality of wafer sensors and the main controller, respectively, and is used to monitor the output signals of the plurality of wafer sensors, and record the level change timestamp when the output signal of any wafer sensor changes level; The main controller is communicatively connected to the driver and the input acquisition unit, and is used to obtain the encoder value, speed value and sampling time collected according to the sampling period from the driver, obtain the level change timestamp from the input acquisition unit, and execute the wafer correction method according to any one of claims 1 to 14 based on the encoder value, the speed value, the sampling time, the level change timestamp and the calibration position of multiple wafer sensors in the robot coordinate system.

16. A control device, characterized in that, Used as the main controller of a wafer transfer device, the wafer transfer device includes a robot arm, a driver, multiple wafer sensors and an input acquisition unit, the input acquisition unit is used to monitor the output signals of the multiple wafer sensors and record the level change timestamp when the output signal of any wafer sensor changes level; The control device includes a processor, a memory, and a communication interface. The communication interface is used to communicate with the driver and the input acquisition unit respectively. The memory stores a computer program. When the computer program is executed by the processor, the processor obtains the code disk value and speed value fed back by the driver, the sampling time, and the level change timestamp recorded by the input acquisition unit through the communication interface, and implements the wafer spin correction method according to any one of claims 1 to 14.

17. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by the processor of the control device, causes the processor to obtain the code disk value and speed value fed back by the driver in the wafer transfer device, the sampling time, and the level change timestamp recorded by the input acquisition unit through the communication interface, and implements the wafer correction method according to any one of claims 1 to 14.

Images