Feedforward Wafer Stage Tuning for Faster Settling Across the Workspace
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wafer positioning systems in semiconductor manufacturing face limitations in motion control performance due to the reactive nature of feedback control, which is slow to react to errors, and manual tuning of feedforward parameters is time-consuming and not robust across different systems, requiring extensive expertise.
Innovation Solution
A method for automatically tuning feedforward controller parameters using a cost function that minimizes the difference between simulated and actual positioning performance, ensuring stable motion control across the workspace of the wafer positioning system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If feedback control is used to maintain wafer position, then the system can reject disturbance forces and track motion commands, but the reactive nature of feedback control limits tracking performance and response speed
Solution Approach 1:
The patent applies preliminary action by implementing feedforward control that anticipates required forces based on desired motion trajectories before errors occur. The controller proactively generates control signals based on known motion commands and system dynamics models, rather than waiting for position errors to develop. This allows the system to prepare corrective actions in advance, improving response speed while maintaining position accuracy through the combination of feedforward prediction and feedback correction.
2Manufacturing precision
If manual tuning of feedforward parameters is performed, then motion performance can be improved, but the process is time-consuming and requires extensive expertise
Solution Approach 1:
The patent implements self-service through automated feedforward parameter tuning that performs system identification and optimization without human intervention. The controller automatically executes identification maneuvers, collects motion data, fits dynamic models, and optimizes feedforward parameters using algorithms. This eliminates the need for manual tuning by experts, reducing tuning time from hours or days to minutes while achieving comparable or superior motion performance through systematic automated optimization.
Solution Approach 2:
The patent replaces the manual mechanical tuning process with an automated computational system. Instead of experts manually adjusting parameters based on experience and trial-and-error, the system uses computer-based system identification algorithms, mathematical modeling, and automated optimization routines to determine optimal feedforward parameters. This substitution of manual expertise with automated computational methods eliminates the time loss associated with manual tuning while maintaining high motion performance.
3Speed
If feedforward control is implemented to improve tracking performance, then proactive control action is achieved, but precise system knowledge is required for effective parameter selection
Solution Approach 1:
The patent applies feedback by implementing a closed-loop system identification process that uses actual measured motion data to refine the system model and optimize feedforward parameters. The controller executes identification maneuvers, measures actual position and motion variables, compares them with commanded values, and uses this feedback information to update the dynamic model and adjust feedforward parameters. This feedback mechanism eliminates the need for precise prior system knowledge by automatically learning system characteristics during the identification process.
Solution Approach 2:
The patent implements parameter changes by systematically varying feedforward control parameters during automated tuning to optimize tracking performance. The system identification process tests different parameter values, evaluates their effect on motion accuracy, and selects the optimal set of parameters that minimize tracking errors. This automated parameter optimization replaces the need for expert knowledge of system dynamics with a systematic computational approach that automatically determines the best parameters based on measured system behavior.
Data Source
AI summary
Methods and systems for automatically tuning parameters of a feedforward based controller employed by a wafer positioning system in semiconductor processing equipment are described herein. One or more feedforward controller parameter values are selected to provide a desired motion performance at any location in the workspace of the wafer positioning system. The tuned feedforward controller operates in combination with a feedback controller to provide stable control of wafer motion. In one aspect, feedforward controller parameter values are selected to minimize a cost function including a simulated positioning performance associated with a first set of feedforward controller parameter values subtracted from a sum of a simulated positioning performance associated with an updated set of feedforward controller parameter values and a measured positioning performance associated with the first set of feedforward controller parameter values. In some examples, positioning performance is characterized by settling time.