Mask-Printing Parameter Optimization Through Stencil Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mask-printing processes are complex and require manual expert assessment to identify and correct errors, lacking guidance on parameter adjustments to prevent future errors.
Innovation Solution
A machine learning process is employed to analyze stencil images before and after printing, establishing relationships between process parameters and print quality, enabling automated optimization of parameter settings to improve print results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual expert assessment is used to identify and correct printing errors, then error detection capability is improved, but process complexity and time consumption increase
Solution Approach 1:
The patent replaces manual expert assessment with an automated image analysis system using machine learning algorithms. The system captures images of the stencil before and after printing, automatically analyzes them to detect printing errors, and provides guidance on parameter adjustments. This substitution eliminates the need for manual expert intervention while maintaining or improving error detection capability.
Solution Approach 2:
The printing system incorporates self-diagnosis and self-correction capabilities through automated image analysis. The system automatically identifies printing defects, determines their causes, and suggests parameter adjustments without external expert intervention. This self-service approach reduces process complexity and accelerates error correction.
2Manufacturing precision
If traditional SPI inspection steps are implemented to check print results, then manufacturing precision is improved, but productivity decreases due to additional process steps
Solution Approach 1:
The patent performs image capture of the stencil before the printing operation completes. By analyzing the stencil state in advance, the system can predict potential printing quality issues and provide guidance on parameter adjustments before the actual printing occurs. This preliminary action prevents defects rather than detecting them after printing, eliminating the need for subsequent SPI inspection steps.
Solution Approach 2:
The automated image analysis system enables skipping of traditional SPI inspection steps by providing real-time feedback during the printing process. The system captures and analyzes images rapidly, allowing continuous production without interruption for separate inspection steps, thereby maintaining manufacturing precision while improving productivity.
3Manufacturing precision
If multiple printing parameters are adjusted to optimize print quality, then manufacturing precision is improved, but the number of parameters to manage increases complexity
Solution Approach 1:
The patent implements a feedback mechanism where image analysis results are used to automatically guide adjustments of printing parameters. The system compares actual stencil images with reference images, identifies deviations, and provides specific guidance on which parameters to adjust and by how much. This closed-loop feedback simplifies parameter management by providing data-driven adjustment recommendations rather than requiring manual trial-and-error of multiple parameters.
Solution Approach 2:
The system focuses parameter adjustments on specific critical parameters identified through image analysis rather than requiring optimization of all possible parameters. By using machine learning to identify which parameters have the most significant impact on print quality based on actual stencil images, the system reduces the effective number of parameters that need to be managed while maintaining or improving manufacturing precision.
Data Source
AI summary
For determining optimized parameter values of a mask-printing process, a quality value for the substrate is determined in a testing process following the printing operation, and an image of the upper side of the template is generated. A relationship between the parameter, the quality value and the template image are ascertained, and the relationship is used to determine an optimized value for the parameter. In the printing operation, the optimized value is set, and the printing process is monitored by using the respective template image, dispensing with the need for the testing process.