Multi-Head Semiconductor Positioning With Camera Offset Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In semiconductor manufacturing apparatuses with multiple processing heads, image distortion due to individual camera differences leads to uneven arrangement of processing objects, deteriorating the quality of semiconductor devices.
Innovation Solution
A manufacturing apparatus and method that includes multiple processing heads with cameras, utilizing a controller to calculate and cancel out optical and machine offsets by taking images of a reference and object, ensuring precise positioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple processing heads with individual cameras are used to perform processing concurrently on one substrate, then processing time is significantly shortened, but image distortion varies between cameras causing positioning variation and uneven arrangement of processing objects
Solution Approach 1:
The system captures images of a reference pattern with each camera, calculates the actual positions of reference marks based on image analysis, compares these with ideal positions, and generates correction values that are fed back to adjust the positioning of each processing head. This closed-loop feedback mechanism compensates for individual camera distortions and positioning variations.
Solution Approach 2:
The system changes the positioning parameters of each processing head based on calculated correction values. By adjusting position coordinates and orientation angles of processing heads according to measured deviations, the system compensates for camera-specific distortions and ensures uniform processing quality across all heads.
2Ease of operation
If individual cameras are used on each processing head for positioning, then each head can be independently positioned, but individual differences between lenses or imaging elements cause image distortion that differs for each camera
Solution Approach 1:
Each camera's image of the reference pattern is analyzed to detect deviations from ideal positions. These deviations are used to calculate correction values that are applied to adjust the positioning of corresponding processing heads, creating a feedback loop that compensates for individual camera characteristics.
Solution Approach 2:
The system uses a physical reference pattern with known geometry as a master copy for comparison. By capturing images of this reference pattern with each camera and comparing against the known ideal positions, the system creates a basis for calculating correction values that account for each camera's specific distortions.
Data Source
AI summary
A manufacturing apparatus of a semiconductor device includes: multiple processing heads provided apart from each other in order to perform predetermined processing in different positions on a common lead frame, the processing heads respectively including cameras; and a controller controlling driving of the processing heads, the controller positioning each of the processing heads at a position where at least an optical offset is canceled out. The controller takes an image of a paste of one island by the cameras and acquires optical inspection images respectively corresponding to the processing heads before the positioning, and calculates, as the optical offset, an amount of difference between the processing heads in a relative positional relationship between the island and the paste shown in the optical inspection images.