Background and Customer Needs
In electronics manufacturing, ensuring precise glue dispensing before the PCB enters the soldering machine is crucial for product quality. A customer required an automated inspection system to verify the accuracy of glue application, as defects in dispensing could lead to product failures.
After glue is dispensed front dispensing machine, issues such as incomplete application, leakage, or unintended contact with video memory components may occur, requiring inspection. Previously, the inspection relied on manual monitoring, which was prone to human fatigue, inconsistencies, and a lack of traceable inspection data.
Challenges
- Manual Inspection Limitations: Operators needed to constantly monitor the process, increasing the risk of fatigue and misjudgment.
- High Staff Turnover: Different inspectors applied varying criteria, making standardization difficult.
- Lack of Traceability: Manual inspection did not allow for image storage or data tracking, complicating production management.
Solution & Key Technologies
Techman Robot’s cobot arm provided an automated inspection solution, integrating advanced vision and AI classification technology to ensure precise glue application detection. Effectively solving the above drawbacks and achieve automated workflow.
- Imaging & Detection
- TM AI Cobot were deployed to inspect glue application on PCBs before the soldering machine process
- The cobot arm detected correct and incorrect dispensing, distinguishing between “OK” (properly applied glue) and “NG” (glue missing or misplaced)
- All images and inspection results were recorded in a database for traceability and process optimization
- AI Model Training
- Training Time: Approximately 10 minutes
- AI Function Used: AI Classification
- Dataset Size: 130 OK images, 80 NG images
- OK Products: Passed inspection and continued to the next production stage
- NG Products: Flagged for rework, preventing defective units from advancing in the process
- Performance
- Inspection Speed: 108 visual tasks completed in just 30 seconds, significantly improving efficiency.
- Achieved 99.9% inspection accuracy
- False alarm rate of less than 0.1%
- Overkill rate of less than 0.1%
Application Scenarios
- Automated verification of glue dispensing before the PCB enters the Soldering machine
- Identification of missing or misplaced glue to prevent product defects
- Real-time inspection with automated data collection for quality control
Benefits
- Enhanced Accuracy & Consistency
- Eliminated human error by automating the inspection process.
- Achieved 99.9% accuracy with AI classification, ensuring defect-free products.
- Increased Efficiency
- Fully automated inspection reduced reliance on manual operators.
- Processing time was significantly shortened, enhancing production speed.
- Improved Traceability
- AI-powered inspection stored all test results and images for quality management.
- Barcode reading and serial number tracking enabled efficient production monitoring.
Conclusion
The cobot successfully transformed PCB glue dispensing inspection by integrating smart automation, AI-driven classification, and advanced vision technology. This solution not only enhances accuracy and efficiency but also provides valuable traceability data, ensuring higher manufacturing standards and lower operational costs. The case highlights how manufacturers can leverage AI to optimize production lines and maintain superior quality control.