Hanul Semiconductor's AI MLCC Six-Sided Inspection Achieves 13,000 Units per Minute and 98% Accuracy

Hanul Semiconductor has achieved a testing speed of 13,000 units per minute and 98% defect detection accuracy with its artificial intelligence (AI)-powered six-sided external inspection equipment for multilayer ceramic capacitors (MLCCs).

According to Hanul Semiconductor on September 1, 2026, this performance test involved capturing images of all six sides of ultra-small MLCCs moving at high speed, with AI making real-time defect judgments. In the company’s in-house testing, defect detection accuracy was improved from approximately 95% to about 98%, while maintaining the inspection speed.

A 0603-size MLCC is an ultra-small component measuring approximately 0.6 millimeters (mm) by 0.3 mm. During inspection, images from multiple angles must be rapidly captured to identify micro-cracks, chipping, foreign substances, and electrode abnormalities. Especially in high-speed inspection environments, the technology that minimizes both missed and excessive detections while maintaining throughput is critical to equipment performance.

This equipment incorporates Hanul Semiconductor’s proprietary AI inspection platform, “Hawaii,” and its AI analytics engine, “Alohanet.” Hawaii utilizes a hybrid inspection structure, combining traditional rule-based visual inspection with AI deep-learning analysis. Defects that are consistent in position and shape are identified through conventional image processing, while micro-cracks and irregular surface defects that are difficult to judge are analyzed by AI.

The accumulated inspection data is used for AI learning, allowing the platform to further enhance defect judgment performance according to each customer’s product specifications and defect types. By implementing on-device AI, the system is being developed to enable AI inference directly within the equipment without external network connections, thus maintaining inspection speed and reinforcing production data security.

The company is currently working on additional development to increase the AI-powered inspection capability from 13,000 units per minute to 15,000 units per minute within this year. The scope of Hawaii’s application is also being gradually expanded to other in-house MLCC post-processing inspection equipment, such as complex measuring devices, indentation inspection equipment, and ultrasonic nondestructive inspection devices.

A Hanul Semiconductor representative stated, “This performance upgrade is significant in that it increases the inspection speed to 13,000 units per minute while further improving AI defect detection in actual high-speed production environments,” adding, “We plan to enhance equipment reliability by objectively verifying inspection speed and accuracy through assessments conducted by external specialized institutions.”

This content was produced with the assistance of AI translation services.