close

AAEON’s MAXER-2100 as Cornerstones of Integration

The effectiveness of AI-assisted quality inspection solutions relies not only on the inferencing models used, but also on how effectively the hardware deployed within quality inspection solutions can process, execute, and optimize the software used.

One vendor deployed AAEON’s new MAXER-2100 as an AI inference server within a comprehensive quality inspection solution, the Hawkeye. The reasons for this were due to the MAXER-2100’s ability to run the company’s inferencing software and use the results of the solution’s subsequent quality reports to retrain and optimize the model to increase the accuracy and reliability of the process.

Challenges in Traditional PCBA Manufacturing Processes

AOI systems with high false alarm rates cause inefficiency and inconsistency.

Manual inspection is time-consuming and can lead to slower production.

Manual inspection can lead to eye fatigue and increased rates of human error.

Scaling production often requires increased manpower, resulting in higher operational costs.

Inconsistent manual inspection standards across different operators can lead to variations in defect detection.

Factory labor shortages pose challenges in meeting production demands while maintaining inspection standards.

How the MAXER-2100 Solved these Challenges

Utilized multiple GPUs to accelerate AI inferencing performance, resulting in the identification of component defects over 99% of the time.

Reduced the time needed for defect analysis from between 2 and 3s per photograph via manual inspection to just 0.05s per photograph.

Enhanced detection accuracy by optimizing the vendor’s inferencing model via machine learning, based on data acquired while in operation.

Reduced labor by 67% through the removal of

elementary inspection tasks from employee duties.

Achieved an impressive 80% defect detection rate

with just 5 photos during training.

AAEON’s HawkEye: A Tried and Tested Solution

The process architecture of the Hawkeye solution includes the following main steps:

1. The PCBA product undergoes an initial inspection by the AOI system, identifying potential defects.

2. The AOI system sends defect images to the MAXER-2100 AI inference server.

3. The MAXER-2100 utilizes GPU acceleration for AI image recognition, swiftly and accurately analyzing the defect types.

4. Based on the AI analysis results, it automatically determines if manual review is needed, significantly improving inspection efficiency.

5. Simultaneously, the analysis data is fed back into the AI model, continuously optimizing the accuracy of defect recognition.

How AAEON Can Benefit Your Quality Inspection Process

Enhance Efficiency

Process high quantities of data faster and more efficiently, with both the MAXER-2100 and BOXER-6843-ADS accommodating 12th and 13th Generation Intel® Core™ processors and up to 128GB of Dual-Channel DDR5.

Improve Inferencing Accuracy

Both the MAXER-2100 and BOXER-6843-ADS can host multiple NVIDIA® GPU cards capable of running sophisticated AI defect detection models.

Introduce Machine Learning for Continual Improvement

Through their high capacity to process data and use it to run inferencing software within quality inspection processes, the MAXER-2100 and BOXER-6843-ADS are able to retrain and optimize your AI model based on the information received during operation.

Our Offerings

MAXER-2100

The MAXER-2100 is a 2U Rackmount Controller specifically crafted to serve as an inference server for solutions demanding the utmost precision. Engineered to fine-tune AI models and execute intricate AI algorithms for AOI in chip and PCBA manufacturing, the MAXER-2100 is compact, expandable, and signifies a novel approach to enhancing industrial computing capabilities.

BOXER-6843-ADS

The BOXER-6843-ADS Fanless Embedded Controller harnesses 12th and 13th Generation Intel® Core™ processors, offering flexible system memory, storage, and expansion options for industrial applications. Through innovative engineering, the BOXER-6843-ADS facilitates seamless integration across industrial environments, bridging legacy infrastructure with cutting-edge technologies.