PCB manufacturing embraces AI: Part One

PCB has evolved from the bulky and outdated "printed circuit board" in the past to the fine-line design used in today's high-density interconnection (HDI) PCB board and IC substrate (ICS). The manufacturing process has also changed from manual assembly to highly automated production. With the further development of manufacturing technology, the manufacturing process has become more and more complex and sophisticated. It is now possible to check and repair defects to avoid panel scrapping. Today, the PCB manufacturing industry is facing historical opportunities to use artificial intelligence (AI) to optimize the production process and the entire PCB manufacturing facility.

PCB manufacturing usually relies on experts who have accumulated many years of experience. They are very familiar with and understand every step of the manufacturing process and know how to use professional knowledge to optimize production and increase yield. However, human limitations, such as errors and fatigue, can hinder the smooth progress of the work of experts. Operator error, or misjudgment of PCB defects (i.e. false alarms), may affect the yield due to excessive processing, and even damage the PCB itself. If AI is integrated into the manufacturing process (pic 1), machines will add value by taking over certain "learned" tasks, and human experts can continue to undertake more complex tasks that require thinking and interaction, while optimizing and "training" Artificial intelligence system. The combination of humans and artificial intelligence has improved overall efficiency and operations. This is the biggest opportunity for AI expert systems.

The future PCB manufacturing plant will fully integrate the Industry 4.0 system, with artificial intelligence at both the global level and the manufacturing system level. The "global" level includes all systems in the entire factory, not just a single manufacturing system. Industry 4.0 provides an automation and data exchange infrastructure that enables real-time production analysis, two-way communication and data sharing, traceability, and on-demand data analysis. In any particular factory, AI can improve the manufacturing process by using data obtained from various manufacturing systems and machines, as well as data collected through Industry 4.0 mechanisms such as traceability and two-way communication. AI benefits the factory by analyzing massive system-level data, optimizing factory setting parameters and achieving the highest level of productivity and yield. AI analysis and self-learning are achieved through artificial neural networks. A truly fully automated factory will be established within a few years, completely eliminating manual intervention.

This new PCB manufacturing model requires all systems in the factory to be fully connected and to adopt AI-enabled monitoring and decision-making mechanisms. There are some patents and technical challenges that limit the PCB factory to achieve full automation, so AI is now applied in a single system as much as possible, such as automatic optical inspection (AOI) solutions. The use of global AI models in production facilities has many advantages. For example, PCB defects can be reported more reliably, that is, "real defects", the root cause of the problem can be found through the feedback loop, and the factory process can be automatically modified to eliminate related defects.

A subset of AI, including machine learning and deep learning, will enable PCB factories to move towards full automation. Machine learning uses algorithms that allow computers to improve task performance through data and examples that have been collected and learned from them, without the need for explicit programming. As far as PCB manufacturing is concerned, machine learning can increase yield, improve manufacturing operations and processes, and reduce manual operations, thereby helping to more effectively handle factory assets, inventory and supply chains.

Deep learning takes AI to a more complex level, which is very useful at the global factory system level. Deep learning is similar to the ability of the human brain to use multi-faceted, multilayer artificial neural networks for learning, understanding, and inference. In the PCB factory, the software expert system learns effectively from the understanding of the collected data, the pattern and the complex expression of the environment. This learning becomes the basis for automatic improvement processes in PCB manufacturing.

The realization of machine learning and deep learning provides PCB manufacturers with the ability to surpass human understanding. In other words, artificial intelligence systems find new opportunities for improvement by digging deeper into areas that humans may not even think of. The AI expert system is very efficient. It additionally uses some more complex parameters to monitor the factory system at a global level, reducing the number of human experts that may require, improving efficiency and obtaining best practices.

From simple read and write functions, to advanced tracking of manufacturing process parameters, to the smallest PCB unit, in the entire PCB manufacturing process, Industry 4.0 sensors (data sent from the device) and systems can generate global-level data. Manufacturing process parameters may include etching, resist development, and chemical material concentration. Using deep learning can analyze this type of data to optimize manufacturing methods and parameters, identify patterns, and make informed decisions about changes required in the process. All of these can be performed automatically 24 hours a day, 7 days a week without interruption.