Nowadays the manufacturing of PCB has developed into a new stage where HDI and ICS is widely used, the process now is much automated other than human involvement. The next stage could be more complicated and the defect inspection might face new challenges--but you have to check: a defect could result in a whole PCB useless. So PCB manufacturers are now testing to use AI to optimize their production procedures and finally the whole PCB making flow.
The PCB making normally rely on experts that with years of experiences, they know every single step in the whole process and know how to make use of their knowledge to optimize production flow and increase production rate. One factor that restrict the increase of efficiency is man-made mistakes: mis-operation or just getting tired. If a operator over reacted with a small defect in a PCB that may reduce the good-ratio in production. Now we introduce AI to the manufacturing process, the device/equipment can be capable of the job by "study" cases, then human can focus on those more complicated tasks. The combination of human and AI can increase the overall efficiency and operation in PCB making, it is the best chance for AI technology.
AI and Industry 4.0
The ultimate trend of PCB making is a plant with full industry 4.0 system. This system adopt AI technology in global and manufacturing level. The global level means not just manufacturing, but including all systems and flow in the plant. Industry 4.0 could provide automation and data exchange basic function, it can realize real time production analysis, both way communication and data sharing, tractability and data analysis on demanding. At any specific factory, AI can help to optimize the production flow using the data from all manufacturing systems and equipment. It benefits the factory because all data are collected at global level and it can help to adjust and optimize the set ups and achieve highest production rate and yield. In a few years, it will not need man operator help and a truly automated factory can be in real.
This totally new PCB manufacturing mode requires all systems in PCB making is connected and shared, using AI as the supervisor and decision role. Currently there are varies challenges that restrict PCB factories achieve 100% automation but the AI technology has been added in separate units as possible as they can--such as AOI solution. The benefits of changing to AI including it can find and track to the reason why this defect happens and then it can automatic correct the relevant manufacturing process--to solve the issue from the top.
The subset of AI, including machine learning and deep learning, it will help a PCB factory be 100% automated. The machine learn feature can help the robot learn from existing data and analysis how to achieve better performance without input programme. For PCB making, machine learning can help to increase volume of production and to help factory make a better management in stocks and supply chain.
The implement of machine learning and deep learning can provide the insights that human can not achieve, it will explore to the area that human ignore or unable to see. Using the sensors and data system of Industry 4.0, to the entire PCB making process, from simply reading and writing to advanced tracking of parameters, as small as the minimum PCB units, the data can be created and studied globally. The AI system can keep running at all time without stop and rest, so the optimize action can be taken 24*7 hours on job.
AI in AOI
There is a considerable impact on using AI in AOI of PCB making. The human made mistake is reduced at large scale. The AI can detect and recognize short circuit and break circuit, even over plated copper. It can detects very minor defects while these defects are not visible at human work, or missed by human eyes--that is a natural result by repeating works.
Without AI, normally we will find 20~30 defected boards for every 100 PCBs, among 75% of them are false alert. But the rule is to check every possible defect 100%, that means a lot of time was wasted in the false defects. Yet in the correction procedure, it may hurt PCB again. By install machine learning on AOI system, the false defects are largely reduced thus the maintenance work is reduced too. Less false alert means less work on the board, higher efficiency. Also the AI can classify the defects in specific categories which is free from human thinking restriction, the false alert rate can be reduced as much as 90%. In the meanwhile, AOI procedure is the most labor intensive step, using AI can reduce labor cost apart from a higher work result.
Challenges
Although AI is developing at a great speed, but the challenge of PCB making is increasing in the meanwhile, or even faster. There are two major problems in defect detection: one is flex material and another is geometrical shape for smaller circuit design. The composited material of next generation such as LCP and MPI has given new problems to manufacturers of imaging recognizing, processing, shape track and slimmer lines. Flexible PCB could be a major types of PCB benefited from the use of AI, because the manufacturing will be done in a more strict scope.
Another major type of PCB is HDI that can be benefited by AI, as HDI PCB normally requires narrower circuit and and irregular shape and strict parameters. This made the defect exam harder for human workers and by machine it can be both promising and challenging in the same time.
Considering all those un-predictable challenges of PCB making, the factories with AI technology will play a key role in the future.
Inno Circuits, as the biggest PCB manufacturer in western China, welcome your inquiry at all time.
Comments