EIPC’s Winter Conference in Lyon, France: Day 2 Review

Estimated reading time: 7 minutes

Additional Reads:
EIPC’s 2018 Winter Conference in Lyon, Review of Day 1
Part 2: EIPC’s 2018 Winter Conference in Lyon, Review of Day 1

The first day of the conference had started and ended in the dark—a long and technically intense day! After a convivial conference dinner and a good night’s sleep we were back on the bus, this time in daylight at the slightly later time of 8:00 a.m., for the journey from downtown Lyon back to Alstom’s conference facility in Villeurbanne, and it wasn’t raining. In fact, there was even some sunshine later in the day!

Nine presentations in two sessions this second day, the first session on process improvements introduced and moderated by EIPC board member John Fix, manager and director of marketing and sales at Taiyo America.

His first presenter was Steve Payne, manager of European Operations for iNEMI, the International Electronics Manufacturing Initiative, who discussed future PCB fabrication and material requirements for the global industry segments. He explained that iNEMI was a not-for-profit, R&D consortium of approximately 90 leading electronics manufacturers, suppliers, associations, government agencies and universities. An important function of iNEMI was to roadmap the future technology requirements of the global electronics industry, to identify and prioritise the gaps in technology and infrastructure, and help to eliminate those gaps through timely, high-impact deployment projects.

The iNEMI Technology Roadmap, updated every two years, provided focus and direction to the electronics supply chain on technology trends and challenges in manufacturing rigid, flexible and optoelectronics substrates for the next ten years. The roadmap considered seven market sectors: aerospace and defence, automotive, high-end systems, IoT, medical, consumer and office, and portable wireless, each of which had specific requirements for future PCB technologies and materials, including miniaturisation, durability in harsh environments, increasing signal speeds, high density interconnect and other factors which were often interrelated. The iNEMI Technical Plan originated from a GAP analysis of the Technology Roadmap, with the objective of predicting and quantifying the technical requirements of a PCB over a ten-year period for each sector.

Payne examined in detail the issues associated with miniaturisation and high-density interconnect, embedded components, optical PCBs, flexible and stretchable circuits, and the role of Industry 4.0 and the Industrial Internet of Things in the PCB fabrication supply chain.

He commented that, historically, PCB fabrication had generally been undertaken by large electronics OEMs with substantial R&D resources, whereas it was now mostly outsourced to a few large and many small-to-medium sized companies, most of which had limited resources for development work. Consequently, they tended to rely on materials and equipment suppliers for incremental improvements to their technical capability. Payne believed that PCB fabricators could benefit from participation in collaborative projects to leverage collective expertise and involve the complete supply chain.

Using raw data to develop an intelligent manufacturing solution for the problem of registration control within the PCB smart factory was the subject of the presentation of Andrew Kelley, CTO of XACTPCB and a long-standing expert in the management of the PCB registration process. He began his discussion of smart manufacturing by quoting Arthur C. Clarke: “Before you become too entranced with gorgeous gadgets and mesmerising video displays, let me remind you that information is not knowledge, knowledge is not wisdom, and wisdom is not foresight. Each grows out of the other, and we need them all.”

He explained that the future capability of a PCB plant would depend upon an ability to collect, collate, integrate and understand process data, and the key to registration control was to develop lots of disconnected raw data into an intelligent manufacturing solution. Data was a strategic business asset, but the real value was in the information, insight and actions derived from analysis of the data. And without data, it was not possible to derive the information required to make decisions and improve products and processes. Data was often collected within factories but its true value was rarely realised. Kelley emphasised the point that all data related to past events, whereas all actions affected the future. And data tended to be used more to provide management reports than to benefit manufacturing requirements. Connected data was fundamental to the concept of the smart factory, with machine networking and automated data capture essential. The goal of manufacturing intelligence was to capture, collate, integrate and understand process data, and to provide a bridge between people generating data and people using it.

Focusing specifically on the optimisation of multilayer registration, he asked: “For a single-lamination, 8-layer product how many different material/process combinations could be involved?” In fact, there were more than 40 million potential combinations without considering the options of copper thickness, copper pattern or process routes. It was not feasible to acquire the amount of data required to always know which innerlayer scale factors to use, and historical data alone would never be able to provide all the answers required.

Kelley used the pictorial analogy of building with LEGO® bricks: starting with a foundation of data, building information based on data, building knowledge based on information, building wisdom based on knowledge, and taking action based on wisdom, in a continuous measure-learn-predict cycle. Machine learning was a statistical process that took accumulated data and used it to derive algorithms that explained the data and could be used to predict future data. Self-learning, autonomous systems were becoming a reality in the field of manufacturing due to the availability of data, improved machine learning and algorithms, and more powerful computers.

Page 1 of 2

• Next Page >