Towards Smarter Use of Data across the Electronics Supply Chain
By Paula Doe, SEMI
Machine learning, artificial intelligence, and predictive analytics are changing electronics manufacturing, just like they’re changing everything else. That means significant opportunities to speed time to yield and profits with better traceability across the supply chain, predictive simulation, and other ways to get key information out of all that data — but only if the sector can find better ways to share the needed data without compromising intellectual property (IP). Experts will tackle this topic at one of the SEMICON West 2017 Advanced Manufacturing sessions “Next Generation Smart Manufacturing: Collaboration towards Enabling Practical Predictive Analytics and Traceability across the Electronics Supply Chain” on July 11 at Moscone Center in San Francisco, Calif.
Simulation starts to speed fault detection, improve yield
The availability of massive low cost compute power and sophisticated analytics are also starting to enable a whole new suite of tools to predict yield. “We have almost an infinite number of variables with extremely complicated interactions that impact the processing of a device, but it is now becoming possible to predict how a change in the input from a particular sensor will change the final device,” says Timothy Miller, Applied Materials deputy director, Controls Marketing, a speaker in the SEMICON West program. “We’re starting to be able to create models of the impact of the sensor data that can speed fault detection by an order of magnitude.” This takes combining the customer’s device knowledge with the tool suppliers’ knowledge of the equipment process and failure modes across multiple customers, and balancing what can be automated and what needs human expertise. It’s opening the doors to new approaches such as running our analysis on data at the customer site, or even cloud solutions to get the value out of the data that would have been unheard of before.
“We really need to better connect the whole chain from design through production, since reducing the errors in the design stage leads to reducing defects in manufacturing,” notes Tim Hewitt, director of Industry Services, Siemens, who will discuss the “pain and gain” from connecting design from Siemens’ Mentor Graphics acquisition into the full smart factory automation system.
This growing ability to use data to predict outcomes and trace back to root causes of problems is driving demand for more access to more data across suppliers and customers along the whole electronics supply chain, while still securely protecting intellectual property. “We need security measures to support data partitioning but still allow collaboration between fabs, equipment makers, and increasingly even sub-component suppliers to figure out what went wrong, as customers like automakers demanding full traceability, down to the equipment that made the part,” notes Supika Mashiro. As manager of the System Development Center at Tokyo Electron and a co-chair of the Factory Integration team for the International Roadmap for Devices and Systems, Mashiro will also present at the SEMICON West session. She’s co-chairing this year’s update to the roadmap, which will address security issues for the first time. “The roadmap committee doesn’t have the solution, but we want to start the serious conversation on how to enable the needed collaboration without compromising IP,” she says. “We are in the process of learning, and are soliciting input to get a wide range of ideas for the best solutions to facilitate the advance of smart manufacturing in the IC sector.”
Genmark Automation president Carl McMahon, another speaker, will update on some other collaborative projects initiated by the SEMI Smart Manufacturing Advisory Council, including a project with McKinsey & Co. to identify common, industry-wide areas where smarter manufacturing approaches could most reduce cost of ownership as a collective reference, and then aim at a collaborative demonstration project.
The “Next Generation Smart Manufacturing: Collaboration towards Enabling Practical Predictive Analytics and Traceability across the Electronics Supply Chain” is on July 11 program at SEMICON West. Hosted by SEMI’s Smart Manufacturing Advisory Council, the program will also feature Brice Achkir, Distinguished Engineer at Cisco Systems; Hem Takiar, VP of Packaging & Assembly at Western Digital; Chao Jin, NSF/UCRC Center, Intelligent Maintenance Systems; and a panel discussion with the speakers on the potential solutions for sharing data without compromising intellectual property. Visit www.semiconwest.org for more information on all SEMICON West programs. Other upcoming SEMICONs in 2017 include: SEMICON Taiwan (September 13 - 15, 2017); SEMICON Europa (November 14 - 17, 2017); and SEMICON Japan (December 13 - 15, 2017).
June 20, 2017