Issue link: https://iconnect007.uberflip.com/i/1197000
JANUARY 2020 I SMT007 MAGAZINE 47 with our internal customers and peers, as well those who are designing new connectors to help explain how design features could impact the PCBA manufacturing process. Taking that knowledge and sharing it both internally and externally has been a good benefit. Going back to your comment about how a design engineer, or anyone for that matter, finds out how to engage with us. One of the things that we do globally is we affectionately refer to them as "tech days." We work with a customer, and they will invite us in and seques- ter a conference room or a foyer; we'll bring in subject-matter experts, like Jim and the team, with various skill sets and knowledge domains and have an open house/knowledge share. We often go on-site to our customer locations or do this at trade shows, either on the show floor or privately. The tech days have provided an inti- mate environment where the customer doesn't leave their location, and anyone who wants to come down to have a conversation with Molex is welcome and encouraged to do so. Johnson: For customers who are going to engage with Molex—either as a connector supplier or to plug into your services—what are the most important things that customers need to know after the tech days presentation? Hines: In terms of trending in the data center space, everything needs to be faster. The speed of circuitry, connectivity, and design is para- mount. The hyperscalers, edge computing, and traditional service providers are all looking for faster, so speed is one important factor; signal integrity would be another one. The last one that I will highlight is everything needs to be smaller, lighter, and lower cost, and have less of a footprint. When you combine all of those, it makes for a very challenging design process. Ruffino: My lens is in the reliability and perfor- mance of a product, and I think it's important for users to understand their use-case of the product, whether it's a connector or a solu- tion. Reliability testing could be designed and focused on that use case because if you over- specify, which is generally what's done, you end up running the costs up because you're testing it in much more severe environments than it's going to be used in. You add cost to balance testing of the product and reduce the likelihood of a product to survive those tests if it's being used in a new application. Often, we see those test requirements overspecified; they need to be sized correctly. Even more importantly, we need to incorporate all signifi- cant application physics-of-failure modes in a reliability test regimen. This includes SMT processing; not addressing degradation mecha- nisms or under testing could result in under- performance in the field. Hines: And as you said, Frank, that could drive costs, but then we've also seen that the application space is truly getting more severe, depending on the market, whether it be auto- motive or mobile products. Ruffino: They are, but we need to understand the severity of the application, the physics of failure, and a right-sized testing regimen. The application and environment should be cate- gorized to a point where we can simulate that use-case in modeling and empirical testing. Understanding the application environment and developing a test plan that's the correct fit or evaluating existing product data supports this goal. Rickett: That translates back into the SMT space where people start looking at more robust conformal coating materials, and then it drives the processes for depth-depositing those and so forth. Johnson: How do these dynamics influence the design considerations for connectors them- selves? Hines: When we meet with our internal design engineers, one of the things we try to stress is that there's more to a PCBA than the connector. Let's consider things like coplanarity, package miniaturization, more pin counts, and reduced terminal pitches on the board. For example, with a BGA component, if we make a connec-