Issue link: https://iconnect007.uberflip.com/i/1424540
NOVEMBER 2021 I SMT007 MAGAZINE 37 testing with salvage components would be important for every finished product with salvaged components. Neves: Sure. But, again, that's today's number. You really don't understand what the long- term reliability is because you're undergoing more rework procedures than you would nor- mally do. e board, the component, and oth- er things are being subjected to processes and heats—not necessarily saying they're bad— but they're just not typical for the process that people have qualified to. ey qualified a prod- uct to a specific process and then a bandwidth around that process. But now, with the com- ponent shortages, people are doing things outside that bandwidth and the best they can hope is do that functional test and say, "Yeah, it works today," but there's really a question of, "Is it going to work in one, five, or 10 years?" We really don't have that information, and no- body will invest the time, money, and energy to figure that out because this is likely a tempo- rary issue. ey will just bury it and deal with the problems as they come up. at's my opin- ion; I don't have any hard data to back that up, but that's just my take on things. Johnson: So, what questions should OEMs and EMS providers be asking? Neves: You need to ask the questions to assess the additional risk that's coming up from all these interesting things that are happening due to the supply chain issues. People are looking at secondary markets to source components. ey're changing components and all sorts of things that will obviously affect the product. But is it going to affect the product in a way that is an acceptable risk? at's the question to be asked. Johnson: ere's this move toward increased automation, smart factory, that sort of thing. Along with that is talk of increased inspection in the middle of the line as a part of the auto- mation. Is that real? Is that valuable? Is that un- necessary? Neves: e one thing that's hard to quantify is that when you have humans in the process, there's a lot of things they see that don't get quantified by automation. ey notice things. e human brain works in such a way that you can notice things happening that you may not be able to program into an AI, or into an auto- mated process of doing things. You're expect- ing things to go the same way when you have an automated process, and where you try to control the variables. en, you expect them to go the same way as before the automation and you have so many inputs, but a human can process a lot more inputs than just the cameras and other data input can. I think you lose some intuitiveness when you pull the human out of the process. So, the idea of having more inline inspections is to add intuitiveness back in. It's a double check just to make sure we didn't miss anything be- cause we're not covering everything a human would cover as far as inputs are concerned. Let's add more inspection to make sure that we verify that we're not losing something by taking away some of the inputs that the hu- man would notice. Johnson: I'm looking at some survey data here, on capital expenditures in this case, from the EMS providers. I asked them, "What is driv- ing your capital expenditure plans for 2022?" Tied for first place was market opportunities and additional capacity. Automation is in sec- ond place, followed by additional capabilities. So far, nobody has said that they're adding ad- ditional capabilities. It's all market opportuni- ties and capacity. Neves: Right, it's reducing cost and increasing revenue. Let me put my IPC hat on for a sec- ond and say that the IPC popped in with CFX at just the right time, and it's really taken off. I think that's an industry standard that has taken