Issue link: https://iconnect007.uberflip.com/i/1463464
18 SMT007 MAGAZINE I APRIL 2022 ica. You don't need $700 million; these are the kinds of things that you could and should be working on." Matties: Right. Step one is building the digital platform because you need to have that for all the other dependencies between equipment— your AI, your communication, and so on. Kelly: at's why all the "cloud wars" have been going on for 10 years or more. Everybody saw the cloud infrastructure as something we'd be needing. What's the Tipping Point? Matties: Matt, with an event like IPC APEX EXPO, you had your Factory of the Future set up on the show floor. What feedback came out of the show? Kelly: We see some nice trends showing up. For example, companies approached me regard- ing machine language and artificial intelli- gence within design. at's very good. But the quality of some other areas in Factory of the Future, I would classify as medium. We are still very much in an advocacy and education posi- tion right now. Johnson: Happy, can you recall an industry- wide compelling event in semiconductor, where the industr y said, "Ok ay, now we need to do this"? Happy Holden: Yes. It h a d t o d o w i t h t h e rapid increase in the complexity of semi- c o n d u c to r s . W h e n I was in college, we had six transistors per chip. Forty years later, we're now talking about bil- lions of transistors. e c o m p l e x i t y a n d t h e density pushes you over the edge, and we've discussed that with semi-additive technology and additive technology. We're at the geome- tries now in microns that the semiconductor guys used to be in. I think that's a tipping point for us. If you're going to do six- and seven-mil lines and spaces, you're probably not going to invest too much in digitization, but if you're looking at features in the microns range, it's a necessity. If you're going to start playing in the game of 30-micron, 20-micron, 10-micron, or 8-micron, then you're going to have to do some- thing different. e semiconductor industry did that because of complexity, and other fac- tors like predictive maintenance. Not preven- tive maintenance, but predictive maintenance, because the equipment is getting so expensive, you can't afford any downtime. at focused a lot of automation and digitization for the semiconductor guys because of the value. But then right aer that came the liquid crystal dis- play, then photovoltaic that adopted the same strategy. Dave brought up the example of a stencil. e logical next step with a stencil is that your AOI looks at the output of the solder paste step. en, a little further down, aer reflow, is there any tombstoning or something like that anticipated? rough various models, and uti- lizing the sensors and statistics, a cause and