Issue link: https://iconnect007.uberflip.com/i/1417991
36 DESIGN007 MAGAZINE I OCTOBER 2021 Shaughnessy: I've been wondering why fab- ricators don't just say, "We're only accepting one of the intelligent formats from now on." Acheson: If you look at Axiom, they created a policy that is unique in the industry: If you send us intelligent data, it's this price; if you send us unintelligent data, aka Gerber, it's going to cost you more. ey're seeing a huge increase in people sending them intelligent data. Nolan Johnson: at's an interesting situation because when we surveyed the designers, their response was, "I want to send what's easy, and I send what my manufacturing guy tells me to send." I don't want to sound like using Gerber is a problem, but the adoption of intelligent data has a lot to do with what seems to be iner- tia at the fab. Acheson: Gerber is a good format. It's tried and true, and failure is very rare. But as data, on the unintelligent side, I need a bunch of different pieces to get intelligence out of it. at's where 2581 comes in; all that intelligence is there at once. Johnson: Ed, if I can make this analogy, you'll get it, and Dan will too. Gerber is basically structured like Unix from the '60s. Shaughnessy: Right. If you liked Unix, you didn't want to switch to Windows. Acheson: Yes, that's true. Shaughnessy: One of the respondents in our survey said, "We output in all three formats and hope for the best." Really? I assume that's because they don't know where it's going to be fabricated. Acheson: Early on in our surveys, when the consortium was promoting the adoption of IPC-2581, we did an audience poll and asked, "How many of you in the group use strictly Gerber?" A good percentage, I'd say 60 or 70%, raised their hands. "How many of you send ODB++?" e remaining 30% raised their hand. ey said, "Of those of you who send ODB, how many of you send Gerber with it?" Of that 30%, 95% raised their hands. Whenever there was a conflict, guess what the answer was? Use Gerber. Shaughnessy: Does it seem to be a generational thing? Do you think the young people, these digital natives coming into the industry, are going to drive the adoption of "intelligent" data formats like 2581 and ODB++? Acheson: Yes, I think they will. We know the board designer generation is getting older, and the old "I don't like change" people are slowly stepping out of the scene. What we're seeing in our consortium is the younger people are more adamant about using it because of what they can have in it. e depth of understanding and the education comes through the different pushes that we do at the trade shows, the dif- ferent things we do in articles to promote it, and why it is better for the industry. We're see- ing a lot of traction in moving toward adoption of it. Holden: Our columnist Dana Korf is a big sup- porter. But Michael Ford, who's a big name in assembly, is pushing assembly-level IPC so- ware that is going to be based on IPC-2581. My hope is that it pushes it over the hill because you're going to need to use 2581. Acheson: Yes, there are a couple things that will push the adoption of 2581. One, as you mentioned, is CFX, the Connected Factory Exchange. CFX uses 2581 as the descriptor of that model. In fact, 2581 is being looked at as a digital twin of the physical board. So, the data for physical boards is based on what's in 2581. On top of the ability to push the assembly data, and the details needed for the CFX, in Revision C we're introducing a standard