Issue link: https://iconnect007.uberflip.com/i/1417991
OCTOBER 2021 I DESIGN007 MAGAZINE 39 I thought he was going to tear the connector out of the back of his laptop trying to get the projector so that I could plug it into my machine and show him how it worked. I uploaded my 2581 data in my Allegro database, imported that 2581 into a Wise Viewer and showed him all the data and the detail in that format. He sat down in awe, and said, "I never thought I'd see this day." Karen McConnell was also elated to see somebody actually produced it. What made my week was walking the show floor with Gary Carter, who was very active with Fujitsu on this. Dieter came walking up with his wife and his entourage. He saw us, stopped, introduced us to his wife, and said, "ese are the guys that made my dream a realization. ey're my heroes." I mean, he couldn't have said anything better. I really miss the guy. Holden:: Nobody had the vision longer than he had. I only go back to '77 or so when I first met him but getting off Gerber onto an intel- ligent and interchangeable format was always his vision. Acheson: Well, that's one of the things about 2581; it was perceived as an achievable data- base that could be opened in any format. ere was an agreement amongst us soware developers at the early part that for the time being, in a gentleman's agreement, we would not import 2581 from each other to gain a competitive advantage. at worked well until a few years back, and I won't mention who it was, but someone started importing 2581 data and then we said, "Well, okay. All bets are off." Holden: at reminds me, Ed, since you're with Cadence, why are the semiconductor guys talking about Gerber files now? I was really blown apart by the fact that we have 2581, but the semiconductor guys are talking about Ger- ber now. Acheson: We've met with several semicon- ductor companies, even some from Japan who were looking for a standard format that they could use to build their substrates and all the different packaging types that they do. IPC-2581 does support quite a bit of it. In fact, we've included definitions about wire bonds, for example. We don't specify the wire bond path and structure itself, but an envelope of where that wire bond would be limited to. We can pass along any of the material's data for how the substrate is structured, whether it's glass, ceramic, or whatever material they're using. We've even addressed the abil- ity to dictate where a bond pad starts on a component vs. a bond pad in an amphitheater part of the board. So, that data is there. Along with that, because of its XML structure you can actually build a subassembly of a subas- sembly in a single file. If I have an interposer stack of components on top of it, I can specify that in 2581. e trick is getting the offering tools to output that data. So, the standard supports it. Shaughnessy: Is there anything you would like to add? Acheson: I think if we keep educating people and allow them to understand and see the advantages that IPC-2581 provides, they'll be convinced—even us old-timers. I was con- vinced. Shaughnessy: anks for your time, Ed. Acheson: ank you, Andy. DESIGN007 Karen McConnell