Issue link: https://iconnect007.uberflip.com/i/1417991
18 DESIGN007 MAGAZINE I OCTOBER 2021 Ki-CAD. Why native EAGLE and native Ki- CAD? Because they are also open, readable formats, and we have programmed our own input processor to transform this data auto- matically into the internal Ucamco format, the DPF format. We skip the Gerber translation in between and we go directly from EAGLE and Ki-CAD into the Ucamco format, which of course is an advantage, because then you don't have this translation step. But from these 2,000 files per day, if I look at ODB here in Europe, I think 10 to 20 times a month I get such a file, whereas from the other, I get 2,000 a day. It's this kind of relation. Why is that? I think it's because many people we are dealing with, the bulk of the market, are not using the big CAD licenses. e big CAD companies are big in number of dollars, but they are not big in number of licenses they have in the market. e number of licenses in the market used by a huge number of small designers are smaller CAD systems, and they don't have these fancy output processors that Karel is describing; they just make Gerber, Gerber X1, and now Gerber X2. at's about it. at's the reason we get this. For EAGLE and Ki-CAD, we solved the matter altogether, so then we don't need it anymore. at's where we are with our prototypes and small series business in Europe. Shaughnessy: You do accept ODB++, also. Stans: Yes. But we would get ODB++, let's say, 10–20 times a month. Everyone knows that you can't misuse Gerber, because if you send us a Gerber data set, the data set is only for one board, for one version of the board, and that's it. You're never doubtful. If you start misusing the ODB container as a filing cabinet, and you keep version number one and version number two of the board, in finalized form and non- finalized form, if you send that to me, the first question I ask is, "Which of the stuff in there should I make for you?" It gets messy because people have the freedom to start fiddling around in it. But you don't have that freedom with a one-to-one clear output for Gerber. No discus- sion. It's simple for everybody. If you want me to make three different boards then you need three data sets. If you want me to make three versions of the same board, then send me three data sets. Don't send it all in one big ODB con- tainer where I must answer a riddle and find the weapon that killed the colonel (laughs). Tavernier: Actually, currently, Gerber X3 is complete. I have a few attributes that I would like to add, but that will depend on somebody who is interested in implementing it. I don't like writing a spec that is not implemented. As Confucius said, "You only understand the spec if you have implemented it." And also, if it's not used, it just clutters the spec. ere's nothing in the Gerber spec that is unused, and I want to keep it that way. Stans: ere is one thing missing, my friend. Tavernier: What is it? Stans: at's my long-time dream, to put a D behind it and make it X3D. You could include the STEP files, but that is up to the CAD sys- tems. CAD libraries are usually poorly main- tained, and you can only output what's in them. If there is no stuff in it, you can't output it. But that would be, really, a step forward. We are moving toward electronics manufacturing, meaning we'll do the PCB, the assembly, and by the end of the year, we will start building the enclosure, the foils, and do everything in- house. With X3D, we could then do complete virtual manufacturing at the beginning of the process, doing a full DFM of all these factors that will follow later in the value chain like enclosure, foil, assembly, everything. at would be a big benefit, because a mis- take that you correct at the beginning of the value chain is a very cheap mistake. A mis- take that you need to correct at the end is a very expensive one. And that is the whole idea