Issue link: https://iconnect007.uberflip.com/i/1327102
64 DESIGN007 MAGAZINE I JANUARY 2021 Shaughnessy: Do you think out of ODB++, IPC- 2581 and Gerber that any one of them do a bet- ter job at representing stackup intent? Gaudion: They all do a job. The challenge we have is that because our tool is dedicated for stackup, we can end up putting more infor- mation in our tool than the other formats can take. We can output 2581. We have an XML format which contains all of our stackup infor- mation, but there probably isn't any one CAD tool that can take all of the information. If somebody says, "I think there are tolerances in the plating on this par- ticular layer," and we put that into our tool, it may be that there isn't a container for that in the other format. Where do you put that data that's been requested by some customers that isn't in the CAD tools yet? There's always a lag between us put- ting things into the stackup and the formats out there taking them. Shaughnessy: I just wonder how much of it goes back to designers being incomplete or, as we found out in our fab notes issue, everybody does their fab notes differently and the designer is just providing incomplete or erroneous data. Gaudion: We see a huge range of designers, PCB technologists, and those guys in the big- ger OEMs that work with designers to make sure that they understand the difference between finished thickness, plated-up thickness or unplated thickness. Those skills are increas- ingly important to have, especially when things are running at high frequencies where it is as important to consider the copper finish as it is to consider the base material. Take the best laminate for high-speed performance—if you put a very rough copper on it then it's not going to perform so well because you're going to lose signal in the copper. That's interesting for us. We keep seeing ourselves exposed to different parts of the industry as technology changes. When we started doing impedance, we just calculated impedance for fabricators, and then the laminate suppliers got interested. Past that, the foil suppliers started to get inter- ested. What we're saying now is we're having more and more communication with chemistry suppliers because they're working on chemis- try to promote adhesion that is smoother. Also, moisture is certainly very bad where high-speed designs are concerned. If you've got humidity, water has got very high dielec- tric absorption. If you get any kind of hydro- scopic nature in solder mask and you have traces on the surface, humidity can affect the high-speed performance. You want the outer layers to not be hydroscopic. When you talk to the manufac- turers, they say, "Our ink is really good, but it's actu- ally dependent on how well the fabricator cures it." Again, it's an area you wouldn't have even thought of before, but now there are things that come into play that wouldn't have otherwise come into play maybe 10 years ago. Shaughnessy: I think a lot of it does just seem to come down to communication, like 90% of the problems that we write about. Gaudion: We talked about different countries, and certainly in some countries I've heard designers say, "We really want this specified, and we don't want to speak to a PCB fabricator. We'll get all the data, and all the material data sheets, and we'll put everything together. And if we spec it right, then it will work." It doesn't happen all the time, but maybe in 5% or 10% of the time, a designer would further think that actually by specifying the materials out of the datasheet they will get what they want. As Happy explained, fabrication isn't taking standard materials and gluing them together, you're actually plating, pressing things