Issue link: https://iconnect007.uberflip.com/i/1002143
52 PCB007 MAGAZINE I JULY 2018 these days is when they talk about via plug- ging, as Yuki said. Very few give an actual re- quirement such as tented, filled, plugged, per- centage filled, things like that. Some param- eters are not clearly specified (if at all); you get variable results on that. You know, copper heights, coverage, thickness at the knee, etc. As Tom said, the mask used has a big effect on the actual finished thickness. Andy Shaughnessy: So the more information they can give you, the better? Lindland: Absolutely, yes. Like I said, the hole size, aspect ratio, copper heights, trace heights... Kojima: More information is al- ways better, because, for in- stance, when you say hole plug, what does that mean? Does it mean 80% of the hole needs to be filled? How about the number of holes being plugged? Let's say the panel has 2,000 via holes. And let's say one is missing the ink inside the hole. Is that bad, or is it okay? What percent is al- lowable? People just do not have the answer to these questions, sometimes. Shaughnessy: Sometimes designer don't even specify the copper, whether it's finished cop- per or not. They assume everybody knows what they're talking about. And sometimes they're using slang, or other terms that are kind of tribal within their company. Kojima: That's true. Meeker: At the same time, as an equipment sup- plier trying to meet specifications, we're very uncomfortable saying 100%. Because there's always some fallout for whatever reason, and what percentage is acceptable is usually nev- er defined. Kojima: Another criterion is how to check these holes. Like I said, some circuit boards have probably three or four thousand via holes. I mean, how do you check them? Goldman: That's very true. And what gets me is that you're putting down a thinner mask, but you're also trying to put extra material in the holes. That seems next to impossible. Kojima: That's correct. And yet they want us to control the amount of ink that goes inside the through-hole or, like I said, micro and blind via holes. Lindland: The other thing, too, for the design- ers, is that as you're in the field making panels, and there are very few times that they'll actually specify thick- ness. A lot of times they'll spec- ify a minimum mask thickness, without saying a maximum thickness. The Circuit Automa- tion machines do not have many issues with thickness; we tend to error on too thick of cover- age, which can be easily adjust- ed thinner, and they're rejected because the mask is too thick. And that creates tombstoning or whatever problems at assembly, yet it's never specified. Goldman: Very frustrating. Well, I guess that mask thickness is partly defined by how fine the features are. But you don't always know that either, I suppose. Meeker: And it's also dependent on the sub- sequent process; if they're doing something like deep gold they're going to want the mask thicker for more protection. A lot of the subse- quent processing after ours makes a big differ- ence in how thick or how thin they want the solder mask. Matties: Is this commonly known by design- ers? If they're not specifying thickness, do they understand that there could be assembly chal- lenges, or, as you're talking about, a gold bath issue? How do they know this? Larry Lindland