Issue link: https://iconnect007.uberflip.com/i/1207026
34 SMT007 MAGAZINE I FEBRUARY 2020 Feinberg: Absolutely. We have no way yet to exceed the speed of light with electron flow. Johnson: I wanted to double back to the thresh- old for when you move from subtractive to additive. Is that threshold well-defined, and is it moving? Prasad: I am not an expert in that area. From observing the additive process, I see that many people are using that, whether the reason is in terms of the technology or cost. Imagine if we are able to do the circuit board in one or two mils; the number of layers that we now need, initially being eight or six, it could be two. Then, the cost would go down, but the cost goes up because of the one- or two-mil line, which can probably be done with the subtrac- tive process. Why is that not being done? I don't know the answer. Johnson: We had a number of conversations through 2019 with people in the supply chain, and some of the advice with regard to compo- nents was, "Go as small as you can," because the drive to make the smaller components is happening thanks to automotive and cell- phones. The secret to a longer-life product, according to these bill-of-materials experts, is to use the smallest components you can. If design teams start following these guidelines, does that get them closer to needing to use some other process? Prasad: I don't think the size of the component matters that much. In terms of the size limit, that is already happening with 0201s and 0105s, and my advice has been quite the opposite, gen- erally. You should use the largest component possible to do what you want to do, meaning the pitch. Go with the largest pitch. The only time the size matters is if we're talking about 0402 or 0201—the resistors and capacitors. Johnson: The largest possible pitch makes sense. Prasad: The largest possible pitch is what you want to go with. You can have your cake and eat it too if it's a thin package, and you are going to use that as a BGA-type package ver- sus QFP. Matties: What is your knowledge around active embedded? Prasad: There's quite a bit of effort. There is some movement, but I don't know that that much is happening in that area for active embedded. Think of the passive resistors and capacitors and what it would take to get them incorporated. We have been talking about embedded passives since the '80s, but how many embedded passives do you see today? If that was happening, we wouldn't be using the 0201 or 0105 because, in any particular board, you have hundreds of these passives being placed, depending on the size of the board. That's a lot simpler than trying to embed the active ones. Matties: Why do they not use the passives as embedded? Prasad: It's all about money. Matties: It costs more to have it embedded? Prasad: Yes. And remember, each board is unique; you could make one, five, 10, 20, 100, 1,000, or 10,000. When you make compo- nents, you make millions and billions. You can never have the efficiency and yield of these processes that make millions and billions ver- sus even if you make 10,000. Ten-thousand boards a day is high speed for most people, and that's nothing for a component supplier. If you compare that to the volume of the com- ponents and passive components, they're in the billions. It's hard to get that efficiency in a low-volume environment and make them cost-effective. Matties: It's always enlightening to talk with you, Ray. We appreciate your time. Thank you. Prasad: You are welcome. SMT007