Issue link: https://iconnect007.uberflip.com/i/1441485
JANUARY 2022 I DESIGN007 MAGAZINE 75 eling. Basically, it's good that people will start looking at these aspects and realize that you can't have it all. Like you said, there are OEMs putting restrictions like 20 passes on a reflow, and that's probably more of an academic curi- osity than a real requirement. People are say- ing, "Well, if it passes 20, you'll definitely never fail," but in the process of trying to make that material work, you'll induce a lot of other fail- ure points. Ritchey: at's the mechanical side of the prob- lem. And I agree where my aerospace clients are wanting to make things work that currently you can't really do. For example, three stacked microvias is almost guaranteed to fail. Shaughnessy: Is that because all the pressure is on that one spot? I know the IPC committee on via failure is still looking at the stacked vias that pass in the lab and fail in the field. Ritchey: Well, it's because the three layers of laminate expand. e temperature puts a strain on the stacked copper vias, or between the bottom plated-through microvia and the copper layer on which it is mounted, which is not very strong. And so that stress when you warm up the board pulls on that and pulls it loose. So, when the board is in your machine and running until it gets nice and hot and it fails, and you pull it out and go into the lab where it's not so hot and it works. So, when I had this problem with thermal vias like this, we would call those rubber band boards. ey go out and fail in the field, and they come back, and they'd pass in the lab. at's what Tarun is talking about, how we've got to figure out how to model. Matties: How does the industry prepare for what's next? Ritchey: We never reflow more than twice. Six, seven, or 10 times is just a way to see how tough the thing is. Amla: Like Lee said, you should have the need for one per side if you've got components on both sides. at it's just rework, which means that the board reliability is kind of compro- mised at that point. But it's a factor of safety within six, but 10 has been the standard requirement for a lot of companies. Matties: at makes sense. But to bring this all around, what should the industry be think- ing about? How do they prepare for what they need to be doing in the near future? Ritchey: I think there are at least two answers to that question. In my piece of the industry, we're not really having any reliability problems with any of the laminates we use. We're having a problem figuring out how to solve those two parameters I mentioned at the outset. How do I keep loss where I can live with it and how do I maintain minimum skew? I still have two prod- ucts when we did our test boards in 2013 that did that, and we don't know why they didn't turn them into a product, but when we built test boards from those, we couldn't measure any skew. At the time, I guess, skew wasn't on the radar of most people. Isola didn't make a product out of that, and my sense is that Tarun is taking advantage of the knowledge he has. I think we're going to find a material that does that. Is that right, Tarun? A m l a : Yes, Lee. Fundamentally dif ferent technolog y from anything we were doing at Isola, but this is what we're working on at intronics. How do I keep loss where I can live with it and how do I maintain minimum skew?