Issue link: https://iconnect007.uberflip.com/i/1116895
32 PCB007 MAGAZINE I MAY 2019 One of the issues we have to solve when we go from one generation of material to the next is how does a material supplier convince a glass copper or resin supplier that this re - ally is the next big thing? I've had that ques- tion posed to me by my customers and their customers—the end users—several times, and there's no easy answer. The real change comes when the structure changes and mon - ey starts being spent. Until there are actual designs that go to some level of production, then you can start talking about cost because the cost reduction is completely driven by vol - ume. It doesn't have to be the same volume of production as the FR-4 type materials, but it has to be a significant production for at least a couple of suppliers in the supply chain. The raw materials that are being used to make the resin, glass, or copper are commodities, for the most part. The chemicals are in the res - in and the way the glass is formulated drives down into a small production volume. That small production volume drives the margins all along the supply chain. Our challenge with advancing technology is to figure out when and how much. Companies in the PCB business that have survived are surviving for a reason. Within our business, one of the reasons we've sur - vived is because we adopted a more conserva- tive business model than in the late '70s and early '80s. They were very bold risk takers. It paid off often enough that it worked for the industry at the time. Those people have had a tough time supplying our business without a significant amount of risk. I'll give you an example. In the '80s, one of the largest problems we had in the circuit board business was the reduction in the size of the features, the spacing, and the thickness of the dielectrics. We were going to controlled im- pedance. The boards were getting thicker, the cores were getting thinner, and the space re- quirements and true position tolerances were getting tighter and tighter all the time. The cir- cuit board business hadn't figured out how to compensate artwork to make up for that. Then, a company came along that made what seemed like a common-sense effort to make a unidirectional prepreg. This prepreg basically had zero movement along the axis of the glass because it was not woven. If you put one ply one direction and turned another ply the other direction and made a symmet- rical construction, you could achieve almost zero movement on a standard FR-4 compos- ite. But it had some limitations and problems that I'm sure Ed Kelley at Isola could tell about you every one because he was involved, but this product worked. Everybody I ever heard of that tested this product had amazing registra- tion far beyond what they needed. The prod- uct was a significant threat to FR-4. If it had become commercially successful, it would've been very disruptive. Unfortunately, it was a low-volume manufacturer. They never signed up a big enough customer to transition over into mass production, and they failed. In the meantime, PCB businesses, which never sit still for long, had been investing in their front-end engineering and equipment. They decided to cover this problem through artwork compensation. By doing artwork com- pensation and dialing it in, they got where they needed to be technically. That solution, which was best for them, was the worst for this par- ticular supplier, and they just disappeared. Those are the kind of things that happen in our business. You're often given a choice of one or the other solution, and the technology that's not chosen usually doesn't survive. This is pretty well-known among people in my busi- ness (laughs). We've been watching it happen for a long time. Johnson: You could take a risk on something years back because you had the margins to support that; now, you don't. That small production volume drives the margins all along the supply chain.