Design007 Magazine

PCBD-June2017

Issue link: https://iconnect007.uberflip.com/i/836472

Contents of this Issue

Navigation

Page 16 of 81

June 2017 • The PCB Design Magazine 17 first came here I would agree with what Happy said. I think that was right, there was certainly a cost issue and tradeoff and how many resis- tors per square inch you need to replace surface mounts, but I think we've moved on. Let me address some of the issues he raised. First I want to be clear, our intent was never to sell Ohm- egaPly as a way of saving money on surface mount components. That model would not fly. Our main thing had to do with performance and density and miniaturization and signal in- tegrity, and so forth. Yes, there were some cost adders at the worst possible place—with the material, because we're basically a material sup- plier. The reason we're still here, being the most expensive product at that time, was because of other technical reasons why people used them. Certainly, if you're building a satellite that's go- ing to cost a billion dollars, you don't care if you have to spend another $100,000 to make it work. But here's what we did and how we ad- dressed those issues. First, we did address cost tradeoff issues, in terms of providing spread- sheet programs with the goal of selecting a sin- gle layer. The single layer is the lowest possible cost adder; if you have to use two layers, say one for the terminating resistors and one for the pull-up and pull-downs, you just doubled that cost-adder. So our goal was to provide a single layer. We would have customers give us the bill of materials, we'd go through the resis- tors and say these are the ones you leave on the surface, these are the ones you can embed using the terminating resistors. For the applications, if we had a BGA we could do something with a BGA that you couldn't do with a surface resis- tor, mainly terminate every lead onto the foot- print of the BGA and free up a lot of space on the surface. All these were ways to reduce cost. The cost wasn't the main thing; the main thing was you had to terminate every lead. The other thing we did was, when some CAD designer ran out of room, the only way they're going to get more parts or more leads was to add layers, and adding layers was the fastest way to increase cost. Using embedded layers was an alternative to adding layers so you wouldn't have to route back to the surface or go down the multiple layers to escape the ar- ray. I think, Happy, you did some of this work when you talked about resistor density limits Figure 1: Cost analysis software developed by NIST and CALCE. A DEEP LOOK INTO EMBEDDED TECHNOLOGY

Articles in this issue

Archives of this issue

view archives of Design007 Magazine - PCBD-June2017