Issue link: https://iconnect007.uberflip.com/i/1256432
JUNE 2020 I DESIGN007 MAGAZINE 19 They'd rather pay to have it done 14 times than pay to have it done right the first time and save themselves $1 million on it. to me, and I said, "That chip is probably going to be used on FR-4 with through-holes. Why are you using these expensive ones?" They ex- plained, "It gives us the best graphs and charts on our app note specs." I responded, "Yes, but blind vias with low impedance is not the way it's going to be used. It's going to be used on FR-4 with through-holes. They're not going to see those curves that you're presenting." Is this common? It's like McDonald's and Burg- er King commercials showing you a big juicy hamburger, but it's not big and juicy like that. Beeker: That's true of ours. I've had a couple of unhappy conversations with some of our app note people who were specifying material for something that didn't need anything like that. On top of it, it costs them business because if their customers go look at the system costs and find out that this material is 5–10 times more expensive than what they're used to, they will find another device. Sometimes, you lose busi- ness you didn't know you had to lose. That's what's so pitiful about app notes is they are important, but nobody puts that kind of impor- tance on them. Holden: I don't think these people recommend- ed that. They were simply using it to get a bet- ter response from their device because of the parasitics. Beeker: They probably described the materials used in their board in their application note, but if they didn't, then they're cheating. The bottom line is if you're trying to show the real capability of the integrated circuit, which is important, and then you will do something like that. You'll put it in a perfect world so that you can see what is happening at the pins. If you can't afford to make it that good, you're going to have reduced performance, it gives you the possibility, but maybe not a practical solution. Matties: I'm still stuck on that story you were talking about earlier, Rick, of the board and the designer following the app notes who thought he was optimizing and drove the cost way up. Is there a peer review process for somebody who doesn't have all the experience that's commonplace? Hartley: Yes. Every engineer in that company reviewed that design, but none of them real- ized it was going to be a problem. Matties: Maybe they have to go to a design bu- reau or a service that is independent of the world that they live in. Beeker: And they're not going to go outside the company. They're proud of these designs. I of- ten get customers requesting help with their EMC issues. They'll send me a snapshot of a two-inch square on the board and say, "What's wrong with this?" They don't want to share the design because it's secret, but there's nothing secret here; they're guarding a dead cow. They don't want to pay for anybody to do any out- side analysis. The money's not there. They'd rather pay to have it done 14 times than pay to have it done right the first time and save them- selves $1 million on it. Johnson: Lest we paint the wrong picture, are there some semiconductor companies that do an excellent job or a consistently good job? What is good about app notes out there? Beeker: I don't have any experience except with my stuff and the ultra-conservative stuff from memory vendors and SMPS application notes. The circuit design is probably okay, and the schematic itself is probably solid. There are still going to be things you need to understand to make it better for your application, but you have a functional starting point.