When it comes to the PCB design community and the challenges facing it, as well as where design has been and where it’s heading, there aren’t many people as qualified to speak on the topic as long-time industry veterans Happy Holden and Charles Pfeil. I joined Happy and Charles at the recent AltiumLive 2017 event in San Diego for a discussion about the current state of PCB design, what the future holds, and what it means to be a designer.
BARRY MATTIES: You both have been in the industry for a number of years. Tell me about how you got your start. Charles, let’s begin with you.
CHARLES PFEIL: I started working summers when I was in high school, so 1966. My first job was doing QA on Rubylith designs for Northrop. Then I eventually became a designer and I job-shopped for probably about 14 years before I started my own service firm.
MATTIES: So it was always design then?
PFEIL: Always design, yeah.
MATTIES: When did you two meet?
HAPPY HOLDEN: Mentor?
PFEIL: No, before Mentor. We probably met at trade shows prior to Mentor. I'm sure of that.
MATTIES: There's a lot of history between the two of you.
HOLDEN: Lost in antiquity. You should've asked us 10 years ago. We would've remembered (laughs).
PFEIL: I was writing my book around 2005, so I know we met before that.
MATTIES: In any case, you started out with circuit design. That was how many years ago?
PFEIL: It was 51 years ago.
MATTIES: What's the most dramatic change that you've seen in circuit board design?
PFEIL: I think it's the impact of miniaturization, the impact of high-speed circuits, and the impact of new fabrication technologies. One of the difficulties has always been keeping up with that, especially with software design. When developing software, we'll start working on something to support a new technology and then, before you know it, it's moved on to some other technology before that method has matured. That's why so often the software companies are behind the curve when it comes to keeping up with technology and fabrication changes.
MATTIES: When you say the impact of miniaturization, what sort of impact do you mean?
PFEIL: The type of design solutions that used to be valid became invalid. A great example was when grids were no longer a valid method for primary routing. When they switched over to surface mount and you had to go between the pins and you're not on a grid, I was working at ASI at that time. We had to make significant enhancements to support that technology. All the algorithms were based on a grid, so you could have a rather rough minimum resolution. But as soon as you go gridless, you've got to have a very fine resolution, which means all the algorithms would no longer work effectively. So that was a big change.
MATTIES: What do you think the greatest challenge is today in PCB design?
PFEIL: I think understanding high-speed circuits and designing them in a way that they'll behave as desired.
MATTIES: I was just listening to Max Seeley, the gentleman from 3M. He's one of the keynotes here. He said it's really the EEs that are going to become circuit designers.
PFEIL: I think Happy would agree with me that the designers who have experience with manual design are retiring or passing away. The good news is that most of the ones coming up behind them have electronics education. They know circuit design. They choose to do PCB design, like Rick Hartley, a great example. For them, the design is really a big challenge, but because they understand the circuit side better, I think it's easier for them.
MATTIES: The tool certainly makes it easier.
PFEIL: But for a circuit engineer to go into design and not understand the fabrication process and or how to constrain it properly and not over-constrain it, that's a challenge.
MATTIES: One of the things that we often hear, and you just mentioned it, is designers not understanding the manufacturing process. I hear that frequently. Is it really their job to understand the manufacturing process, or is their job to get feedback from the manufacturing process?
HOLDEN: I for a long time have advocated that it's useful for designer to get a tour of a PC board shop and sit down with the fabricators. But for a designer magically to understand the nuances of what takes a process engineer an entire career to learn, they're not going to learn in one hour. You're deluding yourself if you think you're going to give these designers this two-hour tour of how PC boards are made and they're never going to make a mistake again.
MATTIES: But we hear all the time that designers need a better understanding of manufacturing.
PFEIL: And you do. In the context of that it helps you understand what kinds of constraints are realistic. I'm a big proponent of not over-constraining. There's a tendency to over-constrain, not only the high-speed circuits, but the board itself. By over-constraining, you frequently end up with more layers than you need. It becomes a higher-cost endeavor.
I was fortunate to work at a service bureau that was bought by a fabricator and I worked with them for a number of years. I got to see and do test measurements and engage with the fabrication process and know the realistic limits that you are working with. And be sure that when a fabricator posts their limits on the web, it's not their limits. Maybe it'd be good to not quote me here, but that's a marketing effort. They can do that, but you're going to pay for it. For their realistic place where they can produce, their sweet spot, they need more space.
They don't want to go down to absolute minimum trace widths. That's the key. What I would recommend for a company that hasn't already done this is make sure they engage with their fabricator and assembler and find out where their sweet spot is. That way if they can work in that, they'll get their boards done cheaper, faster, and both sides will understand what compromises need to be made in order to succeed.
MATTIES: Everybody wins when we stay in our sweet spot. It doesn't mean we're not going to take on our science projects, but the sweet spot is where the customer wins, your employees win, and your process is quite capable and probably the most profitable.
PFEIL: But you have to mix that, too, with doing something innovative.
MATTIES: Those are the science projects.
PFEIL: Yeah. If Apple wasn't innovative and pushing the limits all the time, then they would never have made it this far. Do you realize that the first iPhone was launched 10 years ago? It's unbelievable.
PFEIL: Happy, do you agree that the fabricator can really help designers and let them understand what they're happy working with as opposed to pushing the limits too far? And know that, if you're going to push the limits, it's going to cost extra.
HOLDEN: I advocate that every designer, if they get the opportunity, take a tour of a PC board shop. But I think all we can expect out of a couple hour tour is the designer to walk away realizing, "Boy this is really complicated. I didn't realize it was so complicated." If that means that he'll ask questions and go to a fabricator earlier in his design cycle, then we've achieved. But if you think that giving him a two-hour tour is magically going to make him a manufacturing expert so he won't make design mistakes, no. That's naïve.
The main failure is with the fabricator, because having been a fabricator and then being put into the design side, I realize in the schedules for designers that they have to make so many decisions. As they come to those branches and decision points in their design they have absolutely no information available to them to say, "Well, if I do it this way, how easy is it to build? What does it cost if I do it this way?"
Do they have time to pick up the phone and call? Maybe a couple do. But it'd be so much better if we had taken the detailed design process, put in the milestones where you have to make critical decisions, and then outlined all the possible branches, what those branches cost, and what they do to the manufacturability. That way, as they make the decisions, there's something for them to go to that can show them the best course to take. Designers can't put that together. Only fabricators know that information.
I lay the blame primarily on the fabricators not making their alternatives more visible in the design. Not that the design people didn't learn about manufacturing. How are they going to learn about manufacturing? Their career is design. Only the guys that have spent their careers in manufacturing are going to know about the nuances of the alternatives. That design information still does not exist.
MATTIES: What do you think of what Charles just said about that? Do you think it's practical for fabricators to actually put together that guide?
HOLDEN: We've already done it at HP. We put together a DFM manual that included cost and signal integrity information for all the different choices.
MATTIES: But that's HP. We're talking about the general circuit board shops today.
HOLDEN: Yeah, but after you do it once, it's easy to copy for your company.
PFEIL: The information's out there. Every board is different. For an experienced designer, half of that stuff is already implanted in their brain and they know to look out and take care of the critical aspects. When you're doing new things then, Happy's right, you really need to look at it from a perspective of how do I make this design so it works in fabrication? You're not going to do that by, as you say, a one-hour tour.
Engage with your fabricator all the time, and most of the companies do that. The large companies do that. They have people assigned to do that research and communicate with the engineering teams as to what they do. But it's the small companies, the startups, the makers. They're not really aware of those things so they need that education. There are methods for educating designers, like what Mike Creeden does with the CID program, which is really valuable, especially to a new designer. I think if that can be done for fabrication methods and approaches to solving problems and presented to the designers in a way that they can make intelligent choices, that'd be really valuable.
MATTIES: Wouldn't it just be logical for a board fabricator to offer design services and actually be a design bureau as well?
PFEIL: I don't know how much of that occurs today. There are companies who do both.
MATTIES: Based on what you guys are saying, it seems to me that you would wind up with a better product from a company with that environment.
HOLDEN: I discovered when I was touring Japan that most of the board designers in Japan are electrical engineers who come from manufacturing, but they have to be in manufacturing for at least five years before they get promoted into design. When I saw the Japanese model, I realized, "That's a great way to create board designers." For some reason that's never been the North American model. But at Gentex, as an automotive electronics manufacturer, all the college graduates start on the production floor for the first three years without change. After three years of building the product then you can elect to go into R&D, into design, into robotics, etc.
MATTIES: But you start there.
HOLDEN: You start on the factory floor just like in Japan kind of as an apprenticeship. Because the owner believes that you can never design and help build the future products unless you thoroughly understand our current products and their customers.