What a Long, Strange Trip it’s Been—and It’s a Long Way from Being Over
A Conversation with Industry Icon Harvey Miller
Harvey Miller has been in the PCB industry for more than 40 years, and while I’ll let him disclose his exact age, I think it’s safe to say he’s probably seen it all. I recently sat down with Harvey at a local wine bar in Palo Alto, CA to learn more about his history in the industry and where he sees it going forward from his truly unique perspective. I must admit I was a bit inspired when Harvey arrived to our meeting wearing jogging shorts and running shoes. No surprise after what his doctor recently shared with him.
Barry Matties: Harvey, thanks for joining us today. You have been in the industry for many years. What year did you get started?
Miller: In the Middle Ages.
Matties: There is a lot of speculation as to your exact age. Can you clarify for our readers?
Miller: You know, I have no problem disclosing that. I turned 93 on November 23, actually. A year ago, a cardiologist at Stanford told me that I have the heart of a 40-year-old, so I guess it’s 41 now. According to the actuarial tables, I am right around 79 or something like that.
Matties: So you are here for another 30 years.
Miller: Or I have been dead for about 14 years. [laughs]
Matties: It was a different world when you were born.
Miller: Yeah, it was 1922. The big growth industry was the automotive industry.
Matties: It’s the growth industry today too, isn't it?
Miller: Some things never change, right? It’s a different kind of automobile now—not the Model T. Of course, that is when forced obsolescence was born. William Durant of GM came up with the yearly models and that has been the model ever since. Now they do it by Moore's Law. I don't know what’s going to happen when Moore Law's runs out, because that has been the mechanism for obsolescence. Of course it won’t run out; there will be other things that will take its place.
Matties: What sort of education did you have?
Miller: I studied economics at the University of Michigan.
Matties: Were you born here in the San Francisco Bay Area?
Miller: No, I was born in Baltimore, Maryland but I grew up in Northern Ohio and then went to the University of Michigan. My father was a salesman, actually. He sold furniture. In the 1920s, Middletown, Ohio, was at the center of the automobile industries. Middletown as well as Akron were tributaries to Detroit. Akron backed them with tires, of course. Canton had tin can roller bearings and Diebold made safes and other things, and now they make ATMs.
When I was in the Army they sent me to study electrical engineering for 18 months, but that petered out because the war was over. They had a special program, and I got lucky because I was always good in those intelligence tests, and that is how I spent the war years. I later went back to the University of Michigan.
Matties: So not only were you trained in electrical engineering, you finished up in economics. It sounds like you were very well-suited for analysis in our business. Has your entire career been in circuits?
Miller: I went to work for a consulting company as a precursor. It wasn't as a participant. I never had my hands in the chemistry, but just as an analyst.
Matties: What year did you start?
Miller: I worked for a bunch of OEMs and back then whenever an OEM got to around $100 million they built a board shop. They all had board shops. I guess when I got interested in boards and printed circuit technology it was probably the early ‘70s, maybe 1973 or so.
In those days, probably more than half of the shops were captive shops, and certainly then for the next decade and through the 1980s. It wasn't until Moore's Law forced density and the fabrication got a little too sophisticated that OEMs like IBM, etc., decided that it wasn't their business. That’s when all the big independents started to blossom and bloom, like Hadco and Zycon, etc., in response to the captives getting out of the business.
Matties: In all the years that you have been involved in this industry, what is the most surprising thing that comes to mind?
Miller: I guess the most surprising thing is the growth of the Chinese industry, because nobody would have anticipated that.
Matties: What do we have left in the U.S.? Something like 280 shops?
Miller: Around 300, but it’s the dollars that matter. It’s been totaled around $3 billion, which represents 5% worldwide. That is another surprising thing. The North American share went from 90% to 5% between the years of 1975 and 2015.
Matties: But most of it left around 2001; it was like somebody turned off a light switch. It was just gone. That was surprising, but there was a big race to low-cost labor.
Miller: And that is part of a pattern. It wasn't just the board industry, obviously. People talk about reshoring, but they don't realize it is going to take one of those paradigm shifts. So we are going to do an end run around those guys over there. That is what we should be looking at, not reconstituting the old technology and beating them with that, because we can't.
Matties: It has to be something new, like the Occam process.
Miller: I think embedded is the more general term, and Occam includes that. What Occam and Joe Fjelstad did was add another feature to embedding—which has not made an appearance yet, in general—that separates the two platforms. I’m paraphrasing Joe's work on Occam, but he had one platform for the components and another platform for the interconnect. Except most of the interconnects are relegated to the third dimension. That does a lot of really wonderful things, because it relieves the substrate of the constraints for doing both. They’re two separate functions: holding the components and interconnecting them. They should not be on the same substrate, because one of the drawbacks is you have conductive anodic filaments (CAF). They grew up with the glass cloth idea. They needed a really rigid substrate and glass cloth with epoxy cure made a great substrate.
And that is the backbone of the conventional board industry. But it has the drawback that as the dimensions get smaller, filaments in the glass are conductive, and it shorts. If you separate the two functions of holding components and interconnecting them, then you are free to go to all kinds of other substrates that don't have that drawback, besides the cost savings, etc. So embedding is the general word and I think Occam is a little twist that adds a lot to that, which people haven't seen yet.
Matties: How long do you think it will be before we really embrace embedded technology, because solder joints are the big cause of failure, right?
Miller: Pitch is the problem. As densities go down and the pitch gets smaller there is a bridging problem with solder, so everyone knows solder is the main cause of reliability issues in electronics. I can't quantify it. Werner Engelmaier wrote a lot about that and he wasn't the only one. Then on top of it they throw in this lead-free solder, which had no environmental justification whatsoever. What put it over the top? People blame the European community and environmentalists, but it was really the tin industry and the solder industry—they put it over the top. It was a political thing. Nobody could ever prove that people made fortunes on speculating tin [laughs], so we will never know, right? But I am sure that a lot of money was made that way.
Matties: I know that Joe was a big opponent to the whole thing and was very vocal about it, but it didn't seem to matter.
Miller: Pam Gordon was commissioned to come up with the cost of the transition to lead-free.
Matties: Technology Forecasters?
Miller: That’s right, and I think she came up with $30 billion and still counting. The military can't use it due to reliability and the little tin whiskers.
Matties: Was that move to lead-free a surprising moment in our history to you?
Miller: It was surprising, because I am not a realist. Everybody who was a realist knew it was going to happen. We kept fighting because we are just really good at fighting losing battles [laughs]. We are going to have the last word when they eliminate all solder.
Matties: And they will.
Miller: The lead-free people made a contribution to that. That is the irony of it. By reducing the reliability of all solder they helped seal its fate.
Matties: When you look to the future, obviously you have a lot to draw on from the past. What advice would you give a manufacturer?
Miller: Be flexible; you just have to be flexible. The industry is full of corpses, isn't it? You see them all over, and not just in this industry. Eastman Kodak is a classic case. They knew that electronic imaging was coming. They even had patents on it, and they even brought out an electronic camera which wasn't based on silver halides. But the thing that really killed it was putting cameras on cellphones. So you don't have to wait for film to develop.
Matties: Kodak had a digital camera, but they were just so locked into the old paradigm.
Miller: It is the old classic case. Everybody gets in a rut—companies get in a rut, people get in a rut, it is natural. It is the first law of Newton's Laws, isn't it? Inertia. Especially when they were making money on all that paper being developed. And all the people in power got in power because they were making money on the old technology, so it is very hard for them to see the threat. They sort of wanted to pretend it was going away, because it was such a threat. People don't want to see threats.
Matties: What do you think about all the automation that is coming to factories?
Miller: The more the better.
Matties: It changes the need for the labor pool.
Miller: Well, it is the old story then. People will do more productive things. Let the machines do the work.
Matties: It also makes a case that you don't need to have a specialized skillset, because that is one of the problems in America. If you look at the machining industry, trying to find a skilled machinist today is virtually impossible, but finding someone that can run a computer that runs the machining is quite easy.
Miller: That is a good point.
Matties: And I think we are going to see a shift in labor, but one thing that is surprising to me is that the process of manufacturing a circuit board hasn’t changed all that much. The high-end or the technological shift has been in the equipment we use to plate a circuit board or to image a circuit board, but the process for building a circuit board has not changed that much.
Miller: No, it hasn't changed that much. Ken Gilleo did a great job on delving into the history at Technograph, but it grew out of other industries. New industries always do. It grew out of graphics and printing and borrowed from a lot of other industries and as needs grew, people came up with solutions. Then it grew into something that will probably become obsolete pretty soon. But that is always the scenario.
Matties: It has to be, right? Because one of the problems we see is there are not a lot of young people in this industry. I mean, look at who we are talking to [laughs].
Miller: That’s right [laughs].
Matties: You have lived a rich life, though, Harvey. You have done a lot, building Fabfile for many years. That was a passion and a labor of love and such an important tool that so many relied on.
Miller: It is even going to be more important when the merging of fabrication and assembly takes place.
Matties: If you were to start a business today, would it be circuit board manufacturing?
Miller: No, I don't think so.
Matties: You might pick assembly. I see a lot more assembly shops opening, because with a few pieces of equipment and 4,000 or 5,000 square feet, you could be an assembler.
Miller: We have to go back to Moore's Law. The basic proposition was that every two years it doubles the number of transistors. But the more important corollary was the reduction in cost from dollars per transistor to fractions of a cent. And the periodicity, every two years, represented the period for obsoleting all the electronics that are out there. Like Durant did with his model year, back in the ‘20s.
Matties: What do you think our technology will look like in 40 years?
Miller: You know, there are so many uncountables. If you look at what experts predict, like Peter Diamandis, who wrote a book called Abundance, there are a lot of things that are on the horizon and then so many things that are just unpredictable.
Let me give you an example of one of the unpredictable things that people don’t think about concerning the Internet of Things. One of those things is chickens. Nobody knows how many billions of chickens are around, but they have estimates. Let's say it is 10 billion. It is more than there are people, for sure. Every one of those chickens is a thing that will have an internet connection.
I was talking to a nurse at the hospital for a study I am in and I was pointing that out to her that they will be able to apprehend avian flu right away. Things like the health of the chicken. Farmers could think of all kinds of parameters in order to check up on it. I pointed out to her that they will not be using Facebook. And she came back and said, “No, they will use Twitter.” [laughs] My rejoinder was they already had Twitter. But there are estimates of the number of cows, and the number of pigs, and they will all have Internet connections. Just think about the number of interconnections—that is the future.
Matties: When talking about the future of the industry itself, we know there is a course that it is on. Is the course to a new paradigm or is the industry just going to die away?
Miller: Of course not. There is no limit to human ingenuity. That is the important thing.
Matties: We abandoned manufacturing in America for low cost labor in China.
Miller: Searching for low cost labor is a transitional thing. Automation will open up a whole new way of doing things. The board industry has to be an example. The function of joining electronic components is not going to go away and printed wiring boards serve that function, so they are not going to go away. But the way they are made—that is subject to change.
Matties: It seems that we use a very costly and time-consuming process. A subtractive process. With the advent of inkjet printing, that seems to me like a tabletop manufacturing facility.
Miller: And now we are talking about 3D manufacturing of course—adding layers and programming the deposition. We have only seen the very beginnings of that, but that certainly is one direction.
Matties: Do you think that is something that is going to happen in the next 20 years or so?
Miller: Oh, I think so. I think they will be able to deposit dielectrics and patterns, additively. The history of additive and photo circuits didn’t work out because they didn’t have compatible chemistries, and once something gets in the mainstream it gets pretty hard to fight. Additive wasn't in the mainstream, but subtractive worked and they made it work. That is where the momentum was. The whole new approach to additive, like 3D manufacturing by selective deposition, where they can actually deposit patterns, might be the future.
Matties: It sure seems to make sense to me.
Miller: They have to be able to deposit patterns. As I suggested, joining fab and assembly is certainly one big process-saving step that will change the structure of the industry. But you know, we didn't even bring in photonics, which is another major direction. Vario-optics is doing some and TTM is doing some. That is certainly a large part of the future, and the players are different. It will take components that emit light.
Matties: It is changing on all fronts. We are going to have edible electronics for our pharmaceuticals, and things like that. The medical industry is one area we are going to see substantial change. The prosthetics that we are seeing and the electronics that are connecting to the mind are remarkable.
How do you view the association today?
Miller: It changes as the industry changes, but the IP started out in an environment where the actual fabrication was done by captives. It was really a group of independent shops fighting for the growth of the independent industry. Of course, the captives went away and that is no longer the differentiation.
Matties: Is there anything that we haven't talked about that you would like to share with our readers and the industry as a whole?
Miller: The changes we are talking about with the elimination of solder, embedding die, photonics, carbon-nano tubes, graphene and the replacement for silicon won't happen overnight. People will have time to adjust. They have a long way to go yet, but they will go.
Matties: Harvey, thank you so much for spending time with us today.
Miller: Thank you.