A Day with Pete (Starkey)
Usually the one conducting the interviews, I-Connect007’s own Pete Starkey recently found himself on the other side of the microphone when I spent time with him in his hometown of Market Bosworth, England. There, between hiking and gardening, we found time to discuss Pete’s rich history in the PCB industry and the many changes and surprises he’s seen in the manufacturing process over the years.
Barry Matties: Pete, here we are in your neck of the woods, literally.
Pete Starkey: Yeah, in the woods.
Matties: Why don't we just start with a little bit of your history in the industry.
Starkey: How long have you got?
Matties: Well, the trail's at least three miles long.
Starkey: Right. I studied chemistry at the local technical college and polytechnic and got a first class honors degree in Applied Chemistry. In my hometown of Coventry, a chemicals and textiles group called Courtaulds had big research facilities and I spent some happy years, and quite productive years, working in their research division.
Matties: What kind of research were you doing?
Starkey: At that time it was really into processes and chemistry related to manmade fiber production, but one of their lines was tire reinforcement. At the time they were very heavily into and very well known for their textile tire reinforcement, but it then became fashionable to reinforce tires with steel. They bought themselves into a joint venture that was making steel tire cord for automotive tires, and I'd done a lot of laboratory work and research work into the metal finishing processes that were involved. Although I would happily have stayed in the research division for the rest of my career, they saw the career path being through technical management.
I had a spell as a technical manager in a factory that was producing steel tire cord and the metal finishing processes involved, but I didn't see that as where I really wanted to be for the long term. So I took up a position with a small company in Coventry, called PMD Chemicals, which in fact had been the organization that had introduced a lot of big name brands into the electronics and PCB industry, probably the most notable one of which was Shipley.
They had been the distributors for Shipley until the Shipley guys started up on their own. I ran R&D there, and I was involved in process development and also equipment development. Came a time, I think everyone's ambition in those days was to work for Shipley. A position came open in Shipley in Coventry for a technical service engineer, which I took on, and that was a real introduction to the printed circuit board industry in the UK.
Matties: What year was that?
Starkey: This was mid '70s, and in the UK we had upwards of 400 PCB shops, from the very basic mom and pop shops to the big OEM shops like International Computers, Ferranti, and British Aerospace. I enjoyed my time as a tech service guy, but found I was spending far too long away from home. I took a job closer to home as the technical manager in a PCB shop here in the Midlands until I was invited to go back on the road for another supply company, which at that time was Dynachem, which later became known as Thiokol, which later became known as Morton, which later got absorbed into the same group as Shipley, which later got absorbed into the same group as Dow.
I stayed in that job for a while, and was then invited to join MacDermid where I took a technical sales job, although technical sales at that time was more about technical service. If you wanted to sell anything you sold it not on the basis of price, but on the basis that you could make it work and you would support it. I'd got a pretty good reputation as a service guy. I could make things work and fix things that had gone wrong, whether they were directly related to the product line that I was representing or not. In those days, if you were a good tech service guy, if a week went by without you being offered a position as a technical manager then it was a bad week. You turned down many, many, many offers. But there was a startup company called Forward Circuits that was here in the Midlands, and I joined them. I think I was probably number 12 on the payroll. I stayed with them through several generations until they became one of the most significant technology shops in the UK.
Matties: You were there for 15+ years?
Starkey: Yes, I was there for over 15 years. I started off as technical and quality manager, progressed through technical director and was managing director for seven years. I left shortly before the guys from Viasystems came in and took over the top end of the European PCB industry.
Matties: What were the circumstances of your leaving?
Starkey: There had been some re-structuring at the corporate level above my head, and it had come to a stage where my face no longer fitted as a hands-on, street level technical director/managing director. I was paid off and paid to stay out of the industry in the UK, so I took the opportunity to work as a technical consultant as far afield as I possibly could. I spent some happy times in Europe, some happy times in India, some happy times in South Africa, some happy times in the U.S., and generally made it up as I went along. I sort of fell into technical editorial, basically because during my previous generations I could write good technical reports. That's been my only real paid employment over the last 15 years or so.
Matties: In all that history you've seen a lot of changes to the industry. What strikes you as the thing, in terms of change, that may have most surprised you?
Starkey: Really I don't think anything surprised me, but I can recall a time back in the late ‘70s, early '80s when I was the technical director at Forward Circuits and someone brought to me a circuit board that he was having a problem with and said, "Can you take this to pieces and give me an opinion?" I looked at it and I took it to pieces and it wasn't very well made, or on very good substrate material, and the manufacturing processes didn't look to have been done that well. I asked the guy where it had come from. He said Taiwan. I said, "You're joking. That explains everything." I think probably that was the biggest surprise. You could parallel it with probably with the motorcycle industry in the UK. We were once the world leaders. Nobody could touch us. These people in Japan called Honda, Suzuki, Yamaha, and Kawasaki started making little motorcycles, and the British motorcycle industry looked at them and said, "They're no threat to us." They just carried on. They carried on with equipment and investment that, in most cases, was many years old. The general attitude of the industry was, "We've always done it this way and we always will," and really it was complacency. It was complacency in the motorcycle industry, apart from Triumph, which is the only remaining UK manufacturer, most of whose technology was borrowed from Suzuki and Kawasaki. The British circuit board industry went the same way. I think that people like the venture capital people, like Viasystems, did us no favors by taking out the leading manufacturers, certainly right around Europe. But whether it was a surprise or not? I think I could see it coming.
Matties: A lot of people didn't.
Starkey: From an industry that was 400+ strong in the UK, we now have an industry that, if you count very hard, you may find 30 shops. You may find a couple of hundred shops in Germany. That's really the way it's gone. The only people remaining, really, are the people that were the forward-looking ones, that didn't take the same complacent view as the majority and continued their investment, their development and their specialization. They're specialist companies for the most part, and they're servicing the requirements of industries that, for various reasons, are partly security and partly confidence in quality and service, and prefer to trade with local established credible suppliers than to bring in their stuff from offshore.
Matties: Yeah, I think that is certainly something we're seeing in America as well. From a technology point of view, has the technology progressed as fast or to the level that you might have thought it would have by now?
Starkey: Quite honestly, I think the only significant change in the technology has been the era of HDI, the era of the microvia. In my manufacturing days, all of our through-holes, blind holes, and buried holes were produced by mechanical drilling. It was the sort of revolution that happened during the '90s, and it started off in Japan. I can recall going on several technology missions sponsored by our Department of Trade and Industry to discover the magical land of ”Microvia,” which was supposedly a small island close to Japan.
Again, a lot of that technology was OEM driven rather than free market driven. I think during my existence in the UK industry the big changeover was from the OEM shop, because that's where the technology was, to the free market. For the OEM shop, everyone could write his own set of rules and his own set of standards, but it tended to be quite an expensive operation. So guys set up independently and provided a manufacturing service for these OEMs, and they could do it more cost effectively. As a consequence of that, there had to be standardization so that they could offer the same product, or the same manufacturing service to the same standards, to the general market. The industry had to start adopting standards, and part of my early responsibility was to establish capability approvals to, particularly at that time, British standards—later on to European and to world-accepted standards. That was a change which happened early in my years in the industry.
The technology in the first generation, our accepted plated through-hole and multilayer technology, came out of the OEM shops. The microvia technology came principally out of the Japanese OEM shops. The major technology change has been laser drilling rather than mechanical drilling.
Matties: Now I just saw, I think it was Schmoll, drilling sizes half the dimension of a strand of hair. Was that Schmoll?
Starkey: Yeah, it was. It was part of Michael Weinhold's EIPC conference presentation, because Michael has a very good network in the Far East and has spent a lot of time in Japan and always makes a point of attending the JPCA show. We used to mechanically drill holes at 0.6, 0.5, 0.4, 0.3 millimeter, and 0.1 millimeter was about the limit. Then that sort of technology was overshadowed by the laser drilling technology. Laser drilling sort of blasts holes, whereas mechanical drilling accurately cuts the material away from where you want it. He saw examples of mechanical drilling, and mechanical drill bits—it's one thing having the machine, it's the other thing having the tools to use in it—at 30 microns and trending towards 20 microns, which again opens up a new generation of technology.
Matties: That's the thing, there's the board technology, but there's also the actual manufacturing technology, the equipment that we use to produce boards, and now we're getting into the era of inkjets and spray coatings and such. Has that manufacturing process changed the way that you thought it would by now?
Starkey: Quite honestly, designers tended to operate in a closed environment and they just presumed an awful lot of things about the manufacturing process. They'd read a material data sheet, they would set themselves some design rules, and away they'd go, without really any great depth of understanding about the poor guy on the other end of the manufacturing process. Again, from the next stage down the line, the PCB fabricator didn't normally have an awful lot of idea of what happened in the assembly shop, and the assembler didn't have a great appreciation of what necessarily were the limitations of the materials or the finishes that were industry standard on PCBs. Several of us have made a lot of effort over the years to try to encourage the designer to understand the fabrication process.
I spent many, many years manufacturing jobs that probably, if you looked at actual materials capability and process capability, should not have been manufacturable. I think my background in materials and processes and technical service helped. For manufacturing technology, a lot of the metal finishing processes have shown incremental improvements, but no fundamental improvement.
For imaging technology, the original printed circuit was screen printed and then photoresists were introduced, but these were liquid photoresists, which did a very good job but were very messy to use. Then DuPont invented the concept of dry film photoresist, which really revolutionized it, because it meant that with a very straightforward application process you could do very precise imaging with a dry film resist. Dry film has, again, incremental improvements. Liquid resists were reintroduced and liquid resists have been very successful for fine line innerlayer work.
The actual technique of producing the image on the photoresist has moved forward, whereas in the beginning the thing was always done by photographic contact printing. The photo master had been produced by a guy at a drawing board with a times four sheet of tracing paper and some sticky tape, and some symbols representing tracks and pads. Then the photographer would have reduced that to the right dimension and produced a photo master which would have been used to create the image. That moved on to the design being done on a CAD system, and the CAD output being plotted with a photo plotter, which gave very accurate results but it was slow. Then, in came the laser photo plotter, which instead of drawing feature by feature, drew the whole image as a raster scan image, so that produced a new generation of phototooling.
Also, in terms of the front end of the PCB manufacturing operation, instead of being at the mercy of the photographer, you could receive the CAD data direct. You could do your own pre-production tooling and you could then produce your own working phototool straight off the CAD with your own laser plotter. The next step on from that was to say, "Well, why use a photo master at all?" If you've got the data and you've got the plotter, why not? Then you had guys like Orbotech that came along with direct imaging systems, which took out the photo master. That was, if you like, digital imaging direct to the job, but it wasn't direct imaging, because you were still having to use a photoresist. You were still having to coat the whole surface, expose selectively the bits that you wanted to expose in this generation with a laser imaging system, but you still had to then go through a developing process to take away the unrequired bits from the required image.
In a parallel track, another direct digital imaging system, which is a true direct imaging system, is the inkjet. People have played around with inkjet, and I've been involved in inkjet, for 20-odd years. It's relatively easy to produce a nice looking image on a piece of paper when you're only looking at it with the human eye and you're not that critical as to minor imperfections. The human eye compensates for minor imperfections, but etching machines don't, so the PCB manufacturing process is a lot more demanding than the graphics process. Most of these inkjet systems have been developed from graphics imaging systems, but they're not necessarily the ultimate answer, because the resolution and definition that you can achieve with an inkjet is not the resolution and definition that you can achieve with a photoresist and an LDI system. If you look at the edge profile of an inkjet image, it doesn’t have the sharp square edges that you can achieve with photoresist.
I think we're getting to a stage where we're pushing the limits of PCB technology and there's a gap. If we look at what the semiconductor people have been doing in terms of their imaging technology, the two do tend to be getting closer together, but I think there's a technology gap there that it will take a long, long time to bridge. Really when you look at some of the developments of more recent times with embedded components, that's going some way towards combining the technologies. Then you've got all sorts of concerns about liability and responsibility. If the PCB fabricator is just making the bare board then he can make it, inspect it and test it against the specification. He can sign it off and then it's somebody else's concern and responsibility. If he's going to start incorporating silicon in there at an intermediate stage of the bare board manufacturing process, who's going to be responsible for ensuring that the silicon is fully functional by the time the completed assembly is delivered? There are all of these sorts of concerns.
Honestly, if you look at what was considered consumer electronics 20-odd years ago, consumer electronics was crude printed circuit boards for crude applications. If you look at consumer electronics now, and if you classify things like mobile phones as consumer electronics, which they are, you look at the technology, just the interconnection technology, that is inside a smartphone and it's impressive, but it's scary. How much further can you push that technology with the materials, with the imaging processes, with the metal finishing processes that are currently available?
Matties: I think we're going to find out, because the consumers are certainly demanding more and more.
Starkey: So much more of the interconnect and so much more of the functionality is now on the silicon, but you still need to interconnect those individual silicon devices. Although you are getting more, and more, and more integration on the silicon itself, you're still going to need PCBs in some shape or form as an interconnection system.
Matties: What advice would you give a manufacturer of printed circuit boards today?
Starkey: It's difficult to find a serious answer. I could give some flippant ones. Probably the best thing a manufacturer could do is find someone to buy his company and then live a comfortable and stress-free life on the proceeds.
Matties: That sounds like a serious answer (laughing).
Starkey: But there are fewer and fewer people willing to buy PCB companies that aren't remarkably successful. So many have closed. Some have been taken over and aggregated into groups, but many more have just gone out of existence because they couldn't stand the pace in the market and they hadn't made the investment.
Matties: For the ones who can't find a buyer, or that just decide that they love circuit boards, what advice would you give those people?
Starkey: The future for them is to find a customer that you can communicate with, a customer that you can satisfy, and a customer that will pay you a respectable rate for doing the job of providing a good manufacturing service, because a PCB fabricator doesn't make a product. He just provides a manufacturing service. If he can find the right customer, this is how the guys that are still in existence in the UK and Europe, for the most part, survive. They've found the right customer. They have developed or established the technology that they can satisfy the technical requirements of the customer. They give the customer the service and support that he's looking for, and the customer pays for the service. That's the future. There's no future in just manufacturing volume PCBs in Europe or North America, because there are people elsewhere that are very well invested and very well qualified to manufacture quantities efficiently and cost-effectively.
Matties: Is there anything else you would like to share?
Starkey: We could ramble for the rest of the day, Barry, but I think we've covered an awful lot of ground. I'd get boring then.
Matties: Pete, I certainly appreciate your time and all your service to the industry. It's great having you as a part of our team, that's for sure. We love you very much.
Starkey: You're very kind, Barry. Thank you very much.
Matties: Thank you, Pete.