A Day with Pete (Starkey)

Estimated reading time: 17 minutes

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.

• < Previous Page