The challenges we face as a nation from foreign competition mean that simply sustaining innovation just won't cut it anymore. To put it another way, sustaining innovation is not sustainable. The Chinese, the Indians and the rest of the world are all competing with us in the development of new advances in technology. We need to focus on something called "disruptive innovation," which, by definition, is "cross-cluster--it involves the creation of entirely new clusters by unbundling and re-combining existing ones." (Kip Bergstrom, Rhode Island Economic Policy Council, July 2007.)

Clayton M. Christensen, author of The Innovator's Solution, explains that sustaining innovation is a concept that brings to market a product or service a company can sell for higher margins to its best customers--in other words, it brings a better product to the market. A disruptive innovation brings to market a deficient product that cannot be sold to mainstream customers (Figure 1). This product may be simple and more affordable,but Christensen says, "I call that a disruptive innovation not because it's a breakthrough from a technological sense, but, instead of sustaining the trajectory of improvement that has been established in a market, it disrupts [the market] and redefines it by bringing to the market something that is simpler." (Gartner Interview, April 26, 2004.)
























Figure 1: The effect of disruptive technology on performance over time.


Today's Electronic Interconnection Architecture

After 50 years of evolutionary sustaining innovations, the industry is more than ready for a few disruptive innovations. In 1958, Jean Hoerni invented the planar transistor that led to Bob Noyce's integrated circuit (IC). The IC led to a major disruptive innovation in mainstream electronic interconnection architecture. Prior to the IC's emergence, vacuum tubes were a dominant force-- even though discrete transistors were gaining a share of the market. Tubes were typically mounted in sockets that were, in turn, metal chassis mounted and interconnected by individual wires. That mode of interconnection fit the thermal dissipation (metal chassis) and replacement needs (tube sockets) because vacuum tubes ran hot and wore out rather quickly.

A couple of cases illustrate the interconnection architecture disruption that came with those cool-running, switch-dense and long-lasting ICs. First, all through the 1950s, Zenith Radio advertised their "hand-crafted" TV sets, meaning they contained none of the then "new-fangled" printed circuit boards used for transistors. But the IC and the PCB were natural partners--wires became traces and wave soldering was developed to make all the joints simultaneously, thus replacing the variable quality and much slower, one-at-a-time "hand crafted" solder joint production. All-in-one substrates were the devices, traces and joints. In the 1960s ICs began to penetrate electronic equipment markets. Finally, in 1965, Zenith opened a printed circuit board shop in Chicago. Secondly, in the late 1950s, Tung-Sol, an esteemed tube manufacturer, introduced a 12 volt vacuum tube for the automobile radio market. That introduction staved off transistors for about six months. At that time, Tung Sol began making transistors, but couldn't make the leap to ICs.

For 50 years, the three basic interconnection elements above chip level have lived together on one printed circuit board substrate. Sustaining innovations have helped components, traces and solder joints on one printed circuit "bed" meet the challenge of Moore's Law, which has resulted in increase of IC density a million-fold from 1972 to 2004. However, printed circuit board traces could not keep up with IC feature sizes. It's noteworthy for PCBs to resolve lines and spaces below 25 microns. On silicon, IC line space dimensions are moving to sub-25 nanometers, three orders of magnitude less than traces on boards. This move is one reason why ICs are operating at gigahertz speeds, while speed at the PCB level is still measured in megahertz for today's system level interconnection architecture.

Multilayer printed circuits, finer traces, surface mount technology, cooler CMOS replacing NMOS ICs--all were sustaining technologies that have kept ICs, their interconnecting traces and solder joints together, more or less compatibly, on that laminated printed circuit "bed." The newest group of sustaining innovations can be collectively summed and defined by the initials HDI, or high density interconnects, including microvias and build-up technology. However, they too will ultimately be challenged by increasing IC density.

In conclusion, I should point out that Moore's Law is, in some measure, a human artifact driven by a marketing need to render electronic equipment obsolete every two years or so. It's the electronic industry's counterpart to Alfred Sloan's model year planned obsolescence introduced to the automobile industry in 1922. Of course, the need for competitive system performance improvements is more rational than style changes.

In my next installment, I'll discuss the impacts of lead-free solder and the benefits of HDI. The discussion will include a cluster of emerging disruptive technologies and their potential impact on the entire electronics industry by simultaneously solving the thermal, dimensional, performance, reliability and cost challenges of today's electronic interconnection architectures.

Harvey Miller has been watching the printed circuit industry for over 30 years as an economist at the University of Michigan, analyst and database creator. Prior to that, he built economic input-output models of the electronics industry for 10 years. Harvey began his electronics career as a components engineer for computer and telecom OEMs--Burroughs and GTE among them. At present, he's putting it all together, generating powerful marketing database tools for the global printed circuit board industry at www.FabfileOnline.com.