What will connectors be like ten years from now? Will connectors and connector technology be radically different, or will there be only modest evolutionary changes from where they are today?
These questions are important to system designers because connectors can affect the cost, reliability, and performance of electronic systems. During the coming decade, systems will evolve to higher performance levels, placing ever-greater de-mands on connectors.
Worldwide, industry spends about $20 billion on connectors for electronic systems. By 2006, the connector value in such systems is expected to grow by more than 50% to around $32 billion.
What will $32 billion worth of connectors look like? No one knows. However, by analyzing trends in systems design, driving factors in consumer product needs, and technological developments in a variety of disciplines, it may be possible to foresee some plausible changes in the nature of connectors, and in the connector business, by the year 2006.
Connectors' role. Connectors have historically served one of two functions within a system--whether the system is a self-contained unit or a network.
The first is partitioning within a system. Partitioning divides a system into discrete modules, each with a specific function. It is an essential strategy for designers, builders, and users to cope with complexity and provide flexibility. Partitioning improves manufacturability and allows for convenient testing, maintaining, reconfiguring, and upgrading of a system. The partitions within a system create interfaces that are joined with connectors--both terminations and interconnecting media. The majority of connectors used today are for partitioning within systems.
The second major role of connectors is providing system power and signal input/output. The I/O interfaces take many forms, including wireless, optical, and electrical. I/O connectors are often tied to historical standards, which has led to a proliferation of passive connectors with limited functionality, typically made from metal and plastic.
It is likely that the basic roles of connectors in system partitioning and I/O will not change fundamentally over the next decade. However, the balance between connector use in partitioning versus I/O is likely to change over that time frame, and the nature of the technology embodied in future connectors will be different.
Five trends will shape the future of connectors:
1) There will be fewer ICs per system, but their interconnections will be more complex.
2)Connectors will continue to enable partitioning as new functionality is added to a system. However, system evolution will drive the new functions into improved ICs.
3) The globally competitive electronics systems business will increase pressure for cost reduction.
4) Consumer demand for improved reliability will continue.
5)The convergence of computing, communications, and portability will demand simplification, standardization, and reduction of space requirements for I/O connectors.
The interplay of these factors will help determine the relative roles of connectors in system partitioning and I/O. They will also define new demands for interconnection that will ultimately drive a movement from the traditional metal and plastic connector to radically different technologies.
Future possibilities. Connector technology is likely to remain familiar over the next ten years, although there will be an increase in the use of optical fiber and wireless technologies. An important factor limiting extensive change is the relatively slow pace of evolution in electronic system manufacturing. For example, billions of dollars were invested in new manufacturing equipment before surface-mount technology became a dominant factor in printed-circuit-board manufacturing. Simple economics required this investment be made incrementally over many years.
System design is likely to change more rapidly in response to changes in semiconductor technologies. The Semiconductor Industry Association (SIA) predicts that DRAMs, microprocessors, and ASICs will evolve to reach mind-boggling speeds, densities, and performance levels early in the next millennium.
According to the SIA, internal silicon clock rates will exceed 1 GHz (compared to today's 300 MHz), and DRAM density will increase by a factor of 1,000 from today's 64 megabits per chip to 64 gigabits per chip by the year 2010. ASIC densities will reach 40 million transistors/cm2--20 times today's 2 million transistors/cm2.
Furthermore, the SIA expects board designs to have 475-MHz chip-to-board clock rates (versus 150 MHz today), and multichip modules (MCMs) and chip-to-chip packaging will have clock rates of 550 MHz. The number of package pins or balls for microprocessors and microcontrollers will double from today's 512 to 1024; for ASICs, they will more than quintuple from 750 to 4000.
These changes will drive corresponding changes in connectors and connector technology, and may force fundamental changes in the capabilities and processes of the connector industry in four aspects:
- Scale. Shrinking physical dimensions of future interconnection devices, and the increasing number of interconnection pin I/Os, might exceed the capabilities of current connector technology.
- Speed. Bit rates might exceed the capability of current connector technology to design traditional terminals and connectors matched to the projected requirements.
- Materials. Future connectors may need to use materials and material processing techniques that are not currently common in the connector industry.
- Manufacturability. One or more of the above three aspects may critically im-pede the manufacturability of connectors in 2006. Entirely new techniques--such as bulk manufactur-ing as used in semiconductor manufacturing--might be needed.
Breakthroughs in connector technology could radically change the nature of connectors and even the concept of interconnection. The long-range possibilities for such breakthroughs are limited only by one's imagination. But if we focus on currently active areas of research with potential for successful maturation in the twenty-first century, we will have a much smaller list of possibilities.
Currently in development are noncontact connectors, smart connectors, and molecularly engineered materials.
Current connectors are limited by their need for precise mating: Typically, metal-to-metal contact or optical fiber alignment is needed for a reliable connection. Any material, configuration, or technology that could reduce or eliminate the need for precise, forced mating without sacrificing performance or reliability could radically alter the concept of interconnection. Some possibilities are: memory-less gel materials; self-aligning lightwave coupling; capacitive or inductive coupling; and radio-frequency, digital-modulation coupling.
Just imagine how a non-contact connector for power could alter the world.
Smart connectors incorporate active elements--ICs, sensors, discrete components--within their structure. Such connectors could allow integration of communications hardware into the connector module and simplify the internal system architecture. They could also provide convenience to end-users trying to interconnect myriad communications and computing systems. Smart power connectors could make the type of power source irrelevant, thus enhancing the concept of portable equipment.
Many researchers are investigating a broad array of molecularly engineered materials, both organic and inorganic, that have potential electronic or optoelectronic applications. One category of such materials can create "quantum dots"--clumps of matter less than 20 nanometers wide (about the length of a string of 60 silicon atoms). Quantum dots can form a "quantum well," which can geometrically confine a single electron, and have potential applications in high-power lasers, high-speed switching, and computer logic.
Recent research has shown that the organic compound chlorophyll, used by plants to convert sunlight into sugars via photosynthesis, also works as a molecular switch. It can shift electrons over a small distance in about a picosecond. Other organic molecules have the potential for data storage with densities of 1011 to 1014 bits/cm2.
A massive challenge for systems and connector engineers in harnessing these potentially useful molecular and quantum mechanical devices is to find viable I/O interconnection mechanisms.
As we progress into the next millennium, we are bound to see major advances in bioengineering. A precursor of things to come is the nerve chip being developed by scientists at Stanford University Medical Center with support from the Veterans Administration. It is an interconnection between individual nerve cells and a computer, which can control artificial hands, regenerate damaged nerves, and could enhance human capability in other ways.
Understanding the dynamics of interconnection is as important for system designers as it is for connector designers. The two are increasingly dependent on one another to achieve common goals of reliable, functional, cost-effective systems. Keeping an eye on the evolution of connector technology is particularly important to the success of both disciplines.
A lesson can be learned from Wayne Gretzky, one of hockey's greatest players, who once said: "'I skate to where the puck is going to be, not where it has been." The same principle should be applied to design.