Editor's note: ASSEMBLY magazine celebrates its 50th anniversary this year. To mark the occasion, we have planned a series of articles examining the past, present and future of various assembly technologies. This month, we kick off the series by looking back-and gazing into the future-of electronics assembly technology.
To get an idea of how far electronics assembly has come in 50 years, consider an article that appeared in the November 1958 issue of ASSEMBLY. The article discussed product assembly at the Victor Adding Machine Co. in Chicago.
If you're thinking, perhaps, of through-hole components and wave soldering, think again. No, Victor adding machines were assembled with presses, resistance welders and screwdrivers. That's because adding machines worked with gears and levers in 1958, not integrated circuits and watch batteries. In fact, the integrated circuit had only just been invented in 1958, while the point-contact transistor was introduced just 10 years before ASSEMBLY magazine debuted.
The term revolutionary is overused in manufacturing, but nowhere is it more appropriate than in the electronics industry. Forty years ago, electronics assembly was almost entirely manual. Long lines of assemblers stuffed boards with components that could easily be picked up with one's fingers. Today, the fastest assembly machines pick-and-place parts at the dizzying rate of 60,000 cph, and the smallest surface-mount component, the 01005, is barely visible to the naked eye, measuring just 0.4 millimeter long, 0.2 millimeter wide and 0.13 millimeter thick.
Electronics Assembly Evolves
Although the modern printed circuit board (PCB) was patented in 1943, it was not widely used at first. Vacuum tubes and other components were so large that traditional mounting and wiring methods were sufficient. That changed with the invention of the transistor. With components so much smaller, the circuit board became the ideal platform.Through-hole technology was patented by the Hazeltine Corp. in 1961. Surface-mount technology was developed several years later. It gained momentum in Japan in the 1980s, and became widely used worldwide by the mid-1990s.
"The change from through-hole technology to surface-mount technology was significant, but it was not revolutionary," says Alex Goldstein., director of operations at the Center for Board Assembly Research at the Georgia Institute of Technology (Atlanta). "It was quite obvious that the dimensions of the components would be decreasing and the density of I/O per square inch would be increasing. We could not just waste real estate on the board for drilling holes for leads."
Of the many changes in the electronics assembly process over the years, two stand out as being particularly challenging, according to Malcolm Warwick, global business director for circuit protection and thermal interface materials at Henkel Corp. (Hatfield, UK). One was the development of robust wave and reflow soldering processes; the other was the elimination of chlorofluorocarbons (CFCs) and lead from electronics.
Developing robust soldering processes took many years. "In the early days of mass soldering, a lot of effort was spent to understand and quantify solderability," says Warwick. "Without controlling this feature, processes failed in a way that was almost unknown when people used positive feedback to compensate for performance automatically."
The elimination of CFCs and lead stands out because both occurred in a relatively short amount of time. "In both cases, there was a long period of denial which meant that improvements had to be generated in an unusually short period," says Warwick. "The switch to lead-free has been the most far-reaching and has involved the widest spectrum of the supply chain."
Not only have electronics manufacturers experienced significant process changes, but they've also dramatically changed their business model. Few OEMs these days assemble any of their products, preferring instead to rely on foreign and domestic contract manufacturers. That trend shows no sign of reversing in the future. Indeed, Warwick predicts that the distinction between contract manufacturer and OEM will blur. "Contract manufacturers are beginning to look like the OEMs they have partially replaced," he says. "They want responsibility for design and development to add value."
The capability of overseas manufacturers has made even high-end electronics absurdly inexpensive. For example, in November, the nonprofit organization One Laptop Per Child shipped the first 1,000 units of its so-called "super cheap" XO laptop computers for use by children in impoverished countries. Assembled by Quanta Computer in Shanghai, China, the XOs have a 500-megahertz processor, 128 megabytes of DRAM, 500 megabytes of flash memory and four USB ports, though they will not be equipped with hard disks. The initial units will cost approximately $135, and the price is expected to decrease to just $100 in 2008.
What's Next
If the past 50 years saw quantum leaps forward in the technology for electronics assembly, the next 20 years promise to bring more of the same. "The real revolutionary changes in electronics assembly are going to take place 10 to 15 years from now," says Goldstein.In the future, the circuit board assembly process will be different than it is today. For example, Goldstein predicts that conductive adhesives may eventually replace solder paste as the primary medium for attaching components to circuit boards.
Screen printing, asserts Goldstein, is the biggest constraint in the electronics assembly process today. "We aren't assembling components with a pitch of 1 to 5 mils yet, but we aren't far away," he says. "That leads to the obvious question: 'Can we do things accurately enough?' We can place components with an accuracy of 20 to 30 microns at Three Sigma. So that's no constraint. The reflow process is relatively simple. The real problem is the printer. Sixty percent of the defects in circuit board assemblies can be attributed to the screen printer."
As a result, Goldstein anticipates that today's three-step process-print solder paste, place components, and reflow the boards-may become a new three-step process-dispense adhesive, place components, and cure the boards.
To make the dispensing process easier and less expensive, a new generation of "smart" conductive adhesives will be introduced. Ideally, these adhesives would adhere only to conductive surfaces and withdraw or evaporate from all other surfaces. Thus, assemblers would not necessarily have to invest in high-precision dispensing equipment.
Henkel's Warwick agrees that conductive adhesive will play a larger role in future electronic assemblies, but he's not ready to order a tombstone for solder paste yet. "Conductive adhesives will grow, but I will be very surprised if they break into the majority of the market and displace solder," he argues. "I see conductive adhesives gaining from low-cost 'simple' products using additive technology. ...I also see conductive adhesives moving into the new generation of complex packages. But, in my opinion, conductive adhesives have some fundamental shortcomings at the highest level of assembly, where soldering will continue to dominate."
Warwick believes that changes to the electronics assembly process will be evolutionary rather than revolutionary. "The drive toward cost-reduction and 'green' credentials will encourage the development of additive processes and blur the line between PCB fabrication, semiconductor packaging and PCB assembly," he says.
In the near term, assemblers will increasingly have to deal with packaging modules, such as system-in-package, package-on-package and package-in-package, says Warwick. These modules will still need to be assembled onto some sort of PCB, but it will be different from the boards of today, with many embedded actives and passives.
Development of these modules has been driven mostly by consumer demand for powerful, portable electronic devices, like the cell phone with built-in camera, global positioning system, MP3 player and Internet access. While that trend will continue, electronics assemblers will also face a fragmented market. While some consumers will still crave multifunctional gadgets, others will want a device that does one thing, but does it well. "We already see some backlash against the all-singing device that does everything, when all that is needed is a single function," says Warwick. "As the market fragments...there will be parallel development of materials to optimize different needs. Low-cost, throw-away devices with limited functionality may use different materials and assembly processes than highly functional...devices where price is not the main deliverable."
This fragmentation will challenge assemblers, who will need to find a way to simplify and cheapen their manufacturing processes, while at the same time allowing variability in the product for the consumer.
In the far future, almost anything seems possible with electronics. Researchers are only now beginning to tap the potential of nanotechnology and microelectromechanical systems. In December, for example, researchers at the California Institute of Technology announced that they created complex logic circuits using strands of DNA. Instead of having voltages in a solid-state circuit represent options in a logic circuit, high concentrations of different DNA sequences do the job.
