In a small San Diego, Calif., lab, a group of engineers practices what’s been called the black magic of the wireless technology world. They’re designing radio frequency integrated circuits—tiny chips that process signals for everything from cell phones to televisions. In the lab at RF Magic, a four-year-old chip maker that has raised $40 million in venture capital, engineers test chips that will handle television signals at broadband speeds. They’re the type of chips that are increasingly finding their way into set-top boxes and other television equipment, replacing canned tuners that are slightly bigger than a pack of gum with a piece of silicon smaller than a fingernail. “We’re enabling the next generation of consumer electronics,” says Mark Foley, chief executive of RF Magic.
Wireless chips are becoming the big growth story in the semiconductor industry, taking over from the computer chips that led the industry in the 1990s.
What makes radio frequency design akin to black magic? “The tasks required of today’s wireless circuits are extremely challenging,” says Ian Galton, a professor at the University of California San Diego’s Center for Wireless Communications. “For example, the receiver in a typical cellular telephone must be able to recover a very, very tiny signal in the presence of much larger interference signals. The power of the signals you don’t want can be tens of millions of times more powerful than the signal you do want, so these are incredibly sensitive systems.”
Second, unlike nonradio chips, RF designs can’t be easily tested on a computer. “With some other technologies in the high-tech world, you can actually very accurately simulate it on the computer and know what you’re building before you build it,” says Dave Shepard, chief executive of Sequoia Communications, which designs chips for cell phones. “In RF, you can’t do that. You’re not relying on the software to do the work for you. It’s really the engineer’s brainpower.”
Finally, RF designers must build chips that don’t drain too much from a battery or take up too much room in the cell phone, so the business itself is extremely competitive. “The RF integrated circuit business [has] razor-thin margins,” says Galton, the UCSD professor. “These companies are under intense pressure to minimize the circuit area and power consumption of their chips. On the other hand, volumes can be enormous, so a winning design can be extremely lucrative.”
Three start-up companies—Quorum Systems, RF Magic and Sequoia Communications—are taking different approaches in the RF chip market, which is expected to reach $8 billion in sales by 2008, according to industry experts. RF Magic aims at broadband consumer electronics and already has shipped more than 2.5 million chips, Foley says. Its silicon is used in digital broadcast satellite equipment, broadband wireless modems and cable set-top boxes. With satellite television, RF Magic’s chip allows home owners to watch a show and record a different show on their TiVo, all through one cable from the dish to the set-up box. Without the chip, a second cable would be needed to record on the TiVo.
At Quorum Systems, which has raised $24.3 million in venture capital, the focus is mostly on cellular markets and wireless networking equipment, says Bernard Xavier, CEO, president, and a co-founder of the company. But eventually television signals also might be incorporated into its radio chips. The buzzword is multimode chips. Today, if a cellular company wanted to add a feature to a phone, such as global positioning or streaming music, it would have to add another chip or another layer of circuitry to an existing chip. That process eats up power and takes up room in the ever-shrinking cell phone. With multimode, RF chip designers aim to pack all these features on one chip through shared circuitry. “If you look at Quorum, what’s different about us is we’re focused on the convergence of all these applications onto one chip,” Xavier says. “We converge cellular and networking and eventually television into our radio chips.”
At Sequoia Communications, the focus is more on the next generation of services coming to cellular phones, known as 3G. It is developing chip architecture that aims to shorten the time it takes to design new functions into a phone. The company has raised $20 million in venture capital and employs 38 workers. Shepard, the chief executive, says wireless design talent at big wireless companies is likely to spawn more RF start-ups, as engineers with good ideas leave to pursue projects that are too risky for large companies to fund. “Very big established companies create a feeding ground for start-ups,” he says.