SAN JOSE, Calif. An analyst predicts a shakeout may lie ahead for the many companies designing ultrawideband silicon for the wireless USB interface. Startup Staccato Communications (San Diego) aims to escape that fate with its Ripcord2 announced Tuesday (Sept. 2) that uses 65 nm process technology to hit size, power and cost levels it hopes will attract the kinds of volume uses that have eluded the technology to date.
At least a dozen mainly venture-backed companies have been pursuing UWB for uses such as wireless USB links on consumer and computer gear. To date, relatively high prices as well as performance and regulatory hurdles have limited the market for the wireless links.
"The problem has been the chips are too expensive," said Brian O'Rourke, principal analyst at InStat (Scottsdale, Ariz.). "They all came out at about $15, and by the time they are in an end product that becomes a huge cost adder, and there's been no killer app to drive it," he said.
Less than 100,000 wireless USB chips shipped in 2007. To hit volume markets, UWB chips need to hit a similar $5 price point that drove growth for Bluetooth silicon, O'Rourke said.
"That could happen for wireless USB in late 2009 or early 2010," said O'Rourke. "This year will not be a breakout year for them," he added.
Marty Colombatto, chief executive of Staccato, would not disclose pricing for Ripcord2 which is sampling now and will be in production in the first quarter of 2009. However, he did say that "this is the chip that will enable us to get to that sort of [$5] pricing.
"We're leapfrogging 90 nm to get the smallest die size and the lowest power and cost, Colombatto said, referring to his first-generation 110 nm chip. "Our story has always been about integration and aggressive use of process technology to enable the best pricing," he added.
The 5mm x 5mm Ripcord2 also supports band groups 1, 3 and 6—essentially covering the 3.1-4.7 and 6.3-8.9 GHz spectrum bands to meet diverse UWB rules laid down by regulators in Europe, Japan, Korea and the U.S. Staccato believes the range will also cover proposed spectrum rulings in Canada and China.
The chip anticipates extensions to the wireless USB specification now in the works. Wireless USB 1.1, expected to be complete in mid-2009, will support similar bands. It will also include new techniques to lower power consumption and adopt the approach used by the Near Field Communications specification to automate how devices recognize each other.
Ripcord2 consumes about 600 mW peak in transmit mode and about 200 mW on average, thanks to a half dozen new power management techniques, said Jeff Chang, vice president of marketing at Staccato.
In terms of performance, the chip is currently delivering throughput up to 100 Mbits/second. It should reach up to 200 Mbits/s as engineers optimize host control designs over the next three months, said Colombatto.
The first-generation wireless USB designs were criticized by some testers because they had performance of 50 Mbits/s or less, a fact that vendors blamed in part on early controllers that translated signals to and from the wired USB protocol.
At the recent Intel Developer Forum, competitors such as Alereon showed similar second-generation chips hitting performance levels of up to 150 Mbits/s and supporting the same three band groups. Wireless USB devices on display at IDF included a notebook from Fujitsu-Siemens, an LCD monitor from Taiwan's Asus and a wireless dock from Kensington.
Staccato is not banking solely on wireless USB for design wins. Ripcord2 also supports high-speed Bluetooth, Internet Protocol and other protocols.
Despite the advances, analyst O'Rourke predicts a tough year ahead for the ultrawideband startups.
"This could be a make or break year for many of these companies, some of which have been around since before 2004 and have investors looking for an exit strategy," said O'Rourke.
"We are funded through the end of next year, and an internal round is in the works as we speak with our [existing] seven investors," said Colombatto. "I think we will see UWB and wireless USB taking off this year and next with the major ramp up coming in the second half of 2009," he added.
"These things always take longer than people thought to get the infrastructure and regulatory and silicon issues resolved," he said, noting Bluetooth started taking off in 2004, about the same time some pundits predicted its death.
Colombatto shrugged off competition from a growing group of technologies including streamlined versions of Wi-Fi for short range links, a number of efforts at 60 GHz and a separate ultrawideband effort called TransferJet led by Sony.
"I think a number of these things will fall by the wayside," he said. "60 GHz is still a long way from commercial power, cost and size and has a limited use case, while TransferJet is a different animal with a range of just 3 cm," he said.