Pulse-Link has produced the first integrated circuit for ultra wide-band (UWB), the high-speed software-defined radio technology which could provide very high bandwidth in wireless, wired and cable systems.

The company’s efforts to bypass the UWB standards stand-off through a common UWB "dial-tone", reported here last month, could lead to the formation of an industry coalition later this year.

Pulse-Link's silicon delivery is well-timed, as the company plans to promote UWB at the ITU's meeting in Boston in June. Although UWB promises very high bandwidth (see our feature), the technology is radically different from conventional radio, using "chirps" of broad spectrum signal, making it hard to legislate for.

Although ITU operates independently, approval of the technology there could make it easier for standards makers to operate and for regulators in different countries to find ways to allow it (as the US FCC has done), said Bruce Watkins, Pulse-Link's CEO. "The ITU has no authority over the IEEE, and I don’t think either body would want to see that happen," he said. "The ITU sets parameters under which to operate, and the IEEE decides which is the best tech that can operate under those parameters."

The company has produced a test version of the analogue half of a two-chip system, which is expected to evolve into a product by the middle of 2005. Because of the power demands of early systems, and the regulatory issues, the first applications are likely to be on wired networks, said analyst Dean Bubley of Disruptive Analysis: "So far, the focus has been on UWB as a Bluetooth killer or a wireless replacement for USB, but there may be more traction on fixed UWB, such as cable networks where one company owns the gear at both ends of the connection," he said. "Spectral efficiency is just as important on copper."

Media-friendly
Pulse-Link is avowedly media-independent, with demonstrations of WLAN, powerline and cable TV systems under its belt. All media will be driven by the same chipset, said Watkins, with cable TV links going up to 1Gbit/s and powerline connections going to 200Mbit/s.

"What we are doing is not an alternative to the wireless USB applications pursued by other UWB companies but an entirely different application and market," said Watkins. "Just as Bluetooth and WiFi are two entirely different narrowband technologies that use the same 2.4GHz frequency, what we are introducing is a uniquely differentiated application of UWB."

Watkins expects at least another two "spins" of silicon before the product is ready, but has no doubts about achieving this: Pulse-Link has just announced $30 million in venture capital funding. "We are going to have to identify whatever problems we have, come up with fixes, and go to respin," said Watkins. "We believe we will be ready somewhere towards the early part of Q3 2005. But we will do everything under our skins to get there faster."

UWB is making progress, said Watkins: "I am very rarely having to evangelise on UWB, people are getting it." And the ITU is no more hostile than the early days of talking to the FCC (which has approved UWB at low power), he said.

Shake on it
Pulse-Link has proposed its ideas as the basis for the IEEE's 802.15.4 low-speed wireless PAN, as well as a specification for that standard, which could also serve as a common signalling mode (CSM) for systems to use in handshakes before setting up other radio communications, a kind of UWB dial-tone that machines could use to negotiate the most appropriate communications protocols from a potentially limiltless set of standards.

The CSM approach would allow multiple UWB protocols to be supported, and developed, with machines installing software for each protocol as required. Pulse-Link is a member of both coalitions in the warring 802.15.3a group trying to define a UWB LAN.

After the ITU meeting is out of the way, Pulse-Link is likely to turn its attention to building some sort of UWB forum organisation. "Realistically, this is at the 'we would like to do it' stage," said Watkins. "We are actively working on the development of a forum. Maybe a better idea will come up, but we are following that idea." The guy to do the following will apparently be new marketing vice president Tom Kovanic, newly arrived at Pulse-Link from wireless chip maker Magis (which went bankrupt and was sold to Sanyo earlier this year), and a veteran of forums in both the wireless and DSL worlds.

"No-one has market share or installed base, so let’s deal with any issues on the front end," said Watkins. "We have some IP to contribute, but we are not interested in owning it all. We don’t want to scare people away from the idea."

If it happens, the trick will be to decide whether this is a UWB forum, pushing the technology, or a CSM forum, gathering support for a basic standard which Watkins sees as a bridge to provide a "wake-up mode" for truly software-defined radio, generations beyond UWB, which adjusts itself to the available spectrum and the job in hand: "CSM could support the future evolution of all communications technology," he said. "That is a very tall order. So let's start with UWB."