Friday, June 02, 2006

XG - A Killer New Wireless Technology??

A startup called XG has demonstrated a wireless technology it claims is hundreds of times more efficient than existing wireless technologies. According to the patent behind it, that is.

XG's demonstration used a transmitter not unlike a cordless phone base station, operating in the unlicensed - and crowded - 900MHz band, to send a 3.7Mbit/s data signal to a radius of 18 miles across the suburbs of Miami, using 50mW and an omnidirectional antenna. The system carried 7.4 Mbit/s per MHz per Watt according to Professor Schwartz, professor of electrical engineering at Princeton University. By comparison, GSM would have around 0.0058, and CDMA/EV-DO about 0.0085 Mbit/s per MHz per Watt.

I suggest you take a look at the patent the founder of XG filed on his core technology, to get a better idea of what the excitement is all about.

The technology is very clever and takes the capabilities of radio to their theoretical limit, where every RF cycle can be a bit. It makes very efficient use of the radio spectrum, which is why there is so much interest. Normal radio works by modulating a carrier (the RF or radiofrequency) with a lower frequency (the IF or intermediate frequency) that is in turn modulated in a variety of possible ways by the data. This generates a shmear of radiation around the carrier frequency (called sideband) which is why spectrum is allocated in bands - anything else operating in that band in the same area creates noise that interferes with the reception of the signal, so spectrum bands are managed to reduce interference between users. At the receiver the received signal is demodulated with the carrier frequency to recover the IF with the data in it. The maximum datarate is therefore related to the IF, not the carrier frequency.

All recent developments have been aimed at cramming as much data into as small a part of the spectrum as possible and recovering it reliably. The current state-of-the-art is probably represented by the complex Wimax and US 3G technologies.

XG's technology requires as a minimum just two closely-spaced radio frequencies to transmit a bitstream, which can be at the same datarate as the radio frequency. It sends information by switching between the two frequencies at cycle boundaries (imagine a sine wave with some cycles stretched) and has no need of an IF stage. The economics are interesting - because it requires electronics that operates at the RF rate, I suspect (though I'm no radio engineer) that this is more expensive than the electronics of the IF stage. However, it is definitely hundreds of times more efficient than current radio technologies. It also requires much less power.

Will it kill Wimax? No more than it will kill any other form of radio communication. It takes spectrum cost out because you can cram a lot more data into the same bandwidth. I don't think the companies that have paid billions for portions of the radio spectrum will be too pleased - they will have to write down their spectrum asset. However, that is sunk cost paid for by shareholders, who are used to telecoms losses by now. Wimax operators in general have not bought spectrum, so can deploy this technology quite happily.

To be cost-effective the XG technology has to be converted to silicon, and the economics of silicon are volume-related. I see it being used first in areas where its low power, high bandwidth, small spectrum requirements and range are at a premium and cost will not matter. This will be military applications and remote sensing applications. As the costs of the silicon come down it will head first towards remote control applications and then towards volume markets like RFID rather than targeting relatively low volume applications like Wimax. However, if the technology is as simple as they make out I can see it appearing everywhere, purely to free up the spectrum.


Post a Comment

Links to this post:

Create a Link

<< Home