Coming Soon: Ultra Wide Band

JScarpace writes: "Robert X. Cringely has a new article in which he talks about Ultra Wide Band (UWB), a new wireless communications technology which may allow wireless networking speeds up to a gigabit per second. Read the article."

Comments

Bah! TANSTAAFL.

[phliar]

This keeps coming up every so often.

UWB is ... a series of very short electrical pulses ... on ALL frequencies simultaneously.

UWB requires ultra-low power... [it is] a signal that can't be detected and doesn't interfere.

Here goes: Bullshit!

He (or whoever he got this story from) needs to read a little bit of signal processing. Yes, it sounds very nice, and you can build it, and it's all true... if there's only one such device. You see, what this does to other users of spectrum is raise the noise floor just a bit. No big deal.

But what happens if there's a whole bunch of these devices? Well, let's say you're an FCC licensed user of spectrum. You've been allocated a certain bandwidth. Your channel capacity depends on the bandwidth and the noise floor. If your noise floor goes up, your channel capacity goes down.

Where did that lost channel capacity go? It's being used by these "UWB" devices. As evil as the FCC is, we do need some arbiter of the EM spectrum.

TANSTAAFL, folks. Go read Shannon.

Cringely is an idiot.

the UWB landgrab

[markj02]

There is only so much bandwidth to go around. You can allocate it one way or another. We have chosen to allocate it by frequency channels. That's simple and low-tech.

UWB doesn't give you any unused spectrum, it just degrades that the spectrum there is uniformly for everybody else. In small amounts, that may not be a problem, but in big amounts it is. Think of it like trash: the occasional piece of paper on the street isn't a problem, but if everybody dumps their garbage on the sidewalk, it's a big problem.

If UWB were ever widely deployed, you can think of it as generating noise kind of like one billion light switches turned on and off many times per second. It's best to put a stop to that before it starts. Or, if we are going to throw out frequency based allocation, let's do it consciously (and let's wait for the UWB patents to run out before we do it).

Almost free lunch

[XNormal]

Yes, Cringely doesn't understand 99% of the technology he writes about. That does not make the technology bullshit.

UWB is real. It's as close as it gets to a free lunch, and Claude need not turn in his grave.

you can build it, and it's all true... if there's only one such device

Not correct. UWB devices share the spectrum just fine. In fact, it's a far superior way to share the spectrum than narrowband frequency allocations.

The problems start when different devices use very different power levels: GPS uses extremely low levels, TV stations use very high levels and almost anything is at very high levels if you are close enough to the transmitter.

Spectrum sharing by frequency allocation provides very good separation between bands that use widely differing power levels. It's not too difficult to build filters that reject out-of-band interference by 100db or more. With ultrawideband, the rejection of unwanted signals cannot exceed 40-50db. UWB will work very well if all narrowband communications below 1GHz are shut down. Since that will never happen it will probably remain limited to very low power levels and certain niche applications.

Here's what might happen if all narrowband transmissions *are* shut down:

UWB cells for "last 10 miles" delivery, combined with long range fiber and satellite infrastructure could bring 100kbps to almost any person on earch and 10mbits/second to anyone living in a city. The terminals will use very little power and can have long battery life. Location tracking with 20 centimeter accuracy will be available anywhere in a city, including indoors.

How is all this possible with just 1GHz of bandwidth? The utilization efficiency of spectrum should not be measured in bps/Hz but rather in bps/Hz/square Km. Today's cellular infrastructure uses a very crude form of frequency reuse to optimize this capacity. IS-96 CDMA barely begins to utilize the real advantages of spread spectrum with a bandwidth of 1.25MHz. With 1GHz of spread spectrum things start to look different. And it's not just the bandwidth: 1GHz at a center frequency of 15GHz can only be use for line-of-sight communication. If the 1GHz band has a center frequency of 700MHz it has much better propagation and is immune to fading.

Of course, this will never happen. But not because it is mathematically or technologically impossible.