Unlimited Airwaves

Dan Gillmor has an article concerning the notion of scarcity of the airwaves, which has long been a testament of faith at the FCC. Recent advances in technology may render that testament false.

Comments

Re:Described before?

[dknj]

I seemed to have answered my own question, the article i was refering to was the Ultra Wide Band. Additionally, this article seemed interesting.

-dk

Spread Spectrum combined with TDMA?

[jxs2151]

There is so much more available now than we once thought. Technologies like Spread Spectrum and TDMA and CDMA have the potential to unlock vast amounts of underutilized spectrum.

Think about it:
Vinnie's Cab Company in Newark, NJ is allocated the frequency of 152.125 Mhz and makes use of it maybe 15 total minutes a day. We can improve on that and also allocate 152.125 Mhz to Joyce's Cab Company in Denver, CO so you get more use out of the available spectrum by dividing it geographically.
Now how about if we could take every cab company in the US, regardless of location, and not assign them any frequency at all but provide them with technology such as CDMA or Spread Spectrum that assures no interference. In essence you have freed huge amounts of the 'limited' spectrum for other uses. Once spectrum is freed there is no longer the psychological or bureaucratic limitation on new ways to use spectrum.

The FCC is regulating based on the limited resource model and it is now outdated. Time for a change. With the way that new technologies conserve spectrum we are using a fraction of what is theoretically available.

Re:New?

[Gordonjcp]

Bandwidth. Right, consider a *very* precise signal, at exactly 100MHz. Look at it with a spectrum analyser - you'll see a very very thin spike. Once you modulate it, it will spread out on either side. So, for example, a conventional voice radio channel, with a bandwidth of 300-3400Hz roughly (the same as an ordinary phone) will spread to 100MHz +/- 3.5kHz. To leave room for different channels you need a "guard band", so the channels are usually 10kHz apart.

This is one of the reasons Morse code is still so popular with amateur radio enthusiasts - you can send extremely narrow band signals that allow you to communicate fairly quickly. If someone is really good at it, they can communicate almost as fast as speech, over a channel a few *tens of Hertz* wide.

Re:New?

[SomeoneGotMyNick]

That's an understandable perception and theoretically will work. Consider this enlightenment and not a flame.

Current FM radio modulates the signal above and below the designated carrier frequency. Therefore a 20Khz signal (peak of human hearing) will modulate a 95.3MHz carrier between 95.28 and 95.32MHz. IIRC the full 40Khz deviation accounts for both channels of a stereo broadcast.

There's additional use for Broadcast radio. I forget where I saw it, but I believe there is an offset from the designated frequency to place a mono only 20Khz band away from the stereo part of the transmission for mono FM radios to pick up properly. This may however be an outdated use of the extra bandwidth.

Additional bandwidth can be used for other data/audio signals to be carried independent of the main broadcast for Broadcast FM plus 'padding' between stations.

Re:New?

[Gordonjcp]

The stereo system used on FM radio works like this. Your normal FM carrier has a mono signal (L+R), and modulated on a 38kHz subcarrier is the difference signal (L-R). The mono radio can hear only the L+R signal, the higher band L-R being filtered out. In a stereo decoder, the L-R signal is demodulated, and mixed with the L+R to give L only, which is then subtracted from L+R to give R only.

It's sometimes called M/S (mid/side), so we can express it like this:

M=L+R, S=L-R when transmitting.

L=M+S, R=M-L

Clear as mud, right?

Re:It's a matter of finding things again...

[Coz]

Good synopsis.

One of the problems with his proposal is that there's so much old equipment out there that doesn't tolerate interference - so you have to retire a "use" for a band before you can free it up for spread-spectrum, or the SS signals generate noise for the existing users, who're using 20 year old technology. That's been done successfully once, to my knowledge - when the FCC made the wireless carriers pay to relocate and reequip point-to-point microwave relays in the A band. $Billions spent.

Of course, if you want to go spread spectrum, you have to have codes - and the more users you have, the more complex the codes need to be - there is a limit to the number of users you can have, based on the uniqueness of any given code set.

Gawd, I'm having a grad school flashback!

Re:Radio kindergarten Part II

[Gordonjcp]

Because the higher the frequency of the modulation, the further from the centre frequency you go. For example, given the precise 100.000MHz signal above, if you modulated it with a 1kHz signal, you'd get a spike at 99.999MHz and a spike at 100.001MHz, as well as the original 100.000MHz one, right? If you then increased it from 1kHz to 10kHz, the spikes would be at 99.99MHz and 100.01MHz. If you increased the signal to 1MHz you'd have spikes at 99MHz and 101MHZ.

I recommend that you get hold of the ARRL handbook from your local library, or indeed the RSGB book if you're in the UK. These are the standard works on amateur radio, and explain all these things far better than I can....

Re:This will probably come out "ignorant"...

[Permission Denied]

To implement what you're talking about, you need some sort of time sync. Basically, what you're talking about is very similar to how ethernet works, but there are two fundamental differences between ethernet and radio at one specific frequency: with ethernet, you can both listen and broadcast at the same time (which means you can detect when you have a collision, thus CSMA/CD); with radio, you cannot listen while you're transmitting because your signal will drown out any incoming signal. This means you need some way of saying "OK, you can broadcast now." You could do this either on a time-slice basis (like 802.11) or with a token-passing scheme (and there are some wireless protocols that do token-passing). Another problem is that you might have three radios, like this:

A ---- B ---- C

where A is four dashes away from B and B is four dashes away from C. Suppose that a signal "lasts" for five dashes. That means A and C can't see each other, but B can see both. This brings up other nasty problems with simple protocols (and wireless protocols like 802.11 deal with this).

So, the simple packet-addressing scheme won't work for two-way communication. As for one-way communication, there's no need to "label" the recipient of a broadcast; radio is inherently broadcast, so everyone can hear everything anyway.

Re:i dont get it

[CarlDenny]

> Why cant we just use higher and higer frequencies? 2GHz full? Use 20GHz? Or 50GHz? Or a googlehertz?

Because, the higher the frequency, the shorter the wavelength.

And the shorter the wavelength, the less "penetrating power" the signal has, and the more the signal is absorbed by intervening walls/clouds/.../and eventually air.

In short, 100Ghz signals can't even make it across a room without getting in trouble.

Re:i dont get it

[bugg]

Your reasoning is slightly flawed. Visible light is 400nm-700nm which works out to be 7.5*10^14 Hz and 4.3*10^14 Hz- much larger than 100GHz. I have no trouble seeing the light from my lightbulb across the room.

Re:i dont get it

[Anonymous Coward]

Well, the problem with low frequencies such as 5000hz is noise. There are two main types of noise present in RF and electrical signals. White noise's strength is independent of frequency. And the killer of the low-frequency range, 1/f noise. 1/f noise increases as the frequency decreases-- basically the amplitude is proportional to the multiplicative inverse of the frequency. It kills the low frequency bands because the noise masks the signal. Increasing signal strength improves the signal/noise ratio but there are practical limits to the amount of power used, i.e. battery capacity, electricity cost, human safety, etc.

Space-Time Coding

[femto]

There seems to be a lot of misinformation flying about on this topic. Reed really is talking about a break through, not just about squeezing in more channels by adding repeaters or optimizing the gaps between frequency multiplexed channels.

Shannon's Law says that for a given signal to noise ratio, there is a maximum error free bit rate which can be supported. Recent advances have shown that Shannon's law applies on a per antenna basis. If your transmitter and receiver each have 'n' antennas, it is possible to transmit 'n' times the information which one tx/rx antenna pair can transmit. To my knowledge, there is no limit on how large 'n' can be. Researchers are currently trying to figure out if there is a limit.

Repeating myself in different words. It not only matters at what frequency you radiate (frequency diversity) and when you radiate (time diversity), it also matters where you radiate from (spatial diversity). Since available time and frequencies are limited, it was thought that spectrum was limited. Add space (of which there is lots) to the equation, as recent advances did, and the available spectrum becomes unlimited (though new boundaries may show up with more research).

This is not pie in the sky stuff. Space-Time coding techniques allow such capacities to be realised. Bell labs have already demonstrated a working system in the lab.

John

Re:Ten percent of the spectrum needs to be open

[Kirkoff]

It's called Amateur Radio.

Which is generally not open to experimentation, testing or demonstration of new methods and technologies.

Umm, what Amateur Radio are you dealing with? The license I hold does allow me to experiment with new techniques. For example PSK-31 was invented in 199\8 or 1999 and is widespread. Yes, there are some limits, for example, you can't use more bandwidth then a voice channel on HF for new modes, but that's just common sense. Up above 3GHz, you're pretty much open to anything you want, including spread spectrum. You can do SS on everything above 70cm. If it's currently not allowed, the FCC does grant 6 month at a time experiment permits. If it works, the FCC will allow it. For example, see the ARRL's experiments in the 60m band.

--Josh