A comment from a random on Twitter has brought to my attention that some folks don’t quite understand the limitations imposed by sending data via radio links. This lack of understanding is aided by misinformation about NBN and fibre optics from shock-jocks and Ltd News.

While I could go into the nuts-n-bolts of why wireless can’t ever match the data capacity of optical fibre, I’m going to simplify this as much as possible for the lay audience.

First, let’s talk about radio signals. Any radio transmitter, whether it will eventually send TV, audio or data, begins with an oscillator which generates a carrier wave on the operating frequency. However, the carrier doesn’t occupy only the operating or centre frequency. It has a character known as bandwidth, as well:

An unmodulated carrier is quite narrow, only a few tens of Hertz wide. When modulation is added, the signal becomes wider. A carrier which has been amplitude modulated with low fidelity voice information, as in CB or ham radio, is about 3kHz wide, or 1.5kHz above and below the centre frequency. An AM broadcast band signal is 15kHz wide- 7.5kHz above and below the carrier frequency. An FM broadcast band signal is about 150kHz wide- that’s right, 75kHz above and below the carrier frequency.

The more information you modulate onto a carrier wave, the wider the signal will become. This is why high-fidelity FM radio sounds better than AM- a broader range of audio frequencies is modulated onto the carrier. Better quality sound, but a much wider signal. A single analogue TV signal, with video and audio, is 6MHz wide- that’s 6 times the width of the roughly 1MHz wide AM radio band (550kHz-1600kHz).

If two stations try to transmit on the same frequency, they will interfere with one another and neither signal will be intelligible at the receiving end. The same will occur if the skirts of the transmitted signals (known as sidebands) overlap. Consequently, radio signals’ carrier frequencies must be coordinated so there is enough separation between them to prevent interference.

The same thing happens when you modulate a carrier wave with digital data. A slow data rate will produce a rather narrow signal. Digital data sent via packet radio at 9600 baud will be about the same width as a voice signal- around 3-5kHz wide. The faster you go, the wider the signal.

Radio spectrum is a finite resource. You can only put so many stations on the AM or FM band before they interfere with one another. Same with digitally modulated radio signals- there’s only so much radio spectrum bandwidth available. Similarly, there is a finite limit to how many wireless data streams will fit into each MHz of radio spectrum bandwidth.

Yes, it’s possible for two wireless data users to share a single frequency- but not at the same time. Station 1 sends and receives a few packets of data, then goes silent. Station 2 is then free to use the frequency. If they both try to transmit and receive at the same time, they will interfere with one another and neither will successfully exchange packets with the greater (or wide-area) network (WAN), aka internet. This is called time-division multiplexing. The downside is that with each station sharing the frequency equally, each station can only transmit and receive, at best, 50% of the time. The stations might be capable of 300Mbps (as in 802.11n WiFi), but their effective maximum rate will be 150Mbps. Start adding more stations on the same frequency and you quickly see what happens- data rates drop like a rock as all stations try to share the frequency (or ‘channel’). Wide-area wireless (3G) users are already well and truly familiar with this problem.

It is possible for users to share the same frequency if there’s a large physical separation between the systems and the range of the systems is limited.  Low power wireless systems can be set up on microwave frequencies, which have a very short range, so that a user on one city block won’t interfere with a user on the next block, who happens to be using the same frequency.

Don’t be confused by generation identifiers such as 2G, 3G, 4G, etc. It’s all data via radio link. The modulation method or frequency used doesn’t get around the issue of co-channel interference. The co-channel interference problem is a function of physics and has nothing to do with and will not be resolved by upcoming wireless data communication standards, such as 4G. If you need to provide fast network access to a large number of users in a small physical area, over-the-air radio spectrum bandwidth is a bottleneck which must be eliminated. The solution is modulated lasers, with their light ducted through optical fibre.

In optical systems, a laser is modulated with data and the data is decoded on the receiving end. However, laser light passing through free air is readily blocked or diffused by airborne moisture- rain, fog, etc. While it’s possible to send data on a laser light beam through air, it’s not very reliable. The solution is to contain the laser light within a glass optical fibre.

The greatest thing about a glass optical fibre is that you can jam more than one laser light signal through it. Lasers can produce light of different wavelengths, or colours. These can either be generated in a single multi-coloured laser, or the light from several lasers of different colours can be combined with a prism and fed into a single fibre. The different coloured, modulated laser light streams are then separated on the far end of the optic fibre and fed into decoders so the data can be extracted and sent on to its intended destinations. This means is known as wavelength-division multiplexing. It permits multiple parallel data streams to co-exist on the same fibre, as well as provides future expansion capability. The number of different coloured lasers is limited only by the precision of the prisms.

By now, you’re beginning to see why wireless, aka data over radio links, is not a suitable solution for many thousands of network users in a small geographic area. A wide-area optical fibre network, as is being built for Australia’s NBN, is not only the best solution, but the only solution for providing connectivity as is needed now- and into the future, as it is highly scaleable.

This is not terribly complex stuff. Anyone with enough on the ball to use the internet can understand these concepts. If anyone is trying to convince you that some new wireless communications standard will magically solve the problem of co-channel interference or dramatic slowing due to time-division multiplexing, they either don’t know what they’re talking about or are trying to confuse you for the purpose of driving their political goals (I’m looking at you, Alan Jones and News Ltd).

-weez