Don't kill FM for broadband: Build a new internet

Don't kill FM for broadband: Build a new internet

Summary: Ofcom is planning a digital switch-off for FM radio and eyeing the frequencies for more broadband services. An exciting alternative is already available


Although the move from analogue to digital wireless has seen many new and exciting services, the process of clearing old services is slow and expensive with little innovative in their place. So how about free ultra-reliable nationwide wireless broadband on unused frequencies that are available now?

There's not much sign that big ideas are on the agenda yet. According to The Wall Street Journal, Ofcom is considering liberating the radio spectrum that carries FM stations and giving it over to broadband. The thinking goes that with the move to digital radio, up to 50 percent of the band between 88 and 108MHz could carry wireless data; the frequencies are good at going long distances and penetrating buildings, making them especially suitable for rural areas.

Unfortunately for Ofcom, the FM band is also really good for broadcasting. With FM radios available new for a couple of pounds upwards, universally installed in cars, and capable of receiving four national BBC networks and 10 or so local and independent stations, it's no surprise that the public has been cool towards moving to the new digital radio stations. Also, it's been a long, expensive and complicated ride for the digital switchover for television: there are many times more radios than TVs in households and no real reason to turn them all off.

Band I

Ofcom's argument is that there is huge pressure on spectrum for new services, so the old must give way. Yet this has already happened. Down below the FM waveband, between 30-88MHz, is a nearly forgotten chunk of spectrum called Low Band VHF, or Band I. (FM is Band II, while DAB — digital audio broadcasting — is in Band III, at around 200 MHz.)

Old radios

Rather than eating into FM frequencies to create more wireless broadband, why not delve into the Band I spectrum? Photo credit: bestfor/richard/Flickr

Originally used during the war for Chain Home radar and Bomber Command communications, Band I became the home for two-way radios for the military, the Fire Brigade, the Gas Board, the AA and so on, though the major incumbent was black-and-white 405-line TV. The band was largely vacated in the mid-1980s when the last 405-line TV transmitters were turned off for economic reasons.

By now, the two-way radio traffic, military and otherwise, has largely departed to new, digital systems; apart from a new amateur radio band and a scattering of things like baby alarms, Band I has sat there unloved ever since. Even with gaps for the remaining services, that's getting on for 50MHz of bandwidth, silently waiting.

It does have problems. The lower the frequency, the longer the wavelength; while mobile phones use frequencies with wavelengths of a few tens of centimetres, allowing compact antennas, Band I starts with three-metre wavelengths and goes down to 10 metres. You can use small antennas, but they're not very efficient. Also, as the bottom of the band is close to shortwave frequencies, it's more affected by atmospheric conditions, including anomalous long-distance reception.

But otherwise it's really useful. In-building and past-line-of-site coverage is better even than Band II, it's not particularly prone to atmospheric noise, and it's free. You don't have to boot off The Archers to use it. Still, it's avoided commercial pressure: the other chunks of spectrum are easier to use for mass-market money-making systems, and that's where the focus has been.

It's when you divorce yourself from the model of spectrum being primarily a revenue generator and look at it as a national resource that the game changes.

Mesh network

One interesting model would be a national open mesh network (or "Nomen": why not?). Mesh networks work by each station also acting as a router, relaying data from surrounding stations and steering it towards its destination. As stations appear and disappear from the network, it adjusts the map it has of itself, so each station always knows the most effective way to send messages on.

The system could work by having individuals install a Wi-Fi-to-Band I gateway that automatically configures itself to connect to its nearest neighbours. The range is good enough that a few stations dotted around will have a footprint sufficient to make it worth others joining, and home-based systems will have plenty of room for a discreet external antenna.

One of the secrets of mesh is that the more people join in, the greater the frequency re-use and the greater the bandwidth available, so the more people who use it, the more attractive it becomes to others. Once there's a bit of interest, it will build quickly.

This would produce a large, cheap and above all infrastructure-free network that could operate largely independently of any other wired or wireless system. Gateways into the internet would provide the connectivity most people need, but hosting your own data or running your own services — even in the event of a major network failure elsewhere — would be very viable.

Even remote communities could be part of the mesh, just by pointing a directional antenna at the nearest station: the engineering is simple and cheap enough that community effort will be more than enough.

Hard to jam and easy to reconfigure, it would be unperturbed by anomalous propagation, as stations appearing and disappearing are natural to mesh topology. In terms of a useful addition to the national infrastructure, it would add a new, independent and supremely flexible layer of connectivity, and it would provide the lowest possible cost of access for the digitally disadvantaged.

Low cost

How low a cost? It'd be an extra chip — if that — in a Wi-Fi router; all the protocols and engineering have been done and are freely available, and with no infrastructure there are no bills to pay. With nobody trying to squeeze money out of the system, it will be free to develop along the lines of maximum utility, effectiveness and efficiency, and the engineering is well within the reach of the sort of open hardware experts who have already built an open-source GSM system. A single payment — £50, say — and you're done.

It would never replace the megabits of 3G or cable, but then that wouldn't be its job. Cheap, reliable, usable and flexible: it would be the original internet reborn.

And we could start doing it tomorrow, by Ofcom declaring a chunk of Band I open for business with the lightest of regulatory touches.

All it takes is vision and will. Over to you, Ofcom.

Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.

Topics: Broadband, Emerging Tech

Rupert Goodwins

About Rupert Goodwins

Rupert started off as a nerdy lad expecting to be an electronics engineer, but having tried it for a while discovered that journalism was more fun. He ended up on PC Magazine in the early '90s, before that evolved into ZDNet UK - and Rupert evolved with them into an online journalist.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • This sort of reminds me of the days when CB radio was legalised. Trouble was the rules stated that antenna had to be a size LESS than a half wavelength (half wavelength being the simplest and most efficient size) so loading coils had to be used. The regulations also forced the coils to be at the base thereby taking out the part of the antenna best for radiation.

    If a mesh of computers on FM Band 1 were to be allowed, let's hope the government doesn't screw things up with stupid antenna rules to annoy people into abandoning their FM kit.

    I take up the point about the cost, you suggest £50. Most folk would need to get a device (£££) which connects their network to their antenna (£££), and depending on their physical location, this would cost more for some than for others. Otherwise, your post has a lot of merit. If this were to happen I'd be right in there. It would be great for people with no phone lines too.
    Fat Pop Do Wop
  • NYNET (North Yorkshire broadband) has been given £16 million - I would have thought that the long distances involved in the is very rural county would get well spent on this Band 1 rather than the digging up roads for fibre extensions?
    It'd be the perfect test-bed - Ofcom and Nynet how about it?
  • I would like to correct several inaccuracies on your editorial. Firstly, TV band 1, which is vacant(?) is not, and does not extend fom 10m to 3m. Let's be engineers and talk frequencies, not as politicians and confuse people by talking wavelengths. So lets call it 30MHz to 100MHz. Firstly the area 30-40MHz is special, many TV's, radios and many other systems use this area for IF frequencies and these frequencies should be guarded.

    Secondly 88-100MHz (and on to 108MHz) is VHF Band II, and is internationally allocated to and should be retained for FM radio.

    Third, as you rightly say 50-52MHz and 70-71MHz are internationally allocated amateur bands and shall remain so.

    Many of the remaining frequencies are already allocated to other services.

    Oce we get past all these allocations, we are left with 54-68 MHz, allowing for guard bands, which is hardly enough for any form of wideband data. Secondly can you imagine any form of "home hub" with an aerial of 1m or longer attached. Can you imagine the planning dept of your local authority allowing any form of external aerial without legislation to force them to!

    Please leave Band II alone, Lots of folk rely (like me) on it. Also DAB plain doesn't work. I live 3 miles from my county town, almost on top of a hill, yet I cannot receive my local (BBC) radio station on DAB.

    If we want to provide a decent free radio data service, there is a lot of spare Band IV/V space, the so called digital dividend, and if properly managed, could cover the UK in a mesh system. But of course, it is a cash cow, just waiting to be milked..........
  • Sight. Line-of-SIGHT.
  • Grid has problems.
    Firstly - if the average 'hop' is 5 nodes, then your throughput falls to 1/5th, as 4/5ths of the time you're dealing with other peoples traffic.
    The second problem is that radio is a shared resource.
    You can only avoid transmitting at the same time as another node if you can receive them.
    But, the number of nodes you can't quite hear grows as fast as the distance to them increases - meaning the distant murmur of distant nodes grows.
    This is especially bad as 30MHz transmissions can go many miles, so you can be interfered by hundreds or thousands of nodes.
    Consider a grid network as a football stadium. You want to send a message to the other side. If it's quiet, you can just shout, and get your message to the other side in 4 or 5 relays.
    But if everyone is trying this, you can only just about hear your neighbour, and you need your message to go through a few hundred people. As they are also sending out messages, and dealing with other peoples messages, bandwidth drops through the floor.
  • Neither Band II or band I are at all suited for use for providing Broadband.

    They each amount to, what, 30MHz 50MHz at most.

    That's just not enough to do anything useful with.
  • Thanks for those responses.

    I agree completely that Band II should be left alone. It's very good at what it's doing and the economics of switching over would harm a lot of people.

    One of the key aspects of effective mesh networks is that you adjust your power to an appropriate level for your neighbours. Very dense areas have lots of microcells with lots of low power, giving lots of frequency reuse. Don't forget, all these things are networked together and can co-operate, even when there's anomalous propagation - a station appearing from 400km away looks much like a temporary neighbour.

    I know what the top end of HF/bottom end of VHF sounds like when propagation is running amok (I once talked to a policeman in South Africa while I was carrying a VHF army backpack radio over a tor on Dartmoor) but that's a very poor guide to what you can do with the band when stations are co-operating. You can assume stations know where they are and can communicate indirectly, and that an instantaneous and accurate high resolution propagation model can be created and shared.

    There are many interesting cognitive/white space techniques that allow a lot of spatial reuse; I hesitate to think about MIMO at longer wavelengths, but that may also be a useful tool. Nonetheless, keeping this above 50MHz would be preferable - although with decent adaptive protocols, there's no reason you can't use as much of the band as you find works for you at any one time.

    This is 21st century radio, after all. What the murmur would look like when nodes can identify and adjust themselves, I don't know. It would be fun to model.

    As for the bandwidth not being enough to do anything useful with: you can get upwards of 2 (bit/s)/Hz per site these days, which gives you enough to do a lot of interesting things, even if you and your neighbours are spending most of your time relaying data.

    It would look different to today's Internet. The question is: would this provide a useful service? I think that in terms of making something that is useful, robust and cheap, yes it would.

    Modelling it would be worthwhile, at least!

    (Oh, and how many radios and TVs are still using 30/40MHz IFs? Has that changed with DSO for TVs? Most radios I know use 455kHz and 10.7 MHz, mind)