Tuesday 27 June, 2006, 08:13 - Radio RandomnessLooking back at the analysis of WiFi antenna performance I conducted recently, it struck me that to maximise the performance of a Wireless LAN there are two factors at play. One is the strength of the signal, clearly enhanced by the higher gain antenna. The other is the amount of background noise. It is not for nothing that the quality of the link from the wireless access point to the computers is measured in terms of 'signal to noise'.
Wireless LANs (at least variants 802.11b and 802.11g) operate in the frequency range 2400 - 2483.5 MHz (this extends to 2495 MHz in Japan only, and not all of the band is available in all countries, noteably France which does not have access to frequencies below 2450 MHz), known as the 2.4 GHz band. This band is not exclusively set aside for use by WiFi equipment, but is in fact shared with many different users. The transmitters of each of these users will increase the overall background noise in the band and if strong enough, will cause direct interference.
The 2.4 GHz band is classified as an 'Industrial, Scientific and Medical' or ISM band, meaning that it can be used by a range of non-communicating radio transmitters such as microwave ovens, industrial paint and biscuit drying machines and medical thermal heating devices. These ISM devices typically use very high powered transmitters (900 Watts plus even for a domestic microwave oven) and thus have the potential to cause enormous amounts of interference in their immediate proximity. However the band is also shared with a number of other users including the military, aeronautical radars, wireless video cameras (both professional and domestic), radio amateurs and a whole plethora of low-powered devices in particular Bluetooth. And, of course, other WiFi systems.
If we want to maximise the range of our Wireless LAN installation, it is therefore important to try and select a frequency (channel) which contains the lowest possible number of other users and thus is likely to have the lowest possible amount of interference or noise present.
Wireless LANs, based on the 802.11b/g standard offer us the option of 13 different channels depending on which country we're in (14 in Japan). However these channels are not independent of each other, they overlap very significantly. The 13 European channels have centre frequencies from 2412 to 2472 MHz inclusive, spaced at 5 MHz intervals. However the bandwidth occupied by a transmitter is 22 MHz. This means that transmissions on channel 1 actually extend from 2401 to 2423 MHz. This overlaps with transmissions on channels 2, 3, 4 and 5, meaning the next 'clear' channel is number 6. Thus if our neighbour is using channel 1, we should only use channels from 6 upwards if we are to avoid direct interference. Likewise channel 6 overlaps with all channels from 2 to 10! Thus if we are to avoid interference we can only really use 3 channels, either 1, 6 and 11, or if we have the option, 1, 7 and 13 which will give a bit more separation and hence lower interference. The diagram below attempts to illustrate the situation.
But is there any reason to select any of these channels over another? Which ones might have the lowest inherent level of interference and noise from the other sources in the band? Well of the other users, Bluetooth uses the whole band, so there's no particularly better place to go to be to be protected from this. The military and radio amateurs typically use frequencies from 2400 to 2450 MHz, affecting channels 1 to 10 inclusive, however activity is rare. Certain short-range applications (in particular equipment for the detection of movement as well as high power 'RFID' tags) use 2445 - 2455 MHz affecting channels 6 to 11 inclusive (though the effect is most profound on channels 7 to 10). Emissions from microwave ovens (and most other ISM equipment) centre around 2450 MHz, and as such would also affect channels 6 to 11. So from this simple analysis, it would seem like channels 12 and 13 are the most likely to be clearest of interference.
However, by far the most likely source of interference is... other Wireless LAN users. So which channels are most commonly used by other WiFi bods? On a recent train journey from Manchester to London, I let my laptop and trusty 'Netstumbler' software scan the band for LANs, to see which channel was most commonly in use. The results are shown below:
The most commonly used channel was 11 - given the earlier analysis about the channels least likely to be interfered with, it is perhaps not surprising that this is the default channel on which most equipment operates when initially purchased. Equally, it seems that most people give no further thought to the channel number and leave it on 11. Channels 1 and 6 were also relatively well used - again not surprising given that 1, 6 and 11 are the most common interference free line-up.
So, after all this, which channel is most likely to present the lowest overall noise and interference? If there are unlikely to be any other WiFi users in the neighbourhood, I'd choose channel 12 or 13. If there is likely to be lots of other WiFi use nearby, channel 1 is the best bet.
Tuesday 20 June, 2006, 12:11 - Radio RandomnessHaving gone on about how to extend the range of a wireless LAN using a high gain antenna, the need suddenly arose for the range of my own WiFi connection to be extended so I though I would purchase a 9dBi antenna to see what happened. Being a hardcore engineer, I wanted to try and see whether this antenna really delivered the gain over the original 2dBi antenna which it promised.
Antennas and laptop in hand, I used a programme called 'Netstumbler' to record the signal to noise of the reception of my WiFi connection over about a 1 hour period, changing between the standard 2dBi and the higher-gain 9dBi antenna about half way through. The first graph (below) shows the received signal strength over the period. The orange line is the rolling average over a 2 minute period. Without any further analysis, it is immediately apparent that the signal strength received when the 9dBi antenna is installed is higher than with the 2dBi antenna, indicating that it did have some gain as promised.
A quick calculation of the average over the 2 periods showed that the average signal to noise with the 2dBi antenna was 28.8dB, whereas the average signal to noise with the 9dBi antenna was 34.6dB, an increase of 5.8dB - not quite the 7dB increase that should in theory materialise, but not bad nonetheless. For the statisticians amongst you, the standard deviation in both cases was remarkably close at 3.96 and 3.92 dB for the 2 and 9 dBi antennas respectively, indicating that the signal was equally stable (despite the obvious variations) in both instances.
A further analysis of the results (above) shows the distribution of signal strengths produced by the 2 antenns. In this case it is easy to see that the signal produced by the 9dBi antenna (in orange) is consistantly and significantly better than that of the 2dBi antenna (in dark blue).
Whilst the experiment was less than scientific, taken at face value, it does suggest that the signal produced by the 9dBi antenna is a worthwhile improvement over the signal produced by the standard 2dBi antenna as supplied with most WiFi routers. The claims of a '3 to 4 times' increase in the range of coverage have not been tested - maybe I'll do that one day soon.
Tuesday 6 June, 2006, 08:22 - Radio RandomnessA London Assembly report published yesterday (5 June 2006) into the events surrounding the terrorist attack on London on 7 July 2005 identified that the major problem suffered by the emergency services at the event was the lack of adequate radio communications. It identified that, other than the British Transport Police, none of the radio systems used by police, fire or ambulance services functioned on the London Underground. Further, it goes on to note that the different radio systems used by the different emergency services means that they are unable to effectively communicate from one service to another. It also states that the reliance on mobile phones by the emergency services (in particular the Ambulance service and for communication between senior officers) was effectively an accident waiting to happen.
What's perhaps surprising is that none of these problems were unknown. There have been plans afoot for over 10 years to install radio communications across the London Underground, including in tunnels. The project, known as 'Connect' was awarded to a consortium in 2000. The original plan was to install infrastructure across the underground to support both 'TETRA' based communications for the tube trains themselves as well as to support cellular communications. However technical difficulties have meant that the project has now been restricted to just providing TETRA for cab-to-control communications. The project is still in a 'beta' phase with no actual radios in service.
Since the emergency services first used radio, each of the services had its own radio system. In times of crisis, each service established a mobile headquarters, often in the back of a van, and as long as each services' HQ was parked near the other services, communications between the different services could be achieved by shouting from van to van, or by running hand-written paper notes between them. The events in London, however, occured in several different and diverse locations and with improvements in communication and information technology, there is less of a need for a remote HQ as everything can be dealt with from the regular HQ. As such, communication between the services is now worse than it might have been 10 or 15 years ago and in a situation such as the terrorist attacks on London last July, would have been virtually non-existent. There is light at the end of the tunnel, however, in the form of 'Airwave'. Airwave is a TETRA based radio service being rolled-out across the UK for use by all the emergency services. Initially intended just for the police, Airwave has now been adopted by the ambulance and fire services too. Once roll-out is complete, all three main emergency services will be using the same radio technology and the same radio system so inter-service communications should be much more straight forward.
It is a well understood fact that, in times of crisis, usage of mobile phones rockets. Everyone near or involved in the crisis will wish to contact their relatives to tell them they are OK, or to call the emergency services to try and garner assistance. Mobile networks are not designed to handle such high loads and the inevitable result is severe congestion and in some cases, network failure. With this in mind, the GSM specification includes a number of features to try and ensure that emergency service users can continue to communicate at such times. Access overload class (ACCOLC) is a UK designed system in which ACCOLC enabled mobiles are given priority access to a cellular network. Each ACCOLC mobile phone is assigned a priority from 1 to 15 (normal users are randomly assigned a priority in the range 1 to 10, whilst emergency users have priorities 11 to 15). The network selectively disables the lower priority mobiles to ensure that higher priority communications can continue. Thus priorities 1 to 5 may be disabled, randomly cutting off half of all civilian users and thus freeing the network for emergency communications. There is also a feature of GSM (version 2+) known as enhanced multi-level precedence and pre-emption (eMLPP) - this too allows priority access for certain users, however it is not commonly implemented on many commercial GSM networks being reserved for those networks designed to support railway communications (GSM-R).
So from the perspectives of the communication abilities of the UK emergency services, the attacks came just at the right (or is that wrong) time, during a period of transition and when sub-terranean communications were not at their best. It is to be hoped, therefore, that even if the atrocities had not taken place, the situation would be improving. With the publishing of the London Assembly report, let us pray that adequate communications are in place before anyone else decides to take a pop at the innocent people of our capital city (or any other part of the UK or the world for that matter).
Monday 24 April, 2006, 15:08 - Radio RandomnessThe practice of jamming radio transmissions has been around since radio was first used for military purposes when one side would attempt to inhibit the communications of the other side by transmitting on the same frequency at the same time. Indeed such military jamming is usually termed 'electronic counter-measures' or ECM and a whole industry has developed around it to try and produce radio systems that are resilient to jamming known as 'electronic counter-counter-measures' or ECCM. The picture on the right shows an example of a drone in use today for jamming - in this case one known as a Dragon. The payload of the missile is a high power transmitter which can then be sent to the area of interest and remotely activated. You can see the antenna which unfolds from the front of the device when it is deployed.
However jamming is not strictly a tool of the military. Jamming of broadcast radio stations has been a regular activity for many organisations and - and here's the disappointing bit - continues to take place even today. In the height of the cold war, countries such as Russia used to jam short-wave transmissions from countries such as the UK and USA which carried news or information which the Russians viewed as anti-communist propaganda. To try and compensate, broadcasters such as the BBC World Service, Radio Free Europe and the Voice of America used to transmit their programmes on several short-wave frequencies at the same time (usually 4 or more) in the vain hope that at least one of them would be receiveable in a given area clear of the jammers. The Cuban authorities also jammed Radio Marti, a US Government funded station aimed specifically at the country.
Although they haven't admitted it, China continue to jam Mandarin language short-wave transmissions from the same kinds of stations even today (take a look at this article for example). Tune to the BBC Chinese service (try 9580, 11945, 11980 or 13970 kHz or look at the most recent programme schedule N.B. this page is in Mandarin) and you're likely to hear either the familiar 'wobble' sound over the top of the transmission or 'distorted voices' - both symptoms of a jammed frequency. It's odd that such activities still take place when, with the advent of the internet, the kind of 'propaganda' that the Chinese are trying to guard against is freely available to anyone with a computer. Of course, not all the population of China have access to a computer but with internet cafe's many do (though the Chinese authorities do their best to stop people getting access to such information). There are also case of the Chinese jamming themselves!
Whilst such overt jamming of international stations by such political regimes is virtually taken for granted, there have been deliberate attempts to block transmissions from other stations by apparently respectable regimes. Read the story of Radio Caroline whereby the British Government decided to jam its transmissions to stop them being heard during election times. The British Government pulled another fast trick in 1990 when it licensed a station in London on 558 kHz, the frequency used at the time by Radio Caroline. The choice of frequency was clearly chosen to stop Radio Caroline from being heard in London (though the plan vaguely back-fired as the London station was forced to temporarily simulcast on 990 kHz as the 558 signal was so badly interfered with by Radio Caroline).
More recently there has been a story suggesting that the US Government was deliberately jamming radio communications during the recent disaster in New Orleans. Whilst the US Government clearly could have done more for the residents of this flood-stricken town, it belies belief that they would actively try to scupper rescue attempts.
What is clear is that despite significant developments in radio technologies and in international relations, radio jamming transmitters seem to be as active today as they have always been, mucking up military communications and supressing freedom of speech. Let's hope that one day soon, neither of these uses will continue to be necessary.