For 3G networks to realise their full technical
potential, smaller base stations could provide the key to greater coverage, as Dr Alan Carr explains.
3G network deployments in Europe are at last gathering pace and there will be several commercial launches this year. During the early stages of network rollout, the focus is on achieving good coverage and this is done using large, high power macrocell base stations. However, as networks become more mature and as subscriber numbers grow, the emphasis with established 3G operators is already starting to shift to filling in gaps in coverage, supporting hotspots in capacity and providing service inside buildings.
The same path was followed with 2G systems — such as GSM — and the solution there was to deploy smaller more lightweight pico or microcell base stations in addition to traditional macrocell equipment. As 3G networks mature, a similar process will take place. But this time, smaller base stations will be much more important and will eventually make up a large part of 3G networks. Here, we examine the reasons for this and look at the form such pico or microcell base stations might take.
Changing traffic patterns
To understand the reason why small base stations will be so important in 3G networks, we first need to consider changes in the type of traffic in a 3G network. With 2G, most traffic is voice calls and, of course, voice will remain an important element in 3G networks. But the real goal with 3G is the transition to a true mobile communication system, rather than just a mobile telephone network. As a result, a growing part of the traffic will be made up of data — much of it at higher rates than is possible in 2G networks. In urban areas, where users tend to cluster together in offices, stations, shopping malls and conference centres, this will quickly lead to a situation where the overall density of traffic is substantially higher than in 2G networks. The result is that many more base stations will be needed. But this will be uneconomic and, in many cases, unacceptable due to public opposition to the introduction of new sites. The answer is to introduce a layer of smaller base stations in crowded urban areas to carry much of the traffic.
The problem indoors
Another big driver towards smaller base stations in 3G is the need to get good indoor coverage. The problem starts from the fact that 3G operates at around 2GHz, where losses penetrating buildings are higher than with 2G systems. However, this alone should not to be a major issue when it is considered that GSM has been made to work perfectly satisfactorily in the nearby band at 1800MHz. The more significant factor is that 3G users inside buildings will typically want to make use of higher speed services and that such high speed services normally require a higher quality radio link than would be needed for voice alone. And so, although it may be possible to achieve reasonable voice coverage inside buildings, it will be much more difficult for the more advanced data services that are those expected to be needed indoors. Again smaller base stations can address this requirement.
Technical factors
There are also some more detailed technical reasons why smaller base stations will be more important in 3G than 2G networks. The first relates to the orthogonality introduced to separate users in the downlink. CDMA systems have been carefully designed to maximise this orthogonality since it has an important effect on boosting capacity in the downlink. However, in a macrocell it is common for a substantial part of the orthogonality to be lost because of the complex multipath channels that exist between the base station to the mobile terminal. This results in a corresponding loss in capacity. The effect is much less in microcells or picocells because of the shorter and less complex path between base station and terminal.
The second technical factor in favour of smaller cells is intercell interference. Adjacent cells in CDMA systems normally operate on the same carrier frequency and interference from cells bordering the home cell appear as noise when trying to detect the wanted signal. In turn, this noise-like effect depresses system capacity. Pico and microcells are normally located below rooftop level and there is therefore less scope for interference between cells. On average, capacity loss due to intercell interference is also therefore lower.
Thus, perhaps surprisingly, smaller cells are more efficient than their conventional larger equivalents and are therefore able to support more user traffic.
The cost factor
The final driver to small base stations is simply cost. Macrocell base stations are relatively expensive items of equipment designed for so-called ‘carrier class’ operation. This typically involves building in redundancy to meet reliability targets as well as incorporating expensive high power amplifiers. Special environmentally protected cabins are also often needed. By contrast, the format of pico and microcell base stations is such that a design approach can be used closer to handsets rather than conventional base stations. Reliability requirements are typically reduced, output powers are much lower and special environmental protection is not needed. Maintenance costs for pico and microcell base stations can also be lower.
In terms of operational costs, a weak area for smaller base stations is backhaul, because of the larger number of cell site involved than with macrocells. However, technology is evolving to provide solutions here. For picocell base stations, 3G standards are developing so that IP networks can be used for backhaul. So, for an indoor application such as in an office, an attractive backhaul solution is to use the existing Ethernet infrastructure installed within the building. For microcell base stations that are more likely to be located outdoors, one emerging solution is the 802.16 WiMAX standard which could provide an effective radio backhaul option. There are also other radio alternatives and, in terms of fixed networks, xDSL is potentially a lower cost alternative to conventional leased lines.
Typical deployments
So what will pico and microcell base station deployments look like and how will they connect to the rest of the network? We will look first at picocell base stations. These will typically be located indoors and have an output power of some 250mW. Typically, a base station will be needed for each floor of a building or for each major area as shown. (Diagram 1, above). It will not be possible to achieve good indoor coverage in 3G networks using traditional macrocell sites. A solution is to use indoor picocell base stations within the building connected to the existing IP network infrastructure for backhaul.
Moving on to microcell base stations, these are typically located outdoors, but below rooftop level. The goal is to infill areas such as street canyons and to deal with urban hotspots. It is also possible to achieve some in-building coverage from outdoors. The diagram above shows a possible scenario.
Microcell base stations will typically be located outdoors and will provide coverage infill, deal with capacity hot-spots and, in some cases, provide indoor coverage from outside.
Picocell and microcell
Finally, we consider the likely form of pico and microcell basestations. The diagram (Diagram 2, below) shows a Reference Design available from PA for both pico and microcell basestation products.
PA’s Reference Designs for 3G picocell and microcell basestations, including a scale diagram. The microcell version is shown with a cut-away to illustrate the internal design. They can be used for WCDMA (including HSDPA), cdma2000 or other 3G products
The baseband section in these designs is implemented entirely in software running on the TigerSHARC DSP from Analog Devices.
This enables upgrades in functionality to be loaded entirely over the network. Such an approach is vital since it would be very expensive to carry out a hardware upgrade when thousands of base stations are located in the field. The all-software approach also has advantages in terms of cost, scalability and the ability to improve performance through advanced processing techniques.
The radio design is a flexible one, able to support multiple 3G standards and with a combination of high sensitivity and efficient power amplification. In mechanical terms, both units are entirely convection cooled and have built-in antennas — the microcell design has an option for external antennas.
One interesting factor in the implementation of pico and microcell base stations is that it is not possible to simply scale down techniques that are well established in macrocell base stations.
This applies to the general design approach which, as stated above, is more like a terminal than a conventional base station. It is also necessary to rethink the way the core processing operations in the base station are carried out. This is because the channel conditions experienced are very different to those in macrocell base stations, particularly in areas such as propagation loss, delay spread, Doppler effects and so on. In turn, these can require that quite different processing methods be employed.
Much stronger case
Microcells are widely used in 2G networks and typically make up some 5-10 per cent of the total number of base stations. The case for picocell and microcell base stations in 3G is much stronger and we expect this proportion to be higher. It is difficult to provide accurate forecasts at this time, but eventually a proportion of at least 20 per cent seems possible. This will require good quality base station products fully customised to the changed conditions in which pico and microcell base stations will have to operate. PA has produced Reference Designs for both picocell and microcell base stations to assist with this process.
3G network deployments in Europe are at last gathering pace and there will be several commercial launches this year. During the early stages of network rollout, the focus is on achieving good coverage and this is done using large, high power macrocell base stations. However, as networks become more mature and as subscriber numbers grow, the emphasis with established 3G operators is already starting to shift to filling in gaps in coverage, supporting hotspots in capacity and providing service inside buildings.
The final driver to small base stations is simply cost. Macrocell base stations are relatively expensive items of equipment designed for so-called ‘carrier class’ operation. This typically involves building in redundancy to meet reliability targets as well as incorporating expensive high power amplifiers. Special environmentally protected cabins are also often needed. By contrast, the format of pico and microcell base stations is such that a design approach can be used closer to handsets rather than conventional base stations. Reliability requirements are typically reduced, output powers are much lower and special environmental protection is not needed. Maintenance costs for pico and microcell base stations can also be lower.
In terms of operational costs, a weak area for smaller base stations is backhaul, because of the larger number of cell site involved than with macrocells. However, technology is evolving to provide solutions here. For picocell base stations, 3G standards are developing so that IP networks can be used for backhaul. So, for an indoor application such as in an office, an attractive backhaul solution is to use the existing Ethernet infrastructure installed within the building. For microcell base stations that are more likely to be located outdoors, one emerging solution is the 802.16 WiMAX standard which could provide an effective radio backhaul option. There are also other radio alternatives and, in terms of fixed networks, xDSL is potentially a lower cost alternative to conventional leased lines.
Typical deployments
So what will pico and microcell base station deployments look like and how will they connect to the rest of the network? We will look first at picocell base stations. These will typically be located indoors and have an output power of some 250mW. Typically, a base station will be needed for each floor of a building or for each major area as shown. (Diagram 1, above). It will not be possible to achieve good indoor coverage in 3G networks using traditional macrocell sites. A solution is to use indoor picocell base stations within the building connected to the existing IP network infrastructure for backhaul.
Moving on to microcell base stations, these are typically located outdoors, but below rooftop level. The goal is to infill areas such as street canyons and to deal with urban hotspots. It is also possible to achieve some in-building coverage from outdoors. The diagram above shows a possible scenario.
Microcell base stations will typically be located outdoors and will provide coverage infill, deal with capacity hot-spots and, in some cases, provide indoor coverage from outside.
To understand the reason why small base stations will be so important in 3G networks, we first need to consider changes in the type of traffic in a 3G network. With 2G, most traffic is voice calls and, of course, voice will remain an important element in 3G networks. But the real goal with 3G is the transition to a true mobile communication system, rather than just a mobile telephone network. As a result, a growing part of the traffic will be made up of data – much of it at higher rates than is possible in 2G networks. In urban areas, where users tend to cluster together in offices, stations, shopping malls and conference centres, this will quickly lead to a situation where the overall density of traffic is substantially higher than in 2G networks. The result is that many more base stations will be needed. But this will be uneconomic and, in many cases, unacceptable due to public opposition to the introduction of new sites. The answer is to introduce a layer of smaller base stations in crowded urban areas to carry much of the traffic.
The problem indoors
Another big driver towards smaller base stations in 3G is the need to get good indoor coverage. The problem starts from the fact that 3G operates at around 2GHz, where losses penetrating buildings are higher than with 2G systems. However, this alone should not to be a major issue when it is considered that GSM has been made to work perfectly satisfactorily in the nearby band at 1800MHz. The more significant factor is that 3G users inside buildings will typically want to make use of higher speed services and that such high speed services normally require a higher quality radio link than would be needed for voice alone. And so, although it may be possible to achieve reasonable voice coverage inside buildings, it will be much more difficult for the more advanced data services that are those expected to be needed indoors. Again smaller base stations can address this requirement.
Technical factors