Infrastructure analysis and solutions for 800MHz network deployment

ICT KNOWLEDGE TRANSFER NETWORK Infrastructure analysis and solutions for 800MHz network deployment Page 1

About the ICT Knowledge Transfer Network (ICT KTN) The ICT KTN is an organisation established by an industry-led group of leading players, with funding from the Technology Strategy Board. We seek to bring competitive advantage to the UK by promoting collaboration and knowledge sharing between the users and providers of ICT, and helping to drive innovation in the sector. If you are involved in ICT, and have not yet registered as member of the ICT KTN please visit our web page and register as a member. ICT KTN Russell Square House, 10 12 Russell Square, London WC1B 5EE Email: info@ictktn.org.uk Website: www.ictktn.org.uk Telephone: +44 (0) 20 7331 2056 Contact point for positioning paper: Stuart Revell, ICT KTN Wireless Technology & Spectrum working group Email stuart.revell@ictktn.org.uk John Burns Aegis Systems Ltd Email John.Burns@aegis-systems.co.uk Ian Vance Amazing Communications Ltd Email ian@vance.myzen.co.uk David Barker Quintel Solutions Email david.barker@quintelsolutions.com The opinions and views expressed within this positioning paper have been reviewed by members of the ICT Knowledge Transfer Network Wireless Technology & Spectrum working group. The views and opinions do not necessarily reflect those of the individual members of the ICT KTN or the Working Group or the organisations that the members represent. Page 2

1. Contents 1. Contents...3 2 Table of figures and tables...3 3 Abbreviations and definitions...4 4 Introduction...5 5 Network analysis...6 5.1 Indoor coverage to a USB dongle or mobile handset...8 5.2 Indoor coverage to a desktop modem or router...9 5.3 Outdoor coverage to an external, directional antenna...10 5.4 Coverage improvement example using enhanced UE antennas...11 6 Capacity implications...12 7 Coverage and capacity improvements through adoption of higher gain antennas...14 7.1 Opportunities to improve user equipment (UE) performance...14 7.2 Opportunities to improve base station performance...16 8 Conclusions...22 2 Table of figures and tables Figure 1: Three link budget scenarios used for modelling...7 Figure 2: Additional base stations required for coverage expansion with USB dongle at 800 MHz...8 Figure 3: Additional base stations required for coverage expansion with desktop modem or signal booster at 800 MHz...9 Figure 4: Additional base stations required for coverage expansion with external antenna at 800 MHz...10 Figure 5: Rural cell site example - coverage comparison demonstrating different UE Antenna options...11 Figure 6 UK population density (incremental) vs population coverage...12 Figure 7: Population per 5 km radius cell (incremental) vs population coverage...13 Figure 8: Available bit rate per user assumes 10%of population active mobile broadband users, half of these supported by 800 MHz spectrum...13 Figure 9: Cell Splitting in Azimuth to increase coverage/capacity...17 Figure 10: Using 4-way Uplink Receive Diversity to increase Uplink range (increased Rx aperture and 4-branch Rx de-correlation)...18 Figure 11: Using 4-way Uplink Receive Diversity to increase Uplink range (increased Rx aperture only)...19 Figure 12: Extract from OFDM for LTE, illustrating the link budget differences between 900MHz and 2600MHz....20 Figure 13: Maximising wireless opportunities through Government & Industry collaboration...23 Table 1: Estimated cell size for indoor coverage to USB dongle in 800 MHz band... 8 Table 2: Estimated cell size for indoor coverage to desktop modem or signal booster in 800 MHz band... 9 Table 3: Estimated cell size for indoor coverage to external antenna in 800 MHz band... 10 Page 3

3 Abbreviations and definitions Abbreviation 3G 3GPP 4G C/I db dbi DSL DTT EiRP GSM ITU km LTE Mbps MiFi MIMO QAM QPSK RF Rx sq km UE UMA UMTS USB VoIP WiBE WiFi WLAN Definition 3 rd Generation Mobile 3 rd Generation Partnership Project 4 th Generation Mobile Carrier to Interference (Ratio) Decibels Decibels isotropic (gain of antenna) Digital Subscriber Line Digital Terrestrial Television Equivalent isotropically radiated power Global System for Mobile communications International Telecommunication Union Kilometre 3GPP Long Term Evolution Megabits per second Mobile Wi-Fi hotspot Multiple-Input and Multiple-Output Quadrature amplitude modulation Quadrature phase shift keying Radio Frequency Receive or Receiver Square Kilometre User equipment Unlicensed Mobile Access Universal Mobile Telecommunications System Universal Serial Bus Voice over Internet Protocol Wireless Broadband Extender Wireless Local Area Network technology (WLAN) Wireless Local Area Network technology Page 4

4 Introduction 3G mobile networks currently deployed in the UK claim to cover up to 97% of the population. Such claims typically relate to outdoor coverage at the lowest available bit rate, which with current technology is considerably less than 1 Mbps. This coverage is based on existing 3G spectrum at 2 GHz and requires the deployment of approximately 12,000 base station sites. The availability of good indoor coverage or higher bit rates (corresponding to true broadband rates of 2 Mbps or more) is far more limited. For example, indoor coverage implies typically a further 10 15 db signal attenuation, which will reduce the cell coverage area by over 70%. Doubling the bit rate (by deploying 16QAM modulation instead of QPSK) requires a 6 db signal enhancement, implying a further halving of the cell coverage area. Deployment of 800 MHz or 900 MHz alongside the core band (3G 2GHz, assumed to be 12,000 sites) should yield a big improvement, bringing indoor coverage close to the current levels of outdoor coverage and providing higher bit rates over much of the cell area. Coverage at the minimum bit rate will also be extended further beyond the current 97%. However, the limited availability of spectrum in these two bands means that enduser performance and network capacity may still fall short of future requirements, especially if the mobile network provides the main or only broadband connection in some areas. Maximising coverage and capacity may therefore require a different approach to network and terminal deployment, for example using directional signal boosters or external aerials to enhance the wanted signal and reduce interference from neighbouring cells. In its recent consultation 1 on the 800 MHz band, Ofcom has proposed to apply a coverage obligation of 95% of the UK population on one of the licences, with a minimum specified bit rate of 2 Mbps. However, for mobile networks to provide a viable platform for fixed broadband delivery where alternatives do not exist will require coverage to extend considerably beyond this level, into areas beyond the reach of cable and DSL platforms. The number of additional sites required to achieve true broadband capability (i.e. 2 Mbps or higher) could be significant and depends very much on the architecture deployed. 1 Ofcom consultation published 22 nd March 2011, Assessment of future mobile competition and proposals for the award of 800 MHz and 2.6 GHz spectrum and related issues http://stakeholders.ofcom.org.uk/consultations/combined-award/ Page 5

In this paper we estimate the number of additional base stations that might be required to expand broadband mobile coverage, using 800MHz spectrum, from 90% to 99% of the UK population, assuming that coverage is extended progressively into increasingly sparsely populated areas. The analysis considers various network configurations and the potential user bit rates that might be available. The 90% to 99% model range has been selected to demonstrate the increasing challenge to meet coverage to the last 10% of the population and also to provide a platform to stimulate further discussion and opportunities for innovative solutions to address this challenge. 5 Network analysis In the following analysis we have assumed that coverage is provided using LTE technology in the 800 MHz frequency band, using a 2 x 10 MHz carrier. Note that the bit rates referred to are peak rates in the downlink and do not take account of contention; we have also not included MIMO deployment in our analysis (though this is referred to in a later section). Three link budget scenarios are considered, namely: i) Indoor coverage to a USB dongle ii) Indoor coverage to a desktop modem or signal booster iii) Outdoor coverage to an external, directional antenna (10 dbi) oriented towards the mobile base station. The starting point for the model is 90% population coverage, the rationale for this starting point is that existing coverage is already assumed to be achieving >90% (outdoor coverage) and 800MHz spectrum using these sites will be able to deliver incremental capacity, population coverage and/or coverage depth (improved indoor coverage). The upper limit of 99% has been used as an upper limit of what is reasonably achievable. To estimate the number of additional sites required to achieve the specified population coverage, we have assumed that the coverage area of rural base station sites will be constant but that coverage will be progressively extended to more sparsely populated areas, i.e. each incremental base station will serve a slightly smaller population than the previous site. Whilst this assumption is reasonable for small cell sizes corresponding to indoor coverage at high bit rates, for less demanding link budgets the cell size is more likely to be limited by local terrain features. We have therefore applied a lower limit Page 6

based on existing mobile networks for the indoor coverage scenarios and on the digital TV transmission network for the outdoor aerial scenario 2. The figure below demonstrates the three scenarios used in the modelling. It must be noted that the parameters used in the model for options ii) and iii) have used a conservative gain improvement compared to the direct dongle scenario. Section 7.1 illustrates some solutions for indoor smart antennas that provide 15 to 20dB improvement compared to the 10dB we have used in the model. Figure 1: Three link budget scenarios used for modelling 2 Data for existing mobile networks is based on graphical data illustrating the number of sites as a function of population coverage published Annex 7 of the 22 nd March Ofcom consultation. For the outdoor aerial case a total of 1,100 sites is assumed to be required to extend coverage from 90% to 98.5%, reflecting the projected coverage target for public service DTT in the UK. Page 7

Addditional sites required 5.1 Indoor coverage to a USB dongle or mobile handset In this scenario it is assumed that the terminal device has a zero gain antenna and that there is an average 10 db building penetration loss. The rural cell size (radius) as a function of available peak bit rate is estimated to be as follows (based on a typical LTE link budget and COST-Hata propagation model): Table 1: Estimated cell size for indoor coverage to USB dongle in 800 MHz band Layer 1 Peak Bit Modulation and Maximum path Rate (downlink) coding scheme loss Cell size (radius) 7.2 Mbps QPSK 1/2 117 db 5.5 km 14.4 Mbps 16QAM 1/2 110 db 3.5 km 21.6 Mpbs 16QAM 3/4 106 db 2.7 km 32.4 Mbps 64QAM 3/4 100 db 1.8 km The number of additional base stations that would be required to extend coverage beyond 90% under this scenario is estimated to be as follows: Figure 2: Additional base stations required for coverage expansion with USB dongle at 800 MHz 50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 0 90% 92% 94% 96% 98% 100% Population Coverage 32.4 Mbps 21.6 Mbps 14.4 Mbps 7.2 Mbps Note that for the lowest bit rate an additional 5,000 base stations would be required, increasing to 20,000 for 21.6 Mbps and over 40,000 for 31.2 Mbps. Page 8

Additional sites required 5.2 Indoor coverage to a desktop modem or router In this scenario it is assumed that the terminal device has a 5 dbi gain and is located by a window, reducing the typical building penetration loss to 5 db i.e. an overall 10 db improvement in the link budget compared to the previous scenario. Table 2: Estimated cell size for indoor coverage to desktop modem or signal booster in 800 MHz band Layer 1 Peak Bit Modulation and Maximum path Rate (downlink) coding scheme loss Cell size 7.2 Mbps QPSK 1/2 127 db 10.6 km 14.4 Mbps 16QAM 1/2 120 db 6.7 km 21.6 Mpbs 16QAM 3/4 116 db 5.2km 32.4 Mbps 64QAM 3/4 110 db 3.5 km The number of additional base stations that would be required to extend coverage beyond 90% is estimated to be as follows: Figure 3: Additional base stations required for coverage expansion with desktop modem or signal booster at 800 MHz 14,000 12,000 10,000 8,000 6,000 4,000 2,000 32.4 Mbps 21.6 Mbps 14.4 Mbps 7.2 Mbps 0 90% 92% 94% 96% 98% 100% Population Coverage In this case, coverage at 21.6 Mbps can be achieved with a similar number of base stations to the provision of 7.2 Mbps with a conventional mobile terminal. The number of base stations required to deliver 32.4 Mbps is reduced by approximately 75%. Page 9

Additional sites required 5.3 Outdoor coverage to an external, directional antenna In this scenario it is assumed that an external directional antenna with a 10 dbi gain and 5m elevation is deployed and is oriented towards the base station. This will eliminate the building penetration loss and will provide a significant reduction in clutter loss due to the elevated antenna. A small feeder loss will be incurred (estimated up to 2 db). An overall improvement in the link budget of 8 db relative to scenario 2 and 18 db relative to scenario 1. We anticipate that in many cases a similar improvement may be attained by deployment of an indoor router device equipped with a smart antenna array and located in a favourable reception position (e.g. attic or by upstairs window facing nearest base station) Table 3: Estimated cell size for indoor coverage to external antenna in 800 MHz band Layer 1 Peak Bit Modulation and Maximum path Cell size Rate (downlink) coding scheme loss 7.2 Mbps QPSK 1/2 135 db 18.0 km 14.4 Mbps 16QAM 1/2 128 db 11.4 km 21.6 Mpbs 16QAM 3/4 124 db 8.7 km 32.4 Mbps 64QAM 3/4 118 db 5.9 km The number of additional base stations that would be required to extend coverage beyond 90% is estimated to be as follows: Figure 4: Additional base stations required for coverage expansion with external antenna at 800 MHz 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0 90% 92% 94% 96% 98% 100% Population Coverage 32.4 Mbps 21.6 Mbps 14.4 Mbps In this case there is a dramatic reduction in the number of sites required, and it is possible to provide the highest bit rate with fewer sites than are required for the lowest bit rate in the conventional mobile scenario. Page 10

5.4 Coverage improvement example using enhanced UE antennas The figure below illustrates the potential coverage improvement that might be realised from a single randomly chosen rural cell site in Cumbria using the terminal antenna enhancements discussed above. The projections are based on the ITU P.1812 propagation model and assume a terminal height of 1.5 metres and 10 db building penetration loss for the indoor scenarios. Base station data (location, height and radiated power) is sourced from the Ofcom Sitefinder web site 3. 2 GHz indoor dongle 800 MHz indoor dongle 800 MHz indoor smart antenna 800 MHz external antenna (5 m) 7.2 Mbps 14.4 Mbps 32.4Mbps Figure 5: Rural cell site example - coverage comparison demonstrating different UE Antenna options 3 Ofcom 'Sitefinder' Mobile Phone Base Station Database http://www.sitefinder.ofcom.org.uk/ Page 11

6 Capacity implications The capacity of a mobile network is generally much lower than is implied by the peak bit rates, due to the distribution of users throughout the cell, impact of inter-cell interference and so on. According to the Third Generation Partnership Project (3GPP) which is developing the LTE Release 10 standards, the target average spectrum efficiency is 2 bps/hz, implying a capacity per sector per carrier of 20 Mbps for 2 x 10 MHz. For a trisectored cell the total capacity would therefore be 60 Mbps. Deployment of MIMO, which may be more practical for a desktop modem or external antenna due to the larger terminal size, could potentially increase throughput as much as four fold, i.e. to as much as 240 Mbps per base station. A lot more capacity would be available if additional (higher) frequency bands are deployed, but this would be limited to areas closer to the base station. Nevertheless, it is reasonable to assume that at least half of the capacity in a rural cell could be provided by higher frequency bands for example by ensuring that base stations are located close to towns and villages so that much of the population served is within range of the higher frequency signal. We assume that in practice the majority of sites will deploy multiple frequency bands to maximise coverage and capacity from each site (for the same reason we have also assumed that all macro sites will be sectorised an average value of 3 sectors is assumed). The average population density as a function of the percentage of population covered in the UK is shown below (based on census data): Figure 6 UK population density (incremental) vs population coverage Assuming an average cell radius of 5 km, which our analysis indicates would provide LTE coverage at the minimum bit rate to indoor mobile terminals and coverage at the highest bit rate to outdoor mounted antenna, the average coverage area per cell would Page 12

Available bit rate per user (Mps) Population seved by each additonal cell be 49 sq km 4. The total resident population in each additional cell as a function of population density would be as follows: Figure 7: Population per 5 km radius cell (incremental) vs population coverage 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 90% 91% 92% 93% 94% 95% 96% 97% 98% 99% Population coverage Assuming a take-up rate of mobile broadband of 10 % of the total and a per-cell capacity of 60 Mbps without MIMO and 240 Mbps with MIMO, the bit rate per user that could be supported in each additional cell is shown below. Note that the bit rate per user shown relates to an individual cell that extends coverage to the corresponding value as we have assumed that coverage would be progressively extended into increasingly sparsely populated areas and the cell size is limited by the radio link budget, each additional cell will serve a smaller population and hence the available bit rate per user within that cell will be correspondingly greater. : Figure 8: Available bit rate per user in most sparsely populated cell assumes 10%of population active mobile broadband users, half of these supported by 800 MHz spectrum 10 9 8 7 6 5 4 3 2 1 0 90% 92% 94% 96% 98% Population coverage without MIMO with MIMO 4 UMTS network planning, optimization, and inter-operation with GSM By Moe Rahnema (Wiley-IEEE, 2008) ISBN 978-0-470-82301-9 (HB). Table 5.6 Cell Size calculation = 1.95r² Page 13

Deploying an entirely fixed network could potentially increase the capacity further by ensuring every connection uses the highest available modulation and coding scheme in theory this could raise the total capacity to more than 150 Mbps per carrier, or 450 Mbps per site, yielding bit rates up to 7.5 times higher than those shown above, however this would mean sacrificing the provision of mobile connectivity over the network and would probably require an excessively complex antenna configuration. In practice, the take-up of high speed mobile broadband services is likely to be higher in more remote areas where fixed broadband is less available i.e. the assumption of a constant 10% take up regardless of the coverage level may not be valid. As these remote areas tend to have the lowest population densities and hence fewer users per cell coverage area this may make the mobile network solution more attractive for serving these areas. 7 Coverage and capacity improvements through adoption of higher gain antennas The analysis conducted in this paper has shown that significant savings can be made by improving the radio link budget through using different network topologies. One of the few improvements that can be made to the link budget of a radio system without incurring extra power consumption, or without building many extra base stations is to improve the antennas, potentially this could be at both ends of the link. 7.1 Opportunities to improve user equipment (UE) performance Improving performance at the UE is a significant opportunity since many UE antennas are sub-optimal in several ways. There has been much recent discussion about the performance of Smartphone handsets and this situation is likely to deteriorate further with the addition of new bands that are being made available. Handset manufacturers struggle with the number of different frequency ranges to be covered in one unit and larger handsets are unlikely to find favour with users. UE antennas are typically either integrated into physically small devices (handsets, dongles) or embedded in a laptop, palmtop etc. It is noted that a large percentage of calls and data connections are made from within premises and that users are not truly mobile in that environment. In fact with the preponderance of data services most of the traffic is non-voice data communications to a range of devices such as smartphones, i-pad-like devices, Kindles etc. This range of devices can be better served by an intermediate equipment such as a 3G (or 4G) - Router which can be located at a favourable location to receive the cellular signal from Page 14

the nearest base station and translate it to WiFi, a low power 3G/4G femtocell or even Ethernet cable. Voice calls can then be handled either directly via the mobile network (as the narrow bandwidth required results in substantially better coverage or via the router using VoIP or UMA). The cellular transceiver (router) is thus decoupled from the user which has the following benefits: Elimination of antenna performance degradation arising from the terminal housing, proximity to the user and other physical obstructions Ability to incorporate a larger antenna as size is typically no longer at such a premium Ability to deploy the radio in a better immediate location (e.g. window-ledge or higher position etc) This type of enhanced router will typically have a performance at least 8 to 10dB better than a handset or dongle and thus will give significant improvements in coverage, effective cell size and/or data rate in rate adaptive systems. Typical examples are MiFi type devices and regular routers with built-in 3G or 4G capability However, with slightly larger equipment, Smart Antenna techniques can be used. Here a number of individual antennas are assembled into an array and some extra intelligence is added so that the antenna can be electronically pointed at the best signal. This gives several additional benefits: More gain arising from the directionality giving an overall improvement of 15 to 20dB over a standard dongle, which would approximate to the fixed external antenna model used in scenario 3, section 5.3 Less interference to and from other visible cell-sites resulting in a lower noise floor both at the UE involved and at these adjacent cell site receivers (especially important to users at the cell edge in all the cells) (If available and implemented) the ability of the network management function to balance traffic better by forcing use of a particular cell Such devices are available now, e.g. the Deltenna WiBE 5, and may be expected to become ubiquitous in the future The benefits of such fully featured smart antenna systems dramatically improve the system performance in both range and speed and have a huge effect on the number of 5 Deltenna WiBE information: http://www.deltenna.com/products/wibe.aspx Page 15

cells needed to cover a given area as shown in Section 5 They also fill in black spots wherever they may be in the coverage not just at remote sites and rural areas. In the same way future standards incorporate MIMO techniques. Here multiple antennas are deployed so that independent radio channels are established between the base station and the UE resulting in better system performance particularly higher data rates within a given channel width. MIMO systems require good antenna engineering to establish the independence of the transmissions. Combining this with smart systems mentioned above gives the best of both worlds at some cost in complexity. The effect of deploying this type of system is dramatic and brings the possibility of cellular networks being able to play an important part in delivery of Rural Broadband very much closer. 7.2 Opportunities to improve base station performance The communications link budget can be improved via engineering at the base station end. This can include a variety of base station and antenna configuration techniques and/or exploitation of emerging technologies. Increasing Base Station power can allow greater downlink range and/or capacity to be achieved, but of course range will be limited to the Uplink link budget. The use of higher power base stations should not be used as a panacea; increasing base station power offers diminishing returns, as any increase in power radiated also means increased inter-cell interference. For rural areas, especially where base stations have reasonable isolation from one another, or where terrain provides natural isolation between cell sites, then additional base station power can indeed be useful in particular to enhance Downlink capacity or to increase downlink throughput rates at cell edges. Use of higher power base stations is becoming attractive, especially as data demand in general is highly asymmetric, and Power Amplifier costs are dropping. Such higher power use has been observed in the US following a relaxation in the maximum EiRP limits. We understand that the UK (Ofcom) still maintain maximum EiRP limits, and feel this should be re-considered in particular for the interests of maximising rural services. Cell-Splitting is used as a means to increase capacity in a network using any finite spectrum bandwidth, and includes adding extra cell sites and/or use of higher-order Sectorisation at existing cell sites. Higher-order Sectorisation can be used to enhance both coverage and capacity through the use of higher gain antennas. A typical 3-sector base station can become a 6-sector base station using antennas which have narrower horizontal/azimuth beamwidth; the antennas often being called bi-sector antennas. The drawback here is that additional base station equipment needs to be added, and often only justified for capacity reasons rather than coverage, plus the antennas are wider. However, if 360 degrees coverage around the site is not required, then the use of simple narrower Azimuth beamwidth antennas can be useful for enhanced range Page 16

without the need for additional equipment, e.g. coverage up and down a main road using two highly directional antennas, which might be typical of rural coverage requirements. Using higher gain antennas in this way will require a larger antenna (in width) and will increase the EiRP (as commented upon in the previous paragraph). Figure 9: Cell Splitting in Azimuth to increase coverage/capacity It should be noted that maintaining 3-sectors and increasing antenna gain per sector by narrowing the vertical or elevation pattern beamwidth of the base station antenna is not a practical option. Antenna gain can be increased by increasing the length of the antenna array, but if the vertical radiation pattern becomes too narrow then coverage close to the site can be compromised (due to users falling into elevation pattern nulls), especially if tall sites are being used. Practically, and due to this reason, we rarely see antennas for the 1800MHz and 2100MHz bands over 2m in length (and more typically 1.3m). Elevation pattern null-filling techniques can be used in antennas to mitigate coverage holes close to sites using such very high gain antennas, but null-filling is difficult to achieve over wide bandwidth which is a common requirement of antennas today. Higher order receive (Uplink) diversity offers another coverage enhancement technique. Conventional base station configurations use 2-way Receive diversity for the Uplink channel, normally delivered by the two orthogonal polarisations of a cross-polar antenna array. Uplink signals received on each polarisation are generally de-correlated and hence fade somewhat independently; diversity combining in the receiver leads to an effective Uplink gain (against had no diversity been used). With higher-order diversity, such as 4-way diversity, an additional cross-polarised antenna is used within the same sector and connected to the Receive (Uplink) processing chain in the base station. Diversity combining is performed at baseband and processed independently (multiple and concurrent) for each user being received in Uplink. The use of a second antenna increases the effective aperture, hence provides Uplink gain, and provides additional fading de-correlation branches if the two cross-polar antennas are separated sufficiently. Uplink gains of around 4dB over 2-way diversity are typical for rural environments. The drawback with this technique is that extra antenna positions are Page 17

needed per sector, but no additional base station equipment is required, apart from providing extra ports and additional baseband processing. Figure 10: Using 4-way Uplink Receive Diversity to increase Uplink range (increased Rx aperture and 4-branch Rx de-correlation) We are observing a number of operators with high-band spectrum in 2600MHz band considering, and some in fact deploying two cross-polar antennas, as stated above, but instead of using two spatially separated cross-polar antennas, using them side-by-side within the same radome. This is still a valid 4-way Rx diversity configuration; it won t provide 4x de-correlated fading signals but will provide aperture gain for the Uplink. In essence, this will provide up to 3dB gain for the Uplink without the need for extra base station sector equipment (as would be in the 6-sector cell-splitting case). As the two cross-polar antennas are side-by-side (and in fact less than a l apart) then the Diversity combining can be visualised as adaptive Beam-forming in the Azimuth plane plus usual 2-branch polarisation diversity, for the Uplink. Operators are employing such techniques at 2600MHz band in particular as a means to better match/correlate coverage from 2600MHz with existing 2100MHz network site topology grid and coverage. Page 18

Figure 11: Using 4-way Uplink Receive Diversity to increase Uplink range (increased Rx aperture only) LTE Beam-forming Techniques: Operators are exploiting such close coupled dualcross-polar antenna techniques for benefiting the Uplink now, but 3GPP LTE also promises that the Downlink will be able to exploit this configuration too. 3GPP LTE Release 9 introduces Beam-forming for the Downlink traffic channel (i.e. on a per user basis), which in turn increases downlink data rates and/or cell sector capacity. This is called Dual Layer Beam-forming and in fact is an alternative base station configuration using 4x Power Amplifiers to 4xn Spatial Multiplexing MIMO, or 4-way Transmit Diversity in LTE, without having to resort to using additional antenna positions per sector. As for the side-by-side antenna array with higher order diversity in Uplink, duallayer Beam-forming would promise an additional 3dB downlink coverage, but more importantly a significant increase in downlink service rates; not only would wanted signal be increased, but also the adaptive Beam-forming allows interference to be nulled thus enhancing C/I statistics across the cell and network. The challenge for operators is to support multiple bands at site, particularly on key Macro base station sites. Tomorrow s operators in Europe will need to support 800, 900, 1800, 2100 and 2600 bands on existing sites. More RF spectrum is promised over the next decade in more new bands too. Many base station antenna vendors are innovating to try and support multiple bands with independent beam tilting per band on the same single conventional form factor antenna. Elevation Beamforming is also being explored by some antenna vendors, as opposed to using two side-by-side arrays for Azimuth Beamforming, as discussed above. Such antenna array topology lends itself to multi-banding, e.g. a high-band (>1GHz) and lowband (<1GHz) array elements can be inter-laced vertically with one another with a common ground-plane and symmetric axis. Elevation Beamforming at high-band would also be able to exploit the full length of the antenna (e.g. up to 2.6m) to achieve Page 19

Beamforming gains in the elevation plane. Quintel technology 6 is an example of a company innovating in this multi-banding space, plus offering the Beamforming capabilities for LTE in both Azimuth and Elevation, as discussed above. As an example, and food for thought concerning rural coverage, conventional wisdom says that an LTE800 service will have superior coverage over an LTE2600 service. Based upon LTE for UMTS: OFDMA and SC-FDMA based radio access, Harri Holma and Antti Toskala, P227 7, there is about an 11dB link budget delta between 900MHz and 2600MHz services for the Rural radio environment. Figure 12: Extract from LTE for UMTS: OFDMA and SC-FDMA based radio access, illustrating the link budget differences between 900MHz and 2600MHz 8 If however the 2600MHz antenna receives 3dB coverage gain due to Azimuth Beamforming and 3dB gain due to Elevation Beam-forming, and also receives a further 3dB coverage gain because there is twice the bandwidth for LTE at 2600MHz than at 800MHz, then there is only a 2dB delta. Furthermore, MIMO may perform better at 2600MHz than at 800MHz, and perhaps it could be shown that a 2600MHz service is 6 Quintel Technology: www.quintelsolutions.com 7 LTE for UMTS: OFDMA and SC-FDMA based radio access; Harri Holma, Antti Toskala, First Edition, John Wiley and Sons, 2009. P227 Page 20

better at achieving Rural coverage than 800MHz, for the same antenna aperture size of say 2.6m x 0.3m. Finally, there has been an emergence of active antennas recently. This is where each antenna element in an antenna array has its own Baseband Upconversion, Power Amplifier, filters, etc. and all share a common baseband unit, or simply the power amplifier is close coupled to a conventional antenna array; in effect the base station is integrated into the antenna, and a fibre connects the active antenna to the wider network/core. An active antenna array promises adaptive Beam-forming in the Elevation plane too. Such solutions are inherently single band at present but ideal for streetworks deployments where single band is required, and deployment of Remote Radio Units at the top, or Base Station cabinets at the bottom of the site is prohibitive, and/or where fibre connectivity is plentiful. Also, for active antennas which have elements driven from separate amplifiers, can have many Beam-forming options in Elevation (due to multiple elements; meaning many beams and nulls can be created for the radiation pattern) they are also suited very well for highly dispersive radio channel environments, where a user s signal can arrive over a wide and disparate range of angles; and as such are ideal for Microcell and Picocell type environments. Quintel is also engaged in the active antenna space, whereby Quintel s antenna technology allows Elevation Beam-forming, multi-banding, and support for independent variable electrical tilting for legacy services on a single antenna platform. Active antennas however don t necessarily increase spectral efficiency or increase coverage range. They can use less electricity/power (due to effective removal of RF losses) and with reference to this paper may not be necessarily suited for long range, rural deployments due to higher power RF requirements, cost of deployment, lack of fibre backhaul, and the fact most Rural sites would be expected to be able to accommodate base station cabinets, etc. Page 21

8 Conclusions The above analysis highlights the dependency of both network coverage and capacity on the available link budget, which in turn dictates how many base stations are required to reach a specified level of rural population coverage. Substantial savings (in terms of the network infrastructure required) and performance gains can be achieved by deploying user equipment signal enhancement techniques such as desktop modems, signal boosters or external antenna, compared to reliance on conventional mobile terminals (handsets or dongles). Equally at the base station, techniques such as increased power levels in rural areas, cell splitting, uplink diversity and beam-forming will provide significant benefits. As part of the Ofcom 800MHz and 2600MHz consultation 9 it is recommended that respondents consider the techniques proposed in this paper and the maximum EiRP limit, with a view to increase to improve rural services. A combination of the techniques discussed in this paper are likely to be essential if sufficient network capacity is to be available to provide the main broadband connections in areas not served by alternative fixed platforms. It must also be noted that poor RF performance in the end user equipment increases the demand for spectrum and radio network density. Recent studies have shown a large variation in radio performance for user equipment and in recognition of this challenge the ICT KTN Wireless Technology and Spectrum group voted this as a 2011 priority to address the longer term R&D challenges to improve radio front end technology. The working group is working closely with a newly formed Cambridge Wireless Radio Technology SIG 10 to address this challenge and will be hosting a series of events and workshops to address this topic. Wireless technology development is a major opportunity for the UK to lead the way globally in deployment, technology innovation, policy and regulation. The key to unlocking this potential is collaboration at all levels. If this can be achieved we will be able to realise significant societal benefits whilst enabling our industry to benefit across the whole of the value chain. The concepts provided in this document can significantly help the deployment of rural broadband services and increased capacity in high density areas within the UK and also help UK companies to trial and test concepts to enable global exploitation. 9 Ofcom consultation published 22 nd March 2011, Assessment of future mobile competition and proposals for the award of 800 MHz and 2.6 GHz spectrum and related issues http://stakeholders.ofcom.org.uk/consultations/combined-award/ 10 Cambridge Wireless Radio Technology SIG http://cambridgewireless.co.uk/sigs/radiotechnology/ Page 22

The ICT KTN Wireless Technology and Spectrum working group believe we can act as the catalyst for this collaboration and is keen to take this to the next phase. The diagram below demonstrates how the UK eco-system could provide a more holistic joined up strategy to maximise the benefit for the UK. Driving Innovation INDUSTRY Wireless Technology & Spectrum working group Funding R&D to solve the major challenges Innovation, collaboration and knowledge transfer Identifying & prioritising challenges Feasibility R&D to ensure technology can meet market, spectrum and economic challenges Policy and Regulation Flexibility to implement R&D trials Spectrum, policy & regulation Global exploitation Figure 13: Maximising wireless opportunities through Government & Industry collaboration Page 23