Exposure to Radio Waves near Mobile Phone Base Stations

NRPB-R321

Exposure to Radio Waves near Mobile Phone Base Stations

S M Mann, T G Cooper, S G Allen, R P Blackwell and A J Lowe

Abstract

Measurements of power density have been made at 17 sites where people were concerned about their exposure to radio waves from mobile phone base stations and where technical data, including the frequencies and radiated powers, have been obtained from the operators. Based on the technical data, the radiated power from antennas used with macrocellular base stations in the UK appears to range from a few watts to a few tens of watts, with typical maximum powers around 80 W. Calculations based on this power indicate that compliance distances would be expected to be no more than 3.1 m for the NRPB guidelines and no more than 8.4 m for the ICNIRP public guidelines. Microcellular base stations appear to use powers no more than a few watts and would not be expected to require compliance distances in excess of a few tens of centimetres.

Power density from the base stations of interest was measured at 118 locations at the 17 sites and these data were compared with calculations assuming an inverse square law dependence of power density upon distance from the antennas. It was found that the calculations overestimated the measured power density by up to four orders of magnitude at locations that were either not exposed to the main beam from antennas, or shielded by building fabric. For all locations and for distances up to 250 m from the base stations, power density at the measurement positions did not show any trend to decrease with increasing distance. The signals from other sources were frequently found to be of similar strength to the signals from the base stations of interest.

Spectral measurements were obtained over the 30 MHz to 2.9 GHz range at 73 of the locations so that total exposure to radio signals could be assessed. The geometric mean total exposure arising from all  radio  signals  at  the  locations  considered  was  2 millionths  of  the  NRPB  investigation  level,  or 18 millionths of the lower ICNIRP public reference level; however, the data varied over several decades. The maximum exposure at any location was 230 millionths (0.023%) of the NRPB investigation level, or 1800 millionths  (0.18%)  of  the  ICNIRP  reference  level.  The  exposures  are  therefore  well  within guidelines and not considered hazardous.

National Radiological Protection Board Approval date: May 2000 Chilton Publication date: June 2000 Didcot £15.00

Oxon OX11 0RQ ISBN 0 85951 455 2

This NRPB report reflects understanding and evaluation of the current scientific evidence as presented and referenced in this document.

Contents

1 Introduction 1

1. Background 1
2. Objectives 1

2 Base station characteristics 1

1. Principles of cellular radio networks 2
   1. Providing coverage 2
   2. Power control and network capacity 3
   3. Cell sectors 4
2. General technical aspects 4
   1. Electrical characteristics 5
   2. Frequency allocations 5
   3. GSM signal waveforms 6
   4. Antenna beam shapes 7
   5. Antenna gain 8
3. Specific data on radiated powers 9
   1. Data from standards 9
   2. Total radiated power 9
   3. Effective isotropic radiated power 10
4. Microwave links 10

3 National and international exposure guidelines 11

1. Basis for guidelines 11
2. NRPB guidelines 11
   1. Basic restrictions 12
   2. Investigation levels 12
3. ICNIRP guidelines 13
   1. Basic restrictions 13
   2. Reference levels 14
4. Summary of guidelines 15
5. Exposure assessment 16

4 Measurement equipment and surveying procedures 16

1. Instrumentation 16
   1. Measurement antennas 17
   2. Spectrum analyser 17
   3. Data acquisition 18
2. Data processing 18
   1. Calculation of power density 18
   2. Spreadsheet calculations 18
   3. Example results 19
3. Surveying technique 19
   1. General method 19
   2. Limitations 19
4. Measurement uncertainty 20

5 Results of exposure assessments 21

1. Calculation of exposures 21
   1. Far-field power density calculation 21
   2. Near-field power density evaluation 22
   3. Direct SAR evaluation 22
2. Predicted compliance distances 22
   1. GSM900 base station antennas 23
   2. GSM1800 base station antennas 23
3. Power density at ground level 24

4. Measured power densities 25
   1. Base stations of interest 25
   2. Comparison with theory 26
   3. All environmental signals 27
   4. Statistical analysis 27
5. Exposures produced by base station and other signals 29

1. Exposure due to signals from all sources 29
2. Exposure due to the base station of interest 31
3. Exposure from all base stations 31

6 Conclusions 33 7 Acknowledgements 34 8 References 34 Appendix A Technical Data from Operators 35 Appendix B Measured Power Density Spectra 41

1 Introduction

1. Background

Liberalisation of the UK telecommunications market, combined with the development of digital radio technology in the 1980s, stimulated the development of mobile phones resulting in the proliferation of new radio antenna sites throughout the 1990s. Many of these new antenna sites are close to people's homes and workplaces and this, combined with their rapid deployment, has led to concern that the radio waves transmitted could be harmful to health.

NRPB has carried out many surveys in response to requests for measurements of radio wave signal strength at and around specific new radio antenna sites; the data from 17 different site visits have been drawn together to form this report. Eight of the sites were schools with base station antennas mounted on their roofs and the majority of the remaining sites were residential tower blocks. The first site was visited in Spring 1998 and measurements have been performed at a number of locations at each site.

The selection of sites is inherently biased because the sites were chosen on the basis of specific requests. Furthermore, the locations where measurements were made were chosen to be where the highest power densities would be expected to occur. In view of this bias, care should be taken not to over-interpret the data or extend its results to the UK population as a whole. A much larger study with more carefully selected locations would be necessary to produce results of validity to the population as a whole.

2. Objectives

The objectives of this report are:

1. to provide factual information about exposures,
2. to clarify further the basis for the NRPB scientific position.

2 Base station characteristics

The first generation of mobile phone base stations was deployed in the UK in the early 1980s and used the analogue Total Access Communication System (TACS). At this time, there were only two operators (Vodafone and Cellnet) and the  number of  mobile phone users  was considerably less than that today. For these reasons fewer base stations were required to fulfil operational considerations with consequential lesser impact on the environment. TACS systems are still in existence; however, the number of users is dwindling and the radio spectrum used by these systems is progressively being reallocated. The networks are scheduled to be shut down before the year 2005.

The arrival of second generation systems based on the digital Global System for Mobile Telecommunications (GSM) initially required more sets of antennas to be installed at existing radio sites; however, increasing usage of mobile phones quickly led to new sites being required. It was with the licensing of two more network operators (Orange and One 2 One), and the need to install two completely new networks in the early 1990s, that the number of base station sites began to increase rapidly. It is estimated that there are around 20,000 base stations in the UK today.

Third generation networks using the Universal Mobile Telephone System (UMTS) are now being planned and the Government is to issue five licenses as a result of the recent spectrum auction. Whilst some reuse of existing base station sites is anticipated, it is only when there is no further increase in the use of mobile phones that the number of base stations will stabilise.

This report includes a total of 17 sites visited by NRPB where one or more base stations have been installed. In addition to performing measurements, the operators of the base stations have been approached in order to obtain details of radiated powers, antenna characteristics etc. This section begins with a general description of the operation of cellular radio networks for mobile telecommunications and then summarises the specific information that has been obtained for the sites visited.

1. Principles of cellular radio networks

Most people are familiar with the use of radio to permit wireless communication of signals between transmitting and receiving antennas. Perhaps the most familiar example is the network of very tall towers (eg Crystal Palace, Sutton Coldfield and Emley Moor) that are used to broadcast television signals to the antennas (aerials) that most houses have mounted above their roofs. Mobile phones communicate by radio signals passing to and from an antenna mounted on the phone and antennas connected to the base station. The radio link from the phone to the base station is known as the uplink and carries the speech from the mobile phone user. A separate radio link from the base station to the phone is known as the downlink and this carries the speech from the person to whom the phone user is listening. This principle is illustrated in Figure 1.

Antennas

Radio signals

Uplink Radio signals Antenna on cables

Downlink

Plant room Mobile phone containing

base station

FIGURE 1 Radio signals used for communication between mobile phones and base stations

The antennas connected to the base station tend to be mounted high above ground level because the radio signals would be blocked by buildings etc if the antennas were nearer the ground. Antennas used with macrocellular base stations are generally placed between 15 and 50 m above ground  level  because  they  are  designed  to  provide  communications  over  distances  of  several kilometres. However, microcellular base stations have their antennas mounted nearer ground level as communications are only carried out over distances of a few hundred metres. Antennas tend to be mounted directly on existing structures, such as buildings, when this is convenient, but ground- based lattice towers, shorter masts mounted on roofs, and lamp-post type systems are also used.

1. Providing coverage

Transmitted signal strength falls off rapidly with distance from base stations and mobile phones, but a certain minimum signal strength is required for adequate reception. The current generation of GSM base stations cannot communicate over distances greater than 35 km because the delay in receiving radio signals becomes too great. However, the decline of signal strength with

distance places a practical limit on coverage of around 10 km. This means that a large number of base stations is needed to provide coverage of the whole of the UK by all four current networks.

The use of a number of base stations to provide complete coverage of an area of land is illustrated by Figure 2. The figure shows how the area covered by each base station can be regarded as a hexagon if there is a fixed distance between neighbouring base stations. In practice, the location of base stations is influenced by many factors so cells vary in shape and size.

Mobile phone

Cell

Base station

Radio signals

FIGURE 2 Radio signals travelling via the uplink from a mobile phone to a base station and via the downlink in the reverse direction

2. Power control and network capacity

An important design consideration that tends to limit the radiated powers of base stations is the desire of operators to use the available radio spectrum as efficiently as possible. Network operators have a certain number of radio channels assigned to them and they aim to use these for the maximum number of mobile phone users. This is achieved by reusing any given radio channel many times in a network and carefully controlling base station powers so that signals arising in different parts of the network do not interfere. Figure 3 shows how four different radio channels can be shared between the cells in a network with no adjacent cells having to use the same channel.

FIGURE 3 Frequency reuse in a cellular network. Cells that have the same shading use the same radio channels as each other

The  power  radiated  by  base  stations  has  to  be  carefully  controlled  in  any  frequency reuse scheme to limit the distance travelled by signals. If a base station were to transmit with too much power, its signals might be strong enough to interfere with signals in other cells using the same radio channel. An example of frequency reuse in another area of technology is in the choice of channels for broadcast radio in the FM band; a radio station in one town can use the same frequency as another provided powers are limited so that the effective range does not give rise to interference.

3. Cell sectors

Operators tend to divide the area about a base station into three sectors and then mount three different sets of antennas on a mast such that each set provides coverage of a 120º arc about the mast. Most base station antennas are oriented such that Sector 1 is directed towards grid north, Sector 2 is directed 120º east of grid north (EGN) and Sector 3 is directed 240º EGN. This gives rise to the familiar triangular configuration that is seen at the head of many of the older base station masts, as depicted in Figure 4. In some cases omni-directional antennas are used to provide full 360º coverage.

Sector 1

Antenna 1 Antenna 2

Antenna 2 Antenna 1 Grid north

Sector 3 Sector 2

Antenna 1 Antenna 2

FIGURE 4 Six antennas are arranged on some masts in order to provide

coverage of a cell divided into three sectors. Arrows indicate the directions of signals to and from the antennas

Figure 4 shows how a single antenna can be used to transmit signals, whereas two antennas are used to receive signals from a sector. This arrangement of diversity reception allows con- tinuous operation even if one of the antennas experiences a reduced or faded signal. Some of the more  modern  antennas  effectively  contain  two  sets  of  receiving  elements  that  are  arranged perpendicularly to each other inside a single case. These dual polar antennas give rise to much more compact mast heads since only three are used to cover a cell.

2. General technical aspects

When discussing base stations, it is important to be clear that, in strictly engineering terms, it is the electronic equipment contained in the plant room shown in Figure 1 that is the base station. Nevertheless,  it  has  become  common  practice  to  describe  the  complete  installation,  including antennas and mast, as the base station. The features of a typical base station installation are shown in Figure 5.

Antenna

Gain, G

Loss in cables and

combiner, L Beam

Power density, S, at distance, d

Plant room

Mast

Base station

trx 1

trx 2 Base station contains Combiner N transmitters, each with

an output power, Ptx

trx N

FIGURE 5 A base station contains a number of radio transmitters (trx) whose outputs are combined before being fed to an antenna and transmitted as radio waves

1. Electrical characteristics

Base stations contain a number of radio transmitters and each of these  has the same maximum output power, Ptx. The outputs from the individual transmitters are then combined and fed via cables to the base station antenna, which is mounted at the top of a mast (or other suitable structure). It therefore follows that the total power fed into the base station antenna, Pant, is given by

Pant NPtx10  L/10 (1)

where N is the number of transmitters and L is the loss (in decibels) in signal strength that occurs in the combiner and connecting cables.

The power that is fed into the base station antenna is launched into a radio wave travelling away from the tower and the strength of this radio wave decays with distance from the antenna according to the inverse square law. The power density, S, in the beam thus varies with distance, d, according to the following expression

NP

S  tx 10(G  L)/10 (2)

4  d 2

The antenna gain, G (in decibels), is discussed further in Section 2.2.5 and is a measure of how much the antenna is able to focus the radiated power in the direction of its beam. It should also be noted that equation 2 is strictly only valid at distances greater than around 10 m from a typical sector antenna and will overestimate the power density at lesser distances (see Section 5.2).

2. Frequency allocations

The frequency bands used by current and future mobile phone networks in the UK are as shown in Table 1. Each frequency band contains a large number of channels and these are

shared between the operators according to licenses issued by the Radiocommunications Agency. The analogue TACS systems operate close to 900 MHz, as do some of the GSM systems, which will  be  denoted  as  GSM900  in  this  report.  Other  GSM  systems  operate  close  to  1800 MHz and these will be referred to as GSM1800 in this report. UMTS systems will operate close to 2000 MHz, although the structure of uplink and downlink bands is more complicated than with the GSM systems.

TABLE 1 Frequency bands currently allocated to mobile phone networks in the UK. Base stations transmit on the downlink frequencies and mobile

phones transmit on the uplink frequencies

Frequency band (MHz) Channel spacing Number of System Uplink Downlink (kHz) channels TACS 872–888 917–933 25 640 GSM900 890–915 935–960 200 174 GSM1800 1710–1785 1805–1880 200 374 UMTS Various between 1900 and 2200 5000 –

Vodafone and Cellnet use predominantly GSM900 for their networks, although they do have some channels allocated in the GSM1800 band that may be used at some future date. Orange and One 2 One operate purely in the GSM1800 band.

Each GSM radio channel consists of paired uplink and downlink frequencies that are exactly 45 MHz apart for GSM900 and 95 MHz apart for GSM1800. This principle of paired frequency bands is illustrated in Figure 6.

Mobile transmits Base station transmits (uplink) (uplink)

880 890 900 910 920 930 940 950 960 Frequency (MHz)

FIGURE 6 Structure of paired frequency bands showing how transmit and receive frequencies are separated by 45 MHz with GSM900

3. GSM signal waveforms

GSM base station antennas all transmit at least one radio signal quasi-continuously and this  is  known  as  the  Broadcast  Control  Channel  (BCCH)  carrier  because  it  carries  important signalling  information  that  is  used  to  set  up  calls.  The  BCCH  carrier  can  also  handle  up  to seven  mobile phone  calls  simultaneously,  so  it  may  give  sufficient  capacity  for  base  stations in lightly loaded areas. Where there is a potential need for more than seven phone calls at the same time, a base station can be configured to transmit extra carriers (non-BCCH carriers), each allowing the base station to provide a further eight mobile phone calls. For example, a GSM base station  equipped  to  transmit  four  radio  carriers  could  control  up  to  thirty-one  mobile  phone calls simultaneously.

GSM base stations use Time Division Multiple Access (TDMA) within each radio carrier so a base station communicates  with  any  given  mobile  phone  by  sending  out  217  frames  of information every second. Each frame is divided into eight timeslots, as illustrated in Figure 7, and each timeslot is used for a particular phone call (or for setting up calls in the case of timeslot zero in the BCCH carrier).

The BCCH carrier is transmitted at full power in all eight timeslots, even when no calls are being handled, whereas the non-BCCH carriers (if available) are only transmitted when calls are present.  The  BCCH  carrier  is  described  as  quasi-continuous  because,  although  the  waveform is transmitted  at  full  power  during  each  timeslot,  the  power  envelope  shows  transient  dips between timeslots.

The information carried by the radio waves from base stations is encoded as small changes in the frequency of the underlying sinusoidal radio carrier, hence the signals are described as frequency modulated.

BCCH carrier Timeslot number

0 1 2 3 4 5 6 7 0 1 2 3

Time

BCCH BCCH Non-BCCH carrier

Full power

Time

One frame = 4.615 ms

FIGURE 7 Waveforms of the signals produced by GSM base stations showing how

the BCCH carrier is a continuous sinusoid at full power, whereas non-BCCH carriers can have partial occupancy and also use power control

4. Antenna beam shapes

The radio signals developed by base stations are fed to antennas, which produce beams that are radiated into the cell around the base station. The profile of the beams is carefully chosen by the network planners in order to produce optimal coverage of the cell, but the general principle of beam formation is illustrated in Figure 8.

The beams formed by antennas used with macrocellular base stations are narrow in the plane of elevation with typical widths between 5º and 10º. The beams are also tilted slightly downwards so the top edge of the main beam is approximately horizontal whereas the lower edge is directed up to 10º below horizontal. When considering the heights at which antennas tend to be mounted, this implies that the main beam from base station antennas would be expected to reach ground level typically between 50 and 300 m from the foot of a mast. The antennas used with microcellular base stations have much broader beams in the plane of elevation because they are intended to communicate over much shorter distances.

Base station Antenna

mast

5 10

Beam 15 50 m

50 300 m

Ground

FIGURE 8 Elevation showing the shape of the beam formed by a typical antenna used with a macrocellular base station

Figure 8 shows a simplified version of the directional properties of an antenna illustrating why much lower radio wave strengths are found at the foot of a mast than at distances of around 100 m from the mast (see Section 5.3). The beams from real antennas do not have sharply defined lower edges and some power will be directed at all angles below horizontal. Typically the power in the downwards direction is at least a hundred times weaker than in the main beam at the same distance from antennas.

5. Antenna gain

When considering the directional properties of antennas, it is useful to refer to the antenna gain. This is a measure of how effective an antenna is at radiating power in the direction of its main beam. An isotropic antenna is an antenna that radiates equally in all directions, and if a spherical surface enclosing such an antenna is considered, the power density, S, at a radial distance, d, would be given by

S  Prad (3)

4  d 2

where Prad is the total radiated power.

Any real antenna will employ some means to remove power from undesired radiating directions and channel it into the intended direction of the beam. This means that the power density at a distance, d, will be greater than that given by equation 3 by a factor equal to the antenna gain. Gain is normally quoted in decibels relative to an isotropic radiator in the unit dBi.

The beams from antennas used with base stations are narrow in the plane of elevation (see Figure 8), and this is achieved by mounting a stack of radiating elements vertically above each other inside the antenna cases. The taller the stack is in relation to the wavelength, the narrower the beamwidth that is achieved. Mounting reflectors around the radiating elements inside the antenna case can be used to narrow beam widths in the azimuth plane, to between 60º and 120º in order to produce sector antennas. Both of these design techniques cause antennas to radiate preferentially in a certain direction, and hence form a main beam.

Typical gains for the sector antennas used with macrocellular base stations in the UK are in  the  range  15–17 dBi  for  GSM900  systems  and  16–18 dBi  for  GSM1800  systems.  Omni- directional antennas for macrocellular base stations are much less common than sector antennas, but generally have gains in the range 8–10 dB.

Microcellular base station antennas  are  not  intended  to  communicate  over  such  large distances as macrocellular base stations so their beams tend to be wider and their  gains tend to be lower.

The specific data on antenna gains for the sites considered in this report are given in

Appendix A.

3. Specific data on radiated powers

The power assigned to a given base station is determined from its coverage and capacity requirements; however, the operator's desire to use the licensed spectrum as efficiently as possible will tend to minimise powers used in every cell (see Section 2.1.2). NRPB asked the operators to supply the powers of the base station sites that were visited and this section reviews these data in the context of the licensed powers and technical standards.

1. Data from standards

Published  standards  for  base  station  transmitters  provide  specifications  for  the manufacturers. A variety of different power classes are defined in the GSM standard1 and these are as shown in Table 2. It should be noted that these are the powers at the output of each transmitter and should not be confused with the power radiated by the antenna (see Figure 5). Power radiated by the antenna will be considered in Section 2.3.2.

Table 2 is a classification system but does not imply that base stations exist with every power level. For example, it is highly unlikely that there would ever be a need for the power of a Class 1 or 2 GSM900 base station, given the communication distances involved. It is understood that  20 W  and  40 W  are  the  normal  maximum  powers  for  individual  transmitters  used  with GSM900 macrocellular base stations used in the UK.

2. Total radiated power

As described in Section 2.1.2 and illustrated in Figure 5, base stations often contain more than one transmitter and the outputs of each transmitter are combined before being fed via cables to the radiating antennas. When the signals are combined, the radiated power would ideally be equal to the sum of the output powers from the transmitters, but some loss occurs in the combiner and connecting cables. This  combiner loss (normally taken  to  include  the  cable  loss)  is  generally between 4 and 6 dB so the power radiated by the antennas will be less than half of that produced by the transmitters.

The technical data from the sites visited by NRPB (see Appendix A) suggest that each transmitter gives rise to a maximum radiated power in the region of 10 W at the antenna. Most of

the base stations encountered during the surveys had either one or two transmitters, but two had four and one had five. Site A was quoted as having eight transmitters, but the data given for this site were believed to be generic and not particular to the actual site, as no more than two signals were measured. It will be assumed that the maximum power radiated from a base station antenna is 80 W in this report. This could be taken as equivalent to eight transmitters each producing 10 W at an antenna.

Only the mobile phone industry can give a definitive statement as to the maximum power radiated by base station antennas, as the technology is rapidly developing and power levels may change. Personal communications between the authors and professionals working in the industry indicate that maximum powers radiated from macrocellular antennas in the UK are currently in the region of 25 to 70 W. Microcellular antennas would be expected to radiate no more than a few watts given the short distances over which they are designed to communicate. In a safety guide issued by one operator, it is stated that its smaller base stations radiate no more than 2 W.

3. Effective isotropic radiated power

The power density formed in the beam from a base station depends upon the radiated power and on the gain of the antenna. The product of the power radiated and the antenna gain is known as the effective isotropic radiated power (EIRP) and is usually quoted in decibels relative to a milliwatt (dBm). By convention, the EIRP is quoted in terms of the power radiated by a single transmitter and this should be taken into account when calculating total radiated power.

The licenses allocated to the operators by the Radiocommunications Agency stipulate that no more than 62 dBm EIRP may be radiated (per transmitter). This arises from considerations associated with the possibility of interference with other electrical equipment in the environment, and is not related to electromagnetic field safety. Ensuring compliance with protection guidelines is a matter for the operators through their general safety obligations.

The EIRPs supplied by the operators for the sites visited during this work (see Appendix A) ranged from 44 to 56.7 dBm. Based on this information, the typical powers of transmitters and the range of antenna gains, it is probable that very few base stations in the UK produce appreciably more than 56 dBm EIRP.

4. Microwave links

Base stations must be able to communicate with other neighbouring base stations in order to relay calls between  mobile phone  users in two different cells and connect calls into  other networks. In some cases this is achieved using cables but it is more usual for base stations to communicate via microwave links. Microwave links employ dish antennas that permit point-to- point communications, as shown in Figure 9.

Dish antenna

Conical beam less than 2 wide

FIGURE 9 Pair of dish antennas used as terminals for a point-to-point microwave link

Technical information on the powers  used  with  microwave dishes at the sites visited during this work is given in Appendix A. The frequencies used are mostly in bands spread about 13, 23 and 38 GHz and propagation at these high frequencies is such that a line of sight path must exist between the dishes.

Dish antennas produce narrow conical beams that are 1–2º wide. Typical powers are no more than a few tens of milliwatts because the power is channelled so selectively towards the receiver. The powers used by dish antennas are very much lower than those used by base station antennas so the exposures produced by signals from dish antennas will be negligible in comparison.

3 National and international exposure guidelines

A number of different national and international organisations have published guidelines for the protection of people from exposure to electromagnetic fields and radiation. It is a matter for policy makers to determine which set of guidelines should be adopted; however, this section reviews the two sets of guidelines of most importance from a UK perspective.

1. Basis for guidelines

There is a consensus amongst organisations responsible for providing advice on exposure to electromagnetic fields and radiation. The consensus is as follows.

1. Heating can occur as a consequence of exposure to electromagnetic fields at telecommuni- cations frequencies.
2. The established adverse effects on people's health occur at exposure levels where heating would be expected to occur.
3. Exposures should be restricted to avoid the established effects of exposure to electro- magnetic fields.
4. There is no convincing evidence that adverse effects can occur as a result of exposure within the current protection guidelines.

The views of the National Radiological Protection Board (NRPB) and the International Commission on Non-Ionizing Radiation Protection (ICNIRP) form part of this consensus.

2. NRPB guidelines

The advice relevant to radiofrequency radiation provided by NRPB is contained in published guidelines2 on restricting human exposure to time-varying electromagnetic fields and

radiation with frequencies up to 300 GHz. The guidelines have been developed following comprehensive reviews of scientific data and on the advice of the NRPB Advisory Group on Non- ionising Radiation (AGNIR).

There is no specific legislation in the UK that relates to protection from electromagnetic fields. Nonetheless, there is enabling legislation in the general area of health and safety that places a duty of care' on the operators of equipment generating electromagnetic fields. UK Government Departments and Agencies, including the Health and Safety Executive, have looked to compliance with the NRPB guidelines in order to fulfil this responsibility. Notwithstanding, the status of the ICNIRP guidelines has recently been clarified with respect to limiting the exposure of members of the public within the member states of the European Union (see Section 3.3).

1. Basic restrictions

The  survey  work  described  in  this  report  was  concerned  only  with  exposure  to electromagnetic fields with frequencies greater than 10 MHz. The established biological effects of exposure to such fields occur at levels whereby exposure results in significant heating of part or all of the body. The guidelines advise basic restrictions on the rate of energy absorption per unit mass of body tissue, or specific energy absorption rate (SAR), to ensure that harmful temperature rises do not occur in the body. SAR is quantified in the unit watt per kilogram (W kg1). Compliance with NRPB guidelines is demonstrated by showing that none of the basic restrictions is exceeded.

The four basic restrictions on SAR that apply for exposure to electromagnetic fields with frequencies between 10 MHz and 10 GHz are listed in Table 3. The restrictions permit short-term time-averaging so transient exposures may be averaged over specified periods before comparison with the restrictions. The restrictions on localised SAR vary over different regions of the body and apply to contiguous tissue within the region specified.

TABLE 3 NRPB basic restrictions on exposure to electric and

magnetic fields in the frequency range 10 MHz to 10 GHz

SAR averaged over the body and over any 15 minute period 0.4 W kg1 SAR averaged over any 10 g in the head or fetus and over 10 W kg1

any 6 minute period

SAR averaged over any 100 g in the neck and trunk and over 10 W kg1 any 6 minute period

SAR averaged over any 100 g in the limbs and over 20 W kg1 any 6 minute period

At higher frequencies, energy absorption becomes increasingly confined to the surface layers of the skin and the heating effect is directly related to the power density of the incident radiation. Consequently, for frequencies between 10 and 300 GHz there is a single basic restriction of 100 W m2 that applies to any part of the body. This restriction may be averaged over a period equal to 68/f 1.05 minutes, where f is the frequency in gigahertz.

2. Investigation levels

SARs  are  not  easily  measurable  in  living  people,  therefore  NRPB  has  introduced investigation levels of external electric and magnetic field strength and power density that may be compared directly with measured or calculated exposure levels. Compliance with the appropriate investigation levels in a given exposure situation ensures compliance with the basic restrictions; however, the investigation levels are not limits and exceeding them does not necessarily mean that basic restrictions are exceeded. The investigation levels are frequency dependent and are shown in Table 4 for frequencies greater than 12 MHz. All investigation levels are specified as root mean square (rms) values.

TABLE 4 NRPB investigation levels for exposure to electric and magnetic fields in the frequency range 12 MHz to 300 GHz

f is the frequency in GHz.

The three parameters used for investigating compliance with NRPB guidelines are related to each other and in certain instances it is only necessary to measure one of them. During the survey work in this report, electric field strength was measured and the results were subsequently converted to power density as this quantity can be most readily used to normalise exposure to the transmitter power.

The investigation levels have been derived from the basic restrictions using dosimetric models assuming conservative but nevertheless realistic assumptions of exposure. The assumptions that have been used are as follows.

1. The exposed person is stood vertically in contact with the ground.
2. The ground beneath the person is a good conductor.
3. The electric field is vertically directed.
4. The electric field is uniform over the space occupied by the person.

In practice it is unlikely that all four conditions would be simultaneously fulfilled, therefore the investigation levels will be conservative indicators of the fields required to produce SARs equal to the basic restrictions.

3. ICNIRP guidelines

The International Commission on Non-Ionizing Radiation Protection (ICNIRP) is an independent scientific organisation responsible for providing guidance and advice on the health hazards of non-ionising radiation exposure. ICNIRP develops international guidelines on limits of exposure to non-ionising radiations and the most recent guidelines on limiting exposure to electromagnetic fields were published in April 19983,4.

The profile of the ICNIRP guidelines has recently been raised within Europe because the European Council has published a Recommendation5 on the limitation of exposure of the general

public to electromagnetic fields. The basic restrictions and the reference levels in the document are identical to those advised by ICNIRP; however, implementation and policy matters are also addressed in the European Council Recommendation.

NRPB recently issued advice on the 1998 ICNIRP guidelines6. For occupational exposure the ICNIRP guidelines do not differ in any significant way from those previously recommended by NRPB. For members of the public, ICNIRP has generally included a reduction factor of up to five in setting basic restrictions. It is the view of NRPB that the existing UK guidelines on limiting exposure for the general public provide adequate protection and the health benefits to be obtained from further reductions in exposure have not been demonstrated.

While recognising that there was no scientific basis for exposure limits to avoid potential harm from athermal* effects of exposure to microwaves, a House of Commons Select Committee on Science and Technology concluded in a recent report7 that the Government should adopt the

ICNIRP recommended guidelines for microwave exposure as a precautionary measure'.

1. Basic restrictions

The ICNIRP guidelines specify basic restrictions on SAR which are analogous to the NRPB basic restrictions and state that protection against adverse health effects requires that these basic restrictions are not exceeded'. The basic restrictions on SAR that apply to frequencies within the range 10 MHz to 10 GHz for occupational and public exposure are given in Table 5. All

* Biological effects that have been claimed to arise at exposure levels where heating could not play a role are often referred to as athermal.

restrictions are to be time-averaged over a six minute period. The restrictions on localised SAR permit averaging over a 10 g mass of contiguous tissue.

For exposures  to  radio  waves  with  frequencies  between  10  and  300 GHz,  the  ICNIRP guidelines stipulate basic restrictions on power density alone. The basic restrictions are 50 W m2 for occupational exposure and 10 W m2 for exposure of the general public. Provision is also made

for averaging power density over specified areas of the body surface in this frequency range so the ICNIRP guidelines are complied with if the following two conditions are satisfied.

1. The power density averaged over 20 cm2 is less than the basic restriction.
2. The power density averaged over 1 cm2is less than 20 times the basic restriction.

TABLE 5 ICNIRP basic restrictions on exposure to electric and magnetic fields in the frequency range 10 MHz to 10 GHz for occupational and general public exposure

2. Reference levels

A  system  of  reference  levels,  similar  to  the  NRPB  investigation  levels,  is  given  by ICNIRP and the levels reflect the factor of five difference between the public and occupational basic restrictions. Compliance with the reference levels ensures compliance with the ICNIRP basic restrictions.  The  reference  levels  for  occupational  exposure  to  electromagnetic  fields  in  the frequency range 10 MHz to 300 GHz are given in Table 6. Reference levels for exposure of the general public are given in Table 7.

TABLE 6 ICNIRP reference levels for occupational exposure to electromagnetic fields in the frequency range 10 MHz to 300 GHz

f is the frequency in MHz.

TABLE 7 ICNIRP reference levels for general public exposure to electromagnetic fields in the frequency range 10 MHz to 300 GHz

f is the frequency in MHz.

4. Summary of guidelines

The NRPB investigation level and the ICNIRP reference levels both vary according to the frequency of the radio  waves that  produce  a  given  exposure.  Figure 10  shows  this  frequency dependence and compares the levels.

100 50

20 10 5

2 1

10 100 1,000 10,000

Frequency (MHz)

FIGURE 10 Comparison of the NRPB investigation level and the ICNIRP reference levels for occupational and public exposure

The  investigation/reference  levels  tend  to  be  most  restrictive  over  the  10–200 MHz frequency range because electromagnetic energy couples most efficiently into the body over this range. The NRPB investigation level is around 15 times lower over this frequency range than at frequencies of several gigahertz. The ICNIRP reference level for occupational exposure is higher, ie less restrictive than the NRPB investigation level at lower frequencies; however, this situation is reversed at higher frequencies. The ICNIRP public reference level for power density is between a factor of 3 and 13 lower than the NRPB investigation level and this reflects the reduction factor of 5 in the underpinning basic restriction.

Figure 10 shows the two main frequency bands at which mobile phone base stations used in the UK transmit, ie close to 940 MHz and close to 1840 MHz (see Section 2). The lowest values of the investigation/reference levels over these bands are summarised in Table 8.

TABLE 8 NRPB investigation level and the ICNIRP reference levels for public and occupational exposures at the transmit frequencies of UK base stations

Power density (W m2) Guidelines TACS/GSM900 GSM1800 NRPB (all people) 35 100

ICNIRP workers 23 45 ICNIRP public 4.6 9

5. Exposure assessment

For exposure to radio waves emitted at a single frequency, a dimensionless quantity known as the exposure quotient may be calculated. This exposure quotient is expressed in terms of the measured power density S meas and the power density investigation level S inv using the relation

Smeas

exposure quotient (4)

Sinv

Compliance with NRPB guidelines is demonstrated if the exposure quotient is less than unity (compliance with ICNIRP guidelines can also be shown using the appropriate ICNIRP reference level in place of the NRPB investigation level).

This report is concerned with simultaneous exposure to many different radio signals, each with a different frequency. All of the individual signals will contribute to a person's exposure and the total exposure quotient will be equal to the sum of the quotients for each signal, as expressed by

N Smeas Smeas Smeas S meas S meas

total exposure quotient i 1 2 3 N (5)

Sinv Sinv Sinv Sinv Sinv

i  1 i 1 2 3 N

where  N is the total number of signals. Again, a total exposure quotient not exceeding unity indicates compliance with the guidelines.

Exposure quotients offer a convenient form of expressing the exposure due to multiple radio signals present at any given location. For example, it might be stated that the power density measured at the specified location is equivalent to 10–3 or 1/1000 of the NRPB investigation level. Many of the exposures in the later sections of this report will be expressed in terms of how many millionths of the investigation/reference level they represent.

Exposure quotients may also be used to investigate the contributions of various individual signals (or categories of signals) to the total.

4 Measurement equipment and surveying procedures

This section describes the equipment used to measure the electric field strength of radio- frequency signals, the calculation techniques that are used to derive equivalent power densities, and the assessment of exposures at measurement locations. The limitations of the surveying techniques are discussed, as are the uncertainties in the resulting data.

1. Instrumentation

Hand-held hazard assessment probes can be obtained for measuring electric and magnetic field strengths at levels comparable with the investigation/reference levels given in protection guide- lines. The advantages of this type of equipment are that it is portable and may be suitable for use by non-specialists. The disadvantages are that the equipment lacks sensitivity and will tend to read zero at locations that are not in the immediate vicinity and directly in the beam of base station antennas.

The equipment used for the measurements in this report consisted of a spectrum analyser connected to one of a choice of antennas via a coaxial cable, as shown in Figure 11. This set-up allowed measurements to be made over a sweep of frequencies using a narrow measurement bandwidth, permitting the detection of power densities considerably below 1 µW m2. The data collected by the spectrum analyser were transferred to a laptop computer and stored for subsequent analysis in a spreadsheet program.

Antennas

Biconical

Data cable

Log-periodic

Spectrum Laptop

analyser computer Ridgeguide

RF cable

FIGURE 11 Equipment used for measuring electric field strength of environmental radiofrequency signals

1. Measurement antennas

Three  different  broadband  antennas  were  employed  in  conjunction  with  the  spectrum analyser for measuring electric field strength over different frequency ranges. The antennas were mounted on lightweight tripods providing stable support but permitting their orientation to be varied by hand. The types of antennas that are currently used are listed in Table 9 together with the range of frequencies specified for each model.

TABLE 9 Antennas used for the measurements in this report

The antenna factor of an antenna describes its sensitivity to an electromagnetic field and, in the case of the above antennas, it is defined as follows.

antenna factor  electric field strength at antenna (6)

voltage produced at antenna connector

Calibration  data  supplied  with  each  antenna  gave  the  antenna  factors  at  discrete  frequencies. Interpolations  were  then  performed  to  obtain  the  antenna  factors  at  the  precise  frequencies of interest.

The loss (dB) in the cable connecting the antennas to the spectrum analyser was measured by  substitution  between  a  signal  generator  and  a  power  meter  at  a  range  of  frequencies. Interpolation was used to derive the cable loss at the frequencies of the measured signals.

2. Spectrum analyser

The spectrum analyser was used to measure the strength and frequency of each radio signal over a set frequency range. Many other parameters have to be set on the spectrum analyser in order to measure radio signals with a given waveform correctly. Practically, this means that the radio

spectrum has to be scanned in several different parts and then these parts have to be assembled to produce a table covering the complete frequency range. The complete frequency range covered during the measurements in this report was from 30 MHz to 2.9 GHz.

The frequency resolution of the spectrum analyser was determined by the width of the frequency span used during a given sweep. Wider sweeps gave rise to poorer frequency resolution and  some  imprecision  in  the  measured  frequency  of  individual  signals.  This  imprecision  had negligible effect on exposure calculations; however, care had to be taken to resolve signals with frequencies that were close together.

The amplitude displayed by the spectrum analyser indicated the rms voltage at its input connector during the strongest part of each measured signal, ie when the transmission was at its maximum power. This has implications for signals with power modulation, eg video signals for television, that are discussed further in Section 4.3.2.

3. Data acquisition

A rugged laptop computer with a metal case and a daylight-readable screen was used to collect the data from the spectrum analyser. This computer was used with a PCMCIA card enabling it to communicate directly with the computer inside the spectrum analyser and accept the measured data directly into a spreadsheet on command. The format of the downloaded data was a table of the measured peaks on the spectrum analyser trace, each of which represented an individual radio signal. The first column in the table contained the signal frequencies in megahertz and the second column contained the signal strengths in millivolts.

2. Data processing

1. Calculation of power density

The spectrum analyser produced a list containing the frequencies in megahertz (MHz) and amplitudes in millivolts (mV) of each signal detected with the antennas. The following equation was then used to convert the received voltages, Vrx, into electric field strengths, E, corresponding to each signal.

E = Vrx F10L/20 (7)

where F is the antenna factor (m1) and L is the loss in the cable (dB).

The intrinsic impedance of free space, , was then assumed to be equal to 377 so the following equation could be used to calculate the power density, S, of each signal.

E2

S =  (8)

2. Spreadsheet calculations

Typically there may be a hundred or more radio signals measured at any given location and NRPB uses a spreadsheet template on a laptop computer:

1. to calculate the measured power densities (see equations 7 and 8),
2. to calculate the exposure quotients (equation 4),
3. to sum the exposure quotients to get the total exposure quotient (equation 5),
4. to assess the proportion of the total exposure arising from different signals.

3. Example results

In order to present measured data in reports, a table in the format shown in Table 10 is used. Where there is a specific interest in a given base station (or other source), its signals can be highlighted provided their frequencies are known in advance.

TABLE 10 Contributions to total exposure of the ten strongest radio signals measured at an example location. Note BS is short for base station

It is normally only necessary to tabulate the ten signals that produce the greatest exposure quotients in order to account for almost all of the total exposure. At the location described in Table 10, the total exposure was 8.6 10–6, or 8.6 millionths, of the NRPB investigation level and the ten listed signals account for 94% of this. The five signals from the base station of interest (in bold) contribute 77% of the total exposure at this location.

The source categories that were most frequently identified in the tables were broadcast radio (88–108 MHz), paging (138 and 153 MHz), broadcast television (470–860 MHz) and various mobile (including cellular) communications transmitters.

3. Surveying technique

1. General method

The complex propagation conditions giving rise to measured signals at environmental locations include shielding and shadowing by buildings, multiple reflections from walls and re- radiation from conducting structures that are excited by the radio waves. These effects can cause electric  field  strength  to  be  non-uniform  over  small  regions  of  space  about  the  measurement locations. The direction of incidence and polarisation of the radio waves can also be unpredictable and similarly variable.

One aim was to record the maximum field strength from every signal over a volume of space  in  order  to  calculate  a  worst-case  exposure  quotient  in  the  immediate  vicinity  of  the measurement location. This was achieved by using the spectrum analyser to make a continuous log of the maximum value of the measured signal strengths over a period during which the antenna was carefully manipulated over the region of space of interest.

2. Limitations

For  signals  from  sources  that  transmit  intermittently,  such  as  VHF  mobile  radio,  the maximum instantaneous power density will have been recorded. Exposure assessments based on protection guidelines require averaging of signal strength over either 6 or 15 minutes, consequently

the exposure component arising from intermittent signals will have been overestimated. This would be appropriate for a hazard assessment because worst-case conditions are represented; however, further work would be necessary to determine the extent of any effect on the exposure proportions calculated in Section 5.

The amplitude displayed by the spectrum analyser indicated the rms voltage at its input connector during the strongest part of each measured signal, ie when the transmission was at its maximum power. Some signals from base stations are pulsed, so the measurements in this report will represent the peak power during a pulse for such signals. The average power of a pulsed signal is  the  appropriate  quantity  for  exposure  assessment,  so  the  spectrum  analyser  results  would generally overestimate exposures. Again, this conservatism is appropriate for hazard assessment, but may indicate an enhanced contribution to total exposure from pulsed signals.

4. Measurement uncertainty

There are several sources of uncertainty associated  with the method used to  measure electric field strength. The uncertainties include:

1. electrical factors associated with the calibration of the spectrum analyser and antennas,
2. factors arising from surveying practices, eg positioning and handling of the antennas.

Uncertainties belonging to category (a) are readily obtained from the relevant calibration certificates. The uncertainty in the calibrations of all three antennas was 0.8 dB and the uncertainty intrinsic to the calibration of the spectrum analyser was pessimistically estimated at 2 dB. The uncertainties associated with measuring the loss of the connecting cable would be negligible in comparison with these other electrical factors.

Uncertainties belonging to category (b) are more difficult to quantify. The most significant source  of  imprecision  was  likely  to  arise  from  disturbance  of  the  antenna  factor  during measurements. The antenna factor quantitatively describes the coupling between an antenna and a uniform field. It is determined for the idealised situation of the antenna being a long distance from conducting and dielectric objects such that mutual coupling is negligible. During the survey the antennas were mounted on tripods and manipulated by hand causing potentially significant coupling of the antenna with the operator's body and with other nearby structures such as building fabrics. Typical values for disturbances to antenna factors were estimated to be 6, 4 and 2 dB for the biconical, log-periodic and ridgeguide antennas, respectively.

The  uncertainties  described  above  were  combined  in  quadrature  to  give  the  overall uncertainty in power density associated with the use of each antenna as summarised in Table 11. It follows that actual power densities may have differed from reported values by factors which are bounded to a 95% degree of confidence by the limits tabulated in the two right-hand columns of the table. Consequently, in measurements made with the log-periodic antenna, for example, there is a 95% certainty that the true maximum power density was not greater than 2.8 times the reported value.

TABLE 11 Overall uncertainties affecting power density measurements

5  Results of exposure assessments

In addition to performing measurements of power density from the base station of concern, signals were measured from other environmental radio sources. Technical data were also secured from the operators that allowed calculations of power density to be performed for comparison with measurement. This section begins by presenting calculations based on the technical data supplied by the operators indicating the extent of the maximum distances from base station antennas at which the NRPB and ICNIRP exposure guidelines are likely to be exceeded. Measured data, given in full in Appendix B, are then summarised and compared with theoretical predictions of power density from base stations. Finally, the relative contributions of different sources to the assessed exposures are compared in order to determine  whether base stations are a  material source of exposure at the locations visited.

1. Calculation of exposures

Several different approaches to calculating the exposure of a person in the vicinity of a transmitting antenna are possible. In general, the simpler the calculation approach used, the more conservative will be the outcome and the greater the compliance distance that will result (see Section 5.2). Three different calculation approaches will be mentioned here and results will be presented based on the first two.

1. Far-field power density calculation

If the total power fed into an antenna is known as well as the antenna gain, it is possible to calculate the power density in the main beam by assuming an inverse square law dependence upon distance at all distances from the antenna. The following equation is analogous to equation 2 in Section 2.2.1 and may be used to calculate power density, S, in this way.

P G

S  rad (9)

4  d 2

where d is the distance from the antenna, Prad is the total radiated power and G is the antenna gain (in linear units).

Reflections may increase or decrease the power density from that calculated by equation 9 if the path length travelled by a reflected wave is comparable with the direct distance to an antenna. The most likely situation where this could occur would be where a wave was reflected from the ground, as indicated in Figure 12. Reflections would be expected to increase the electric field strength by a factor of up to two, thus the total power density would be increased by a factor of up to four.

Base station

mast Measurement position

Ground

FIGURE 12 Direct and reflected waves arriving together to increase the power density at a measurement position

The use of equation 9 will overestimate power density in directions other than the main beam, because the antenna gain is effectively less in these directions. It will also overestimate power density at short distances, generally within 10 m of antennas.

2. Near-field power density evaluation

If the detailed electrical structure inside an antenna is known, it becomes possible to perform a more rigorous calculation of power density than is represented by equation 9. Such a near-field calculation is able to evaluate the precise variation of power density (or field strength) at all distances and in all directions from an antenna. Near-field results are presented later in this section that have been calculated using a physical optics approach.

If field strength is measured close to an antenna, calculations accounting for the near-field character of the antenna should be in good agreement.

3. Direct SAR evaluation

The most rigorous approach to calculating the variation of the exposure produced by an antenna as a function of distance is to account fully for the electromagnetic coupling between the antenna and an exposed person. This approach should be accurate for exposure at all distances and will generally give the shortest compliance distances. The approach is similar to that used for assessing the exposures produced by mobile phone handsets and can be used for microcellular base stations which radiate similar power levels.

A computer model of an antenna and of the human body can be used for analysis of energy deposition in the body using a computer code implementing Maxwell's equations. Anatomically realistic models of the human body developed from magnetic resonance imaging form the current state of the art and the finite-difference time-domain method (FDTD) is often used to carry out the exposure calculation.

An alternative approach is to carry out SAR evaluation through measurement by producing a physical model of the human body and placing this in the field of the real antenna. Small field probes can be implanted inside the physical model and, if sufficiently sensitive for the powers transmitted, these can be used to measure the SAR that is produced.

2. Predicted compliance distances

The general approach to evaluating compliance distances is to calculate the variation in some measure of exposure (field strength, power density or SAR) in the space around an antenna. The distance is then ascertained at which the appropriate guideline value is met.

Network operators usually calculate compliance distances in a generic rather than a site- specific way for classes of antennas used in their networks. The calculations are performed using the maximum power that will be radiated by any antenna in that class so that worst-case conditions are represented. As a result, the calculations do not have to be repeated if another transmitter is added or the radiated power is increased. An effective closed surface, often a cuboid, is then defined around sector antennas so that exposure will not exceed guidelines provided people are excluded from this region. Cylindrical exclusion zones are normally defined around omni- directional antennas.

The results in this section serve as example calculations, based on realistic technical data that are expected to reproduce typical worst-case conditions. The results are not intended to supersede or replace the exclusion zones defined by network operators. Results are only given for the direction of the main beam. Compliance distances would be less in other directions.

1. GSM900 base station antennas

A model of the electrical characteristics of a typical 900 MHz sector antenna has been formed by producing a vertical stack of  12  half-wavelength  dipole  elements  with  appropriate separation. Each element was then excited to give an omni-directional antenna with a total radiated power of 80 W and a gain of 13 dB. The radiated power of the omni-directional antenna was then increased by a factor of about three to represent a 120º sector antenna and this effectively increased the gain to 17 dB. The model antenna was then used to produce results that could be used to indicate typical maximum power densities produced by GSM900 sector antennas. The predictions from the  model  are  shown  in  Figure 13,  together  with  predictions  from  a  simpler  inverse  square law calculation.

Figure 13 shows how power density reduces with distance from the antenna and indicates compliance distances for the NRPB investigation level and the ICNIRP reference level. Assuming an inverse square law would give a compliance distance of 3.1 m for the 35 W m2 NRPB investi- gation level, whereas the more accurate near-field model indicates a lesser compliance distance of 46 cm. Similarly, the inverse square law would imply a compliance distance of 8.4 m for the 4.6 W m2 ICNIRP reference level for the general public, whereas the near-field calculation would indicate 2.4 m.

1000 100 10

1 0.1

0.1 1 10 100

Distance (m)

FIGURE 13 Power density in the beam from a typical GSM900 sector antenna showing the results from calculations assuming an inverse square law and a more detailed near- field model of the antenna. The total radiated power is 80 W and the antenna consists of 12 radiating elements giving a gain of 17 dB

2. GSM1800 base station antennas

A similar approach to that described for the 900 MHz antenna, but scaled for 1800 MHz, was used to calculate maximum power densities produced by GSM1800 sector antennas. Again, a radiated power of 80 W was chosen, but a slightly higher gain of 18 dB was used. The predictions from the model are shown in Figure 14, together with predictions from a simpler inverse square law calculation.

Figure 14 shows how power density reduces with distance from the antenna and indicates compliance distances for the NRPB investigation level and the ICNIRP reference level. Assuming

an inverse square law would give a compliance distance of 2.1 m for the 100 W m2 NRPB investi- gation level, whereas the more accurate near-field model indicates a lesser compliance distance of 31 cm. Similarly, the inverse square law would imply a compliance distance of 6.7 m for the

9 W m2 ICNIRP reference level for the general public, whereas the near-field calculation would indicate 5.9 m.

1000 100 10

1 0.1

0.1 1 10 100

Distance (m)

FIGURE 14 Power density in the beam from a typical GSM1800 sector antenna showing the results from calculations assuming an inverse square law and a more detailed near- field model of the antenna. The total radiated power is 80 W and the antenna contains 12 radiating elements giving a gain of 18 dB

3. Power density at ground level

The model GSM900 sector antenna of Section 5.2.1 was adapted to calculate the variation in power density at ground level with increasing distance from a typical mast. The beamwidth produced by the antenna was 6º in the plane of elevation; however, a phase shift was applied to the element excitations in order to tilt this beam downward by 2º. The radiated power was set to give an EIRP of 5000 W, which would correspond to a total radiated power of 80 W from a sector antenna with 18 dB of gain.

The calculation geometry is shown in Figure 12 and the base of the antenna was set at a typical minimum height of 13.7 m above ground level, giving 15 m to the centre of the antenna. Power density was calculated at distances ranging from 1 m to 10 km from the foot of the mast and the direct wave power density was multiplied by a factor of four in order to account for the presence of a reflected wave component. The results are shown in Figure 15.

Figure 15 shows how the power density at the foot of the mast is very much lower than at distances  in  the  range  from  tens  to  hundreds  of  metres.  At  these  distances,  sidelobes  in  the antenna's directional pattern give rise to a series of peaks in power density. The lower edge of the main beam begins to become incident at ground level from distances of 100 m; however, the peak power density of 35 mW m2 is not achieved until a distance of 180 m from the mast. Power density then falls off according to the inverse square law at greater distances.

35 30 25 20 15 10 5 0